document resume ed 101 183 ce 002 952 levine, eugene, ed. · 2014-01-14 · ce 002 952. levine,...

ED 101 183

AUTHORTITLE

INSTITUTION

REPORT NOPUB DATENOTE

AVAILABLE FROM

EDRS PRICEDESCRIPTORS

DOCUMENT RESUME

CE 002 952

Levine, Eugene, Ed.Research on Nurse Staffing in Hospitals; Report ofthe Conference, May 1972.National Institutes of Health (DREW), Bethesda, Md.Div. of Nursing.DREW-(NIH)-73-43-1May 72193p.; Report of the conference (Fredericksburg,Virginia, May 23-25, 1972)Superintendent of Documents, U.S. Government PrintingOffice, Washington, D.C. :n402 (Stock Number1741-00055, :2.35)

MF-80.76 MC-89.51 PLUS POSTAGEAdministrative Problems; *Conference Reports; Costs;Evaluation Criteria; Facility Requirements; *HospitalPersonnel; Human Services; Information Systems;Medical Services; Nurses; *Nursing; Patients(Persons); Personnel Needs; *Research; Research .

Methodology; Research Needs; Speeches; *StaffUtilization

ABSTRACTThe conference brought together 45 persons who have

had extensive experience in nurse staffing research. After theleadoff paper by Myrtle K. Aydelotte, which presented an overview ofnurse staffing research, 10 papers were presented on variablesconsidered to be significant in influencing the quantitative andqualitative demand for nurse staffing. These included patients'requirements for nursing services, architectural design of thehospital, administrative and cost factors, and social-psychologicalfactors. In addition, papers were presented on the evaluation ofquality of nursing care and the impact of computerized informationsystems on staffing. The papers were followed by presentations bydiscussion leaders who addressed themselves to points raised in thepapers. These presentaions, the opening remarks of Jessie N. Scott,and a summary and synthesis of the entire conference are alsoincluded in the publication. Appended are four task force reportsoffering recommendations on future directions for research.(Author/MW)

BEST COPY AVAILABLE in 6

U S. DIPARTMINT OP HEALTH,EDUCATION & WILPARSNATIONAL INSTITUTE OP

EDUCATIONTHIS DOCUMENT HAS SEEN REPROOUCEO ExACTL Y AS RF.CEIVEO PROMTHE PERSON OR ORGANIZATION ORIGINMING IT POINTS OF VIEW OR OPINIONSSTATED 00 NOT NECESSARILY REPRESENT OFFICIAL NATIONAL INSTITUTE OFEDUCATION POSITION OR POLICY

rch on \urse Stactg in HosReport of the Conference

ottals

i'robibitedTit le of the Civil Rights ,\(-t ol I!Hstates; "No etson in the !tined states shall. on the ground of um C.Color, or national origin, he excluded how patticipation III, be

denied the Lic.nefits ol, or he scil.ijected to discrimination 1111(h.)' any,progiam or ii(iiity teceiving Federal financial assistan(e." There-tore, the ntirse stalling progiant, like eve, y program or aclivilyreceiving financial as5istan«e nom the 1)cp11r11tent of I lealth, 1(111-tat ion. and 11'ellarc, must he operated in compliance with this law.

1))1 %.11(' I) 111). S1111e1.1111(1)11(.111 of 11))1I nrnrnt., 41, (,11%rititilr111 Ptin,ingI) ( .1.1)1112 Piny. ':' ';`1

Sin( I, \mullet 1711 mu))

Research on Nurse Staffing in Hospitals

Report of the ConferenceMAY 1972

Conducted by

DIVISION OF NURSING

Eugene Levine, Ph.D.Scientific Editor

DHEW Publication No. (NIH) 73-434

U.S. DEPARTMENT or HEALTH, EDUCATION, AND WELFAREPUBLIC HEALTH SERVICE NATI.:NAL INSTITUTES OF HEALTH

Bureau of Health MON.te"orf Education Division of Nursing

ForewordThis publication contains a report on the Conference on Research on Nurse

Staffing in Hospitals, sponsored by the Division of Nursing and held in Fred-ericksburg, Virginia, May 23 through 25, 1972.

I am hopeful that it will prove to be a valuable addition to the literature onnurse staffing. As the opening paper by Dr. Myrtle K. Aydelotte makes clear,the existing literature has many deficiencies, not the least of which are the manygaps in knowledge. The papers presented at the conference contain the latestfindings on many important areas of nurse staffing research and will help toclose they! gaps.

The Division of Nursing's concern with the problems of nurse staffing goesback many years as we have attempted to improve the quality and quantity ofnursing care in the Nation. We have supported considerable research in thisarea, both by our own staff and through grants and contracts to researchers inuniversities and professional organizations and in hospitals and other healthcare settings. The many publications issued by the Division in addition to studyfindings include descriptions of methodology for studying the use of nursingskills, assessing the quality of nursing care, and planning for better matching ofnursing resources and needs. The conference and this resulting publication arenatural extensions of our efforts to improve nurse staffing.

This publication is a companion volume to Dr. Avdelotte's comprehensivecritique of nurse staffing methodology. These two publications should lie usefulto researchers, educators, administrators, practitioners, and all who are concernedwith the improvement of nursing staffing in hospitals.

Juan M. ScorrAssistant Surgeon GeneralDirectorDivision of Nursing

ill

E.

CONFERENCE, PLANNING COMMITTEE

Dr. Eugene Levine, Chairman

Dr. Myrtle K. Aydelotte

Dr. Bernard Ferber

Miss Marie R. Kennedy

Miss Delores M. LeHoty

Dr. Mary K. Mullane

Miss Jessie M. Scott

Mrs. Margaret Sheehan

Mr. Stanley E. Siegel

Miss Elinor D. Stanford

iv

CONFERENCE PARTICIPANTS

Dr. Myrtle K. AydelotteDirector, NursingUniversity Hospitals and ClinicsProfessor, College of NursingUniversity of IowaIowa City, Iowa

Dr. Alexander BarrChief Records Officer and StatisticianOxford Regional Hospital BoardHeadington, OxfordEngland

Mr. Jose Blanco, Jr.Controller and Assistant AdministratorWashington Hospital CenterWashington, D.C.

Dr. Doris BlochExecutive SecretaryNursing Research ant! Education

Advisory CommitteeDivision of NursingNational Institutes of HealthBethesda, Maryland

Dr. Mario F. BognannoAssistant ProfessorUniversity of MinnesotaIndustrial Relations CenterMinneapolis, Minnesota

Dr. Virginia S. ClelandProf essoa. of NursingWayne State UniversityDetroit. Michigan

Mrs. Margaret Ellsworth763 St. Charles StreetElgin, Illinois

Dr. Bernard FerberDirector, Bureau of Research ServicesAmerican Hospital AssociationChicago, Illinois

7v

Dr. Charles D. Flag leDirector of Operations ResearchSchool of Hygiene and Public HealthJohns Hopkins University HospitalBaltimore, Maryland

Mr. Richard G. GardinerOperations Research SpecialistHospital Association of New York StateA1L.iny, New York

Mrs. Phyllis GiovannettiProject DirectorNursing ResearchHospital Systems StudyL niversity of SaskatchewanSaskatoon, SaskatchewanCanada

Mr. John GriffithProfessor and DirectorProgram in Hospital AdministrationSchool of Public HealthUniversity of MichiganAnn Arbor, Michigan

Miss Verna GrovertStall Assistant to the DirectorHealth Systems Research and

Development ServiceVeterans Administration Central OfficeWashington, D.C.

Dr. Elizabeth HagenProfessor of PsychologyTeachers CollegeColumbia UniversityNew York, New York

Dr. Richard HowlanclWestinghouse Elect sic CorporationRescarch and DevelopmentPittsburgh, Pennsylvania

Conference Participants (continued)

Dr. E. Gartly JacoProfessor, Department of SociologyUniversity of CaliforniaRiverside, California

Dr. Stanley JacobsAmerican Hospital AssociationChicago, Illinois

Dr. Richard C. JelinekVice President for Research and DevelopmentMedicus Systems CorporationChicago, Illinois

Miss Angie KammeraadRegional Nursing ConsultantDivision of NursingNational Institutes of HealthDHEW Region IVAtlanta, Georgia

Miss Marie R. KennedyNurse ConsultantNursing Education BranchDivision of NursingNational Institutes of HealthBethesda, Maryland

Miss Dolores M. LeHotyProgram AnalystManpower Evaluation and Planning BranchDivision of NursingNational Institutes of HealthBethesda, Maryland

Dr. Eugene LevineChief, Manpower Evaluation and Planning BranchDivision of NursingNational Institutes of HealthBethesda, Maryland

Miss Ruby M. MartinDirector of Nursing ServicesThe Ohio State University HospitalColumbus, Ohio

Miss Marylou McAthie .

Regional Nursing ConsultantDivision of NursingNational Institutes of HealthDHEW ".egion IXSan Fran4 isco, California

8

Dr. Wanda E. McDowellAssociate Dean for ResearchUniversity of MichiganSchool of NursingAnn Arbor, Michigan

Dr. Mary K. MullaneProfessor, College of NursingUniversity of IllinoisChicago, Illinois

Mr. George J. MullenSenior EditorEngineering Services and PublicationsGaithersburg, Maryland

Dr. Maria PhaneufProfessor of NursingCollege of NursingWayne State, UniversityDetroit, Michigan

Dr. Lillian M. PierceProfessor, School of NursingThe Ohio State University HospitalColumbus, Ohio

Miss Pamela PooleNursing ConsultantDepartment of National Health and WelfareOttawa 3, OntarioCanada

Dr. Muriel PoulinBoston University School of NursingBoston, Massachusetts

Dr. Reginald W. RevansSenior Research Fellow an -1 Scientific AdviserFondation Industrie-University1030 Brussels, Belgium

Miss Jessie M. ScottAssistant Surgeon GeneralDirector, Division of NursingNational Institutes of HealthBethesda, Maryland

Mrs. Margaret F. SheehanNurse ConsultantNursing Education BranchDivision of NursingNational Institutes of HealthBethesda, Maryland

vi

Conference Participants (continued)

Mr. Stanley E. SiegelAssistant Chief, Manpower Evaluation

and Planning BranchDivision of NursingNational Institutes of HealthBethesda, Maryland

Dr. Donald W. SimborgChief, Clinical Systems DevelopmentJohns Hopkins University HospitalBaltimore, Maryland

Miss Marjorie SimpsonNursing Officer in ResearchDepartment of Health and Social SecurityAlexander Flemming HouseLondon, England

Miss Marlene R. Slawson-Research AssociateOperations Research UnitDivision of Community PsychiatrySchool of MedicineState University of New York at BuffaloBuffalo, New York

Dr. Frank A. SloanDel art ment of 'ZconomicsUri; versity of FloridaGainesville, Florida

7

Miss Elinor D. StanfordDirector, Nursing Research Field CenterDivision of NursingNational Institutes of HealthSan Francisco, California

Miss Bernice SzukallaRegional Nursing ConsultantDivision of NursingNational Institutes of HealthDHEW Region VIIIDenver, Colorado

Miss Deane B. TaylorAssociate ProfessorState Universir' of New York at BuffaloSchool of NursingBuffalo, New York

Dr. Mabel A. WandeltProfessor of NursingWayne State UniversityDetroit, Michigan

Dr. Gerrit WolfAssistant Professor of Administrative

Sciences and PsychologyYale Universit.)New Haven, Connecticut

Dr. Harvey WolfeAssociate ProfessorSchool of EngineeringUniversity of PittsburghPittsburgh, Pennsylvania

vii

Contents

Page

ForewordConference Planning Committee iv

Conference ParticipantsFiguresTables xi

Introduction 1

Opening Remarks 3

Miss Jessie M. ScottState of Knowledge: Nurse Staffing Methodology 7

Myrtle K. Avdelotte, Ph.D.Factors Related to Nurse Staffing and Problems of Their Measurement __ 25

Charles D. Flagle, Dr.Eng.Discussion of Dr. Flagle's Paper 35

Measurement of Patients' Requirements for Nursing Services 41

Mrs. Phyllis GiovannettiDiscussion of Mrs. Giovatinetti's Paper 57

Nurse Staffing Patterns and Hospital Unit Design: An ExperimentalAnalysis 59

Dr. E. Gartly jacoDiscussion of Dr. Jaco's Paper 77

Impact of Administrative and Cost Factors on Nurse Staffing 79

Mr. John R. GrimthDiscussion of Mr. Griffith's Paper 98

Nursing Directors as Participants in Hospital Peer Review Boards 101

Dr. Gerrit WolfSocial Psychological Factors Relating to the Employment of Nurses 111

Dr. Virginia S. ClelandDiscussion of Dr. Wolf's and Dr. Cleland's Papers 122

Evaluation of Quality of Nursing Care 125

Dr. Mabel A. Wandei:Discussion of Dr. Wandelt's Paper 133

Psychosocial Factors in Hospitals and Nurse Staffing 137

Dr. Reginald W. RevansImpact of Computerized Information Systems on N .:se Staffing in

Hospitals 151

Donald W. Simborg, M.D.

ix

10

Contents (continued)

Title Page

Discussion of Dr. Simborg's Paper . _ 164

Concept Modeling in Hospital Staffing 167Dr, Lillian M. Pierce

Discussion of Dr. Pierce's Paper . _ _ _ _ - _ 174

Summary, Synth's's, and Future Directions _ 177Dr. Mary K. Mullane

Appendix: Task Force Reports . 181

1

Figures

Title Page

Scheme of nursing care delivery system 9Diagram of evolvement of methodologies for the study of nurse staffing:

application of logic and increases it abstractions 15

Calculation of nursing staff needs 28Interrelationship of staffing and patient carePatient classification tool ____ 45Graph of patient load index 49Graph of daily workload index 50Nursing: An interdependent variable in the health cattlystem 52Radial and angular hospital units . 61Marginal utility of nursing care 82High and low region medium AM &.S nursing hours per patient day 86Prototype outcomes quality assessment in .trument, patient observation

section, diabetes mellitus 89Flow chat of weight order _ _ _ 153Flow chart of vital sign order 154Flow chart of intravenous fluids order _ 155Physicians' orders study 156

Functional requirements of a new system to carry out physicians' orders __ 156General flow chart of new system 158

System output documents 158Team action sheet 159Patient care model 170

Tables

Title Page

Comparison of staffing ,methodologies on selected features _ - _ 16

Ratios of direct nursing care time received per shift and patientcategory _ _ _ _ _ _

Daily patient load index 47

Conversion of index to staff required _ 47

Major type of nursing care in an acute, general hospital 65

Major type of nursing care, by six nurse staffing patterns ______. 66

Major type of mirsing (are, by nurse staffing patterns and by radialand angular units, percentages only 66

Medical, physical, and social types of direct care, by number ofobservations and percentages, total study, by shape of unit _ 69

Type of direct patient care, by nurse staffing patterns _ 69

Type of direct patient care by nurse staffing patterns, by radial andangular units, percentages o n l y. . . . . . . . . . . . . . . . . . . . . . . . . . . . 711.

Ratios of patient involved to nonpatient involved care, by total staff

and by nurse staffing patterns 70

Ratios of patient involved to nonpatient involved care, by nursestaffing patterns and radial and angular units _..______ _ 71

Types of trips to patients' mums by nursing staff in radial and angularunits, by number of observations and percentages __ _ _ 71

Average daily number of trips to patients' rooms by nurse_ staffingpatterns, by total trips and by types of trips 72

Average, daily number of trips to patients' rooms by types of trips,by nurse staffing patterns, and by radial and angular units 72

Utilization of nursing unit by total nursing staff in radial and angularunits, by number of observations and percentages ___ _ _ 72

Utilization of nursing unit by nurse staffing patterns, by number ofobservations and percentages 73

Utilization of nursing unit by nurse staffing patterns and radial andangular units, by percentage of observations only 74

Location of MST's employed in hospitals in relation to hospital beds:1968 84

Medan nursing hours per adult medical surgical patient day, HASsubscriber hospitals, 6 months ending 12/31/71 85

Percentiles of nursing hours per medical surgical patient day in HASsubscriber hospitals, 3 months ending 10/31/71 87

Interrater reliabilities for all strategy items 107

Mean contribution and influence rating according to role (N 15) _ _ 107

Nursing activity study _ 161_

Summary of repeat studies on experimental floor _ _ 161

Sample nursing order: "NG Tube Care" __. _161

Total number of nursing tasks on one nursing unit on selected days _ 161

xi

Introduction

The Conference on Research on Nurse Staffing in Hospitals brought together.45 persons from a variety of disciplines who have had extensive experience in

nurse' staffing research. The conference was held in Fiedericksburg, Virginia,May through May 25, 1972.

After the leadoff paper by Dr. Myrtle K. Aydelotte, which presented an over-view of rtrse staffing research, 10 papers were presented on variables con-sidered to he significant in influencing the quantitative and qualitative demandfor nurse staffing. These included patients' requirements for nursing services,architectural design of the hospital, administrative and cost factors, and social1.sychologkal factors. In addition, papers were presented on the evaluation ofquality of nursing care and the impact of computerized information systemsoo staffing.

Ti,e papers were followed by presentations by discussion leaders who ad-dressed themselves to points raised in the papers. These presentations, whichwere planned to stimulate the discussion during the period that followed eachpaper, zre also contained in this publication.

Also included are the opening remarks of Miss Jessie M. Scott, Director ofthe Division of Nursing, and a summary and synthesis of the entire conference

by Dr. Mary K. Mean:-In addition to p rtici pa ting in the general discussion, each person attending

the conference was assigned co a task force. The purposes of the task forceswere to delineate areas of needed research in nurse staffing in hospitals, deter-mine research priorities, and explore means of coordinating the work of re-searchers in this field. Reports of the four task forces are contained in theappendix.

Taken as a whole, the material presented in this publication serves severalpurposes. First, it brings together the latest findings on important areas of re-search into factors related to nurse staffing in hospitals. Second, it points upsome of the ;;aps in existing knowledge and suggests how these may be closed.Finally, i; constitutes an important step forward in correcting the meager, diffuse

condition of the existing research in hospitals.

I

Eugene Levine, Ph.D.Chairman, Planning CommitteeConference on Research on Nurse

Staffing in Hospitals

13

Opening Remarks

Miss Jessie M. Scott

Assistant Surgeon GeneralDirector, Division of Nursing

Bureau of Health Manpower EducationNational Institutes of Health

Greetings and Welcome

I am very pleased to welcome all of you to theDivision of Nursing Conference on Research onNurse Staffing in Hospitals. This conference is theculmination of 6 months of planning by a com-mittee made up of our Division staff, whose namesare listed in your programs: Dr. Bernard Ferber;Dr, Myrtle Aydelotte, who will serve as conferencechairman; and Dr. Mary Kelly Mullane, who willbe summarizer. The committee tells me their workwas made much easier by the excellent cooperationthey received from each of you in accepting as-signments to prepare papers or to serve as discus-sion leaders and in making your travel arrange-ments to the conference center.

Each of you was selected as a participant in theconference because of the research you have done

in the field of nurse staffing or because of otheractivities that have advanced the knowledge orunderstanding of the field. We indeed feel wehave a truly distinguished group of participants.You represent different professional backgroundsand different geographic settings.

We are pleased to note the presence of partici-pants from England and Canada. Among you arenurse administrators, nurse educators, and nurseresearchers, economists, industrial engineers, sys-tems analysts, statisticians, psychologists, sociolo-gists, computer scientists, and hospital administra-tors. You have written about and have studiedthe problems of nurse staffing deeply and exten-sively. We hope you will share with us during thecourse of the conference your activities and ideas,

Background of the Conference

The Division of Nursing, which is part of theBureau of Health Manpower Education in theNational Institutes of Health, has as its primarymission the improvement of the quantity and qual-ity of nursing manpower. Although the major

3

thrust of our activities is in the area of nursingeducation, we have been concerned over the manyyears with the improvement of nursing services.To catalog the many activities that the Divisionhas been engaged in over the past 20 years in the

14

4 RESEARCH ON NURSE STAFFING IN HOSITAIS

area of nurse stalling, both inn amtirally and extra-murally. would require much more time than Ihave allotted. but let me cite a few examples.

In the early 1950's, concerned with the severeshortage of nursing personnel in hospitals andthe widespread desire of hospitals and niksingadministrators to have numerical guidelines forstaffing, our Division initiated a study that becameknown as "the green book." This was a designto test the relationship between nursing care andpatient welfare. The central hypothesis of thisdesign was that if nursing care were varied quanti-tatively and qualitatively the effects of these changescould he measured.

This design became the basis for several majorstudies. One was directed by our chairman, Dr.Aydelotte. Another was conducted in Kansas Cityby the Community Studies Organization. A thirdwas conducted by members of our own Division,which attempted to establish a relationship be-tween nursing hours available to patients, andpatient and personnel satisfaction with care.

In the area of industrial engineering applicationsto nurse staffing, the Division of Nursing has sup-ported several important projects that have beencarried out extramurally. One of these, conductedat Johns Hopkins Hospital under the directionof Dr. Charles Flagle, has inspired considerableapplication of industrial engineering methods tothe problems of nurse staffing. These methods in-

dude wot kload measurement, work simplification,task analysis, mathematical modeling, and qualitycontrol.

In recent years the attention of our Divisionstaff has turned toward a broadened look at thenurse staffing problem in hospitals. Our statisticalstudies have shown that quantitative levels of nursestaffing in hospitals are influenced by a greatvariety of factors, including the needs of the pa-tients, tin. administrative organization and admini-strative style, the characteristics of the nursingstaff, the physical plant, the economic structurewithin and outside the hospital, and a host ofpsychological and sociological factors. A recentstudy we supported, known as The Illinois NurseUtilization Study, attempted to measure the effectsof many of these factors on nurse staffing in asample of 31 hospitals. The project director, Mrs.Margaret Ellsworth, is here with us and I am sureyou will be hearing about the methodology andresults of this study in the course of the con-ference.

Also among our current concerns and activitiesis the matter of the expanded role of the nurse.With the health system undergoing rapid changesand the rol.' of the nurse in this system widening,it is important for us to anticipate these changesand to he prepared to provide the educationalbase necessary for such an expanded role.

Objectives of the Conference

We could say the main theme of this conferenceis an examination of the important factors thataffect nurse staffing. We are going to hear aboutsome of the methodology for measuring the impactof these factors and we shall examine the findingsof studies directed toward specific factors suchas patient requirements for nursing services, ar-chitectural design, economic costs, administrativetechniques, and psychological and sociologicalfactors. We shall hear about the problems of measurement in nurse staffing research, including meas-urement of the quality of nursing care. Finally,we shall hear about the work being clone in apply-ing computerization to nurse staffing and the im-pact this could have on the quantity and qualityof staffing.

We are limiting ourselves to nurse staffing inhospitals with the full realization that out-of-hospital health care settings are becoming increas-ingly important in the delivery of health services.We are focusing on hospitals because, frankly, mostof the research has been done there. This is sobecause hospitals are the largest employers ofnursing personnel, employing over 60 percent ofthe 748,000 registered nurses actively employed innursing and over 90 percent of the more than onemillion licensed practical nurses and nursingaides, orderlies and attendants.

I feel that much of the research in hospitalnurse staffing can have useful application in otherhealth care settings. Specifically, I see loin' im-mediate and long range objectives of our con-

OPENING REMARKS

ference. First, I see this o oulerence as a supple.

ment to Dr. yole lotte's study entitled N'nr.se

fug i1fellsmi1ogy: A Review And Critiglw of Se-lected Literature.' In this study, done under con.tract with our Division, Dr. .\vtlelotte has identi-fied a number of important areas of research onthe factors we have cited as related to nurse staff-ing. Her study has identified the researchers inThese areaS, and it was on the basis of her reportthat we were able to identify participants for thisconference and to get a perspective on the workthey have been doing. In fact, the research ofmany of those present here today is discussed inDr. Aydelotte's report. So, in a sense this conference'brings her report to life. The proceedings of thisconference is a companion volume to Dr. Ayde.

lode's study.A second objective is to bring together a group

of knowledgeable researchers, management engi-neers, administrators, consultants, and educators

to exchange information and ideas about the

problems of muse staffing. We shall be particularlyconcerned with the factors that influence nursestaffing, either positively or negatively, as well asquantitatively or qualitatively. It is my feelingthat a great deal of research and methodologicaldevelopment are going on in various places butthat there is not enough communication amongthe different persons engaged in these activities.

One of Dr. Aydelotte c conclusions is that re-in writing up their work. only infre-

quently cite previous work. To me this indicatesa lack of information exchange among those work-ing in the field. Hopefully, this conference will hethe starting point for opening up channels of com-munication that will continue beyond the con-ference.

A third and most important purpose of the con-ference is to set forth a progrun of research onnurse staffing for the future. We shall learn fromour presentations and discussions that some signif-icant research has been conducted. We shall hear

about the methodologies and the findings of theseprojects and from this should be able to identifythe important gaps in 'enowledge and methodology.From our deliberations we can then chart a plan

U.S. Department of Health, Education, and Welfare.Public Health Service, National Institutes of Health. DHEWPublication No. (N111) 73-133, Washington, D.C.: l'.5.

Government Printing Office. January 1973.

5

for the future for the kinds of studies that shouldIe undertaken, including suggestions as to howthey might he initiated and with what kinds of

support.This support does not necessarily have to come

only from the Federal Government. I am oftenpleasantly surprised by how intic'.1 useful researchhas been accomplished on limited resources. Al.

though the problems of nurse stalling are large.useful solutions can sometimes he found throughsmall, modest investigations, including doctoraldissertations and even master's theses, Let us thenkeep in mind as the conference proceeds that wewould like to delineate areas for further research,

and we invite suggestions about the most prac-tical, feasible, and productive ways of pursuingthis needed research. Our small group sessions onWednesday afternoon will be specifically directedtoward this subject and will provide for each of

you an opportunity to share your ideas.

Finally, 1 would hope that the proceedings of this

conference could provide educational materials

for nursing service administrators and other practi-tioners as they face their day-to-day problems ofruntthig their health care instittitions. A.great deal

has been written about research on nurse staffing

in hospitals, but much of it is of a technical andsomewhat specialized nature. As I said earlier,

a major mission of our Division is to help advancethe education of nursing personnel, and I hopethe proceedings of this conference will give ad-ministrators some assistance in managing their in-

stitutions.I hope you will be moialed to participate in

the discussions as much as possible. Each presenta-

tion will he followed by a 45-minute discussion

period. I cannot urge you too strongly to partici-

pate to the fullest extent possible during thesediscussion periods. Lad) of you has been selected

because of your extensive background and interest

in this field, and I am sure you have many worth-

while ideas and much information to share with us.

In conclusion, let me say that I anticipate thisconference will be a rewarding and fruitful ex-perience for all of us and that it will provide guide-

lines for future conferences of this kind and laythe groundwork for continuing dialogue. Now Iwould like to introduce your conference chairman,

Dr. Myrtle Ay0elotte. Dr. Aydelotte is Directorof Nursing, lJniversity Hospitals and Clinics, and

16

6.

a professor at the College of Nursing, Universityof Iowa in Iowa City. As 1 ILive already mentioned,she was the project director of a very importantstudy of the relationship between nurse staffing

RESEARCH ON NURSE STAFFING IN HOSPITAIS

and patient welfare conducted during the late1950's and recently has completed a comprehensivecritique of nurse staffing methodology. 1 place thisconference under her most capable direction.

State of Knowledge:

Nurse Staffing Methodology

Myrtle K. Aydelotte, Ph.DDirector of Nursing, University Hospitals and Clinics

andProfessor, College of Nursing

University of Iowa

Introduction

Many individuals working within the deliveryand education systems of the health field recognizethe need for a well defined body of knowledgethat can serve as a basis for nurse staffing meth-odologies. Researchers from several disciplines haveattempted to study the effect of one or more vari-ables upon staffing, and methodologists have triedto develop specific approaches for determiningnurse staffing needs. Often, however, these effortshave been carried out in isolation and have beeninadequately reported to the profession. Many havealso been fragmented in approach and have notdealt with all the relevant problems. Frequently,such studies have been prompted by the need forthe economic use of personnel and have ignoredvariables other than cost, Unfortunately, no personor group has assembled and assessed the literatureon the subject of staffing.

In May of 1970, I was approached by the Divisionof Nursing with the request that I undertake thistask. Work on the project' began late in June

'Contract N u m her NI H-70-4193.

7

18

1970 and was completed in December 1971. Thelengthy, frustrating search of the literature re-sulted in the identification of over a thousandpieces related to the subject. After screening, 182of those that seemed most important and represen-tative were selected, reviewed, and critiqued.

The primary aim was to "screen, review, andcritique relevant research, studies, and approachesfocused on measuring nursing care needs in thehospitals," Development of the research protocolinvolved preparing statements of objectives, specify-ing criteria for the estimation of relevance andvalue, formulating definitions of terms for researchand classification and constructing a glossary ofterminology. Procedures for review and critiquewere outlined and followed.

Each research paper was examined to determineits value as a research study with regard to thefollowing:

the significance of the problem to the fieldof nursing, to nurse staffing methodology,

the theoretical framework for investigation,

R

the aspects 1' rtinent to the measurement ofnursing are requii villeins and prediction ofnurse staffing needs,

the relevance of the research design to generalproiaents of nurse staffing,the use of specific innovations, including adap-tations of other approaches,

the variables treated in the study that weresignificant to the methodology of staffing,the tools and instruments most easily trans-lated into present day use,

the feasihility of applying any portion of theresearch to staffing methodology.

This review dealt primarily with books, includ-ing monographs, research reports, and theses. Of-literature in this assification, 77 pieces weremajor research reports. Certain other materials,which did tat come under any of these headings.were also examined. Unfortunately, most of thesedid not provide enough detail to make possiblethe application of the criteria developed for the

RFNEARcu ON NuRsE svm; F !NG nosprrAt s

above points. Many of them did not, in any case.Import to he scholarly undertakings. These itemswere read for whatever assistance they could pro-vide in chit ifying and identifying ideas that mightbe used itt approaches to the question of nursestaffing.

The variation in types of literature reviewedmakes a general summation of the analysis dif-(alt. The materials were diverse in many re-spects. They varied in the purposes for which theywere written; in publication outlet; in the detailof documentation of content; and 'above all, inscholarly skill as measured by Organization of con-tent, quality of source materials, logical rigor, styleof writing, and presentation of new knowledge.

There is no dearth of literature on the generalsubject of nurse staffing, but the unniher of tightlyconceived, well structured explorations of the cen-tral issues is sadly limited. There are only a fewcomprehensive, well designed research studiesdirected toward examining it specific staffing modeland its impact on patient care and cost.

General Characteristics of the Literature

Most of the literature reviewed was poor inquality. Though this judgment is of necessityporely subjective. it is at least based on an ap-praisal of each study with respect to the eightpoints mentioned at the outset. Many authors failto give a thoughtfui, carefully stated descriptionof a problem and its analysis. For the most partthey do not provide reviews of the literature orreferences to related research studies, even whenthese are needed to support the presentation. Theo-retical framework, or at least some embryonicstructure for a rationale, is absent in many studiesand reports. Tile lack of acknowledgments, evenwhen similar points are made, suggests that sometopics were explored concurrently by writers work-ing Alependently. It is difficult to determinethe impact of specifit works or the links betweenindividual efforts.

The number of variables treated in differentstudies ranges from a few to hundreds. The serviceoffered in a nursing unit depends upon manyfactors other than the requirements of the patients.It depends upon the purpose of the unit, medicaland nsing programs of care, supporting services

19

made available to the nursing unit, physical layoutof the unit, and competency of the nursing staff.The scheme of total nursing care delivery also in-eludes forces affecting nursing unit production fromthe environments external and internal to thehospital including the many systems other thannursing (see figure 1).

The authors of the various pieces of literaturevary in their style of writing and quality of report-ing. The sophistication of analysis in differentstudies appears to be relates. to the backgroundof the author, the purpose of the writing, andthe nature of the problem. In the research reports,only a few writers state the assumptions underlyingthe selection of the statistics used in the analysisand present evidence that these assumptions aremet, In a feu studies, however, that could serve asmodels, the writing is clear, w al edited. and logi-cally organized. The use of language seems to re-flect the author's understanding of the audienceto which he is ,firecting his ideas and, in somecases, his understanding of the editorial policyof the journal in which the article appears. Whatis disappointing in much of the literature is the

STATE OF KNOWLEDGE: NURSE S l'AFFING METHODOt.OGY 9

Figure 1. Scheme of nursing care delivery system.

Community Environment

Expectation of servicesClientele:Knowledge of medical-hospital system

Health knowledge and utilization

Financial resources of communityPhysician supply and expectationNursing manpower supply

1Hospital Environment

organizationPhysician Eind4 uldual differences

attitudes and expectationsPatient utilization of servicesoccupancy rates and admission,

discharge patternsSpecialization effort: complexity, intensity, and vuriety

Floor layout: types and size of units

Budget and budget controlorganization

Hospital administration management styleleadership

Assignment of supporting services: accessibility, condition,

adequacy, and individuality of itemsHospital policies, cat* programsOrganization tension and authority structureIntent and purpose, philosophy, objectives

Nursing Administration

Status of nursing and nurses in the total structureLeadership and management stylePerception of purpose: philosophy, objectives

Policies and programs: development and executionAuthority structure--decentralization, decision making, delegation

CommunicationCoordinationBudget and budget control

Nursing Unit

Nursing supervision: clinical knowledge and skill managementand training

Amount, type, and kind of nursing staff and their combination

in workEmployment status of staffAbsenteeism, turnover, propensity of staff to leave, illness

Personnel policies: salary

Inservice and staff development programsAllocation and assignment of staffEducation and experience of staffRole clarity, role definition, and expectations of staffKnowledge and skill of staffAffective states of staff: satisfaction, morale, overload,

tension, sense of achievement

20

10

lack of precise wording. the ambiguity of the state-ments, and the absence of a logical developmentof ideas. This criticism, however, should be some-what softened in view of the fact that some ofthe literature was not written for publication. Anumber of pieces were clearly designed merelyas exhibits to be used in an oral presentation.

Study of the literature leaves the reader with theimpression that the amount written on the subjectis vast, highly uneven in quality, and availablethrough many outlets. The quality of the report-ing itself often gives grounds for concern. Thevalue of what is written is often questionable,since the writing lacks sufficient detail to provideeffective support for the argument.

This literature review makes possible severalstatements about what we do and do not knowabout nurse staffing. It points up the gaps in ourknowledge. about many variables generally regarded

RESEARCH ON NURSE STAFFING IN HOSPITALS

as related to staffing. It also indicates that we musturgently. proceed with studies of this importantsubject. It i., aim, highly desirable that the researchresults be cumulative. To hell) toward this goal,individuals studying staffing should attack similarproblems, ask similar questions, and keep intouch with one another. It is imperative that theinvestigations be bases, upon carefully_ designedtheoretical structures.

The generalizations I am going to make areorganized around the elements of a staffing pro-gram: measurement of what nurses do and whatpatients require; measurement of quality of pro-duct produced; factors which influence este workof nurses and, consequently, numbers and kindof staff that must be predicted; and methodologiescurrently used in hospitals to determine staffingneeds.

Measurement of Nursing Activity

In much of the literature reviewed, nursingwork basic to staffing predictions was measuredby the application of four standard techniques ofwork measurement developed by the engineeringprofession:

time study and task frequency count,work sampling of nurse activity,

continuous observation of nurse performingactivities,self-report of nurse activity.

All four techniques involve the estimate of timerequired for performance. Differences in techniquesreside in the methods by which data are collected,how the data are categorized, how amounts of timeare estimated, and how minutely nursing work isdescribed.

There is no question about the feasibility ofthese techniques. Furthermore, they are soundand well tested and can produce reliable dataif properly applied. Nevertheless, there are twomajor objections to the use of these techniquesin the measurement of nursing. The first relatesto the logic underlying their use in the study ofnursing practice. The second is the degree of pre-cision with which the needed data can be or havebeen collected.

First, the conceptualization of nursing practicederived from these work measurement techniquesis limited both in scope and character. Their useimplies that nursing work is distinctly proceduralin character: task oriented with specific beginningsand endings. Such a measurement does not reflectthe sweep of the effort. Nor does it properly conveya sense of the great complexity of knowledge andskill necessary in nursing practice. The applicationof these techniques assumes as an optimal goal thetype of nursing that even through ordinary practicesmay involve considerable interruption and dis-continuity.

Alihough the nursing units selected by the in-vestigator may be those whose staffs are regardedas providing the best present nursing care in ahospital, one can raise questions as to whetherthe nursing practice observed is the most effectivethat can be provided. In none of the literaturereviewed were there models of the optimum nurs-ing practice that could be achieved tinder the con-straints of the situation. Descriptions of nursingactivities obtained by these techniques tend to re-flect mainly the status quo of nursing practice,including the ritualistic and unnecessary, ineffec-tive acts. Omissions of care are not identified.

To design a model of nursing meeting desired

STATE OF KNOWLEDGE: NURSE STAFFING METHODOLOGY 11

yet feasible criteria of quality is the major problemin nurse staffing research. Unfortunately, nursingpractitioners and nursing admintrators have notgiven attention to this basic problem and are oftenunable to identify realistic and feasible criteriaof patient ewe. Furthermore, since nurses are un-able to articulate their concerns about the lackof models, they are unable to give guidance to in-dividuals sincerely interested in helping with staff-ing methodology. In none of the literature re-

viewed was the seriousness of the lack of modelsidentified.

Second, in many reports using these techniques,evidence is markedly lacking that the data pre-

seined are objective, reliable, and accurate. Onlya few of the studies describe the pretest of formsand data collection worksheets. Although severalmention orientation and training of nurse stalland observers for data collection, few have madetests of the agreement of observers or the reliabilityof raters on specific features. The problems oferror in recording or reporting data are seriouslyignored in most studies. Problems of the accuracyof reporting are seldom discussed. Often, samplesize and sampling procedures are not given. Con-sequently, the data presented in nursing activitymeasurement are open to the suspicion of beingbiased and incomplete.

Patient Classification Schemes

Forty-one pieces of the literature referred to ascheme to be used in classifying patients. Conceptually speaking, nursing practice and its de-livery evolve from the patient population that thestaff serves. The type of nursing practice, theamount, and the time and sequence of its deliveryare derived from the requirements of patients. Theintroduction of the categorization of patients is

an attempt to quantify the nursing care workloadcreated by patient care demands.

The number of classes in the patient classifica-tion schemes described ranges from three, the mostusual, to nine. The most frequently used categoriesare selfcare, partial care, and complete care. Somestudies refer to these groups as categories I, II, and111, indicating that I I 1 requires the most atten-tion. The names given by still other authors andinvestigators are related to acuteness of illness andcare requirements, such as minimal, partial, moder-ate, and intensive. The expression "variations inphysical dependency" is applicable to terms used inone study (1) as the title of a set of classes.

In the literature, there seems to be general agree-ment about the variables entering into the patientclassification systems:

(a) capabilities of the patient to care for himself,(h) special characteristics of the patient, related

to sensory deprivation,(c) acuteness of illness,(d) requirements for specific nursing activities,(e) skill level of personnel required in his care,

(f) patient's geographic placement or status inthe hospital system.

The capabilities of the patient to meet his ownneeds are defined by statements that attest to hisability to ambulate, to feed lihnself, to bathe him-self, and to care for his own elimination.

Not all variables are used in all classificationschemes. However, the schemes are usually de-veloped along a continuum of requirements fornursing assistance, moving from little to great need.Items picturing the behavior of the patient andhis special care requirements are ordered alongthe continuum. At specific points, the patient is

judged to he in a different category. Most authorsor investigators describe a scheme and providea form and instructions for its use in the classifica-

tion of patients.The .analysis of the literature describing charact-

eristics of patient classification leaves four impres-sions. First, the schemes are almost exclusively de-veloped along the physiological dimensions of care.Few items in the scales refer to the patient'ssociopsychological behavior or requirements. Sincemany of the schemes grow out of acute care set-tings, the emphasis on these characteristics is notsurprising. Second, little is reported in the litera-ture about the precision with which patients canbe classified by the methods proposed. There is ageneral failure of the authors to assess the degreeof reliability obtained in the classifications of pa-tients. The problems of selecting observers and

12 REtEARCH ON NURSE STAFFING IN HOSPITALS

training them to classify patients are almost totallyignored. Third, only one study reported an at-tempt to test for validity (2). Fourth, patient

classification schemes, despite their limitations andimperfections, are widely used in staffing metho-dologies.

Quality Control

In the-182 pieces of literature reviewed there wasno adequate, operationally defined, model ofnursing itself. Furthermore there were no descrip-tions- of the program of evaluating nursing careused by the departments of nursing service in thehospitals where the studies were being conducted.

Yet the most critical staffing problem element isdefinition and identification of the quality of theproduct (nursing care) and its effect upon thepatient. Often. staffing is viewed as an end in it-self, rather than the means to an end. Measuresto determine the effect of changes in staffing uponthe quality of nursing production are essential instaffing programs.

A most complicated and confounding are?. forstudy is assessment of quality. Though much at-tention has been given to this question, never hasa completely acceptable model been devised. Re-search directed toward tools for assessment of nurs-ing care is vital. Without valid and reliable tools,progress in measurement of care requirements, staff-ing programs, and staff utilization, lags.

The authors varied in the number of measuresused to assess quality. The most comprehensiveworks, involving a large number of variables relat-ing to quality measurement or the criterion prob-lem, are those of Abdellah and Levine (3), Ayde-lotte and Tener (4), the Illirois Study (5),Ingmire and Taylor (6), Jaco (7), Lewis (8), Jelinek,et al (9), Alfano and Levine (10), and Wolfe (11).These studies, in particular Wolfe's, illustrate thedifficulty in measurement and the complexity of theproblem of quality determination.

Reports of hospitals using the CASH (Com-mission for Administrative Services in Hospitals)program suggest that something is not right withpresent methods of evaluation of quality, but thisdoes not necessarily entail the conclusion that thebasic concept of continuing appraisal is wrong.

Instead, it leads one to speculate that perhaps themode! of quality is poorly conceptualized, thevariables selected are invalid, and the instrumentused is lacking in sensitivity, reliability, and ob-jectivity.

These ideas are in part derived from the progressreport of the research study by Wolfe and Breslin(12) in their survey of eight nursing units. Fourranked high and four low, based on the rankingsin terms of the quality of nursing care made by34 head nurses, supervisors, and instructors in amedical center. The rankings of these nurses as toplacement of the units on the high-low continuumof quality were highly consistent. These investiga-tors developed a tool incorporating some of theitems of CASH quality control with other itemsand tested the nurses' rankings. Although the find-ings of Wolfe md Breslin are inconclusive, thereport does raise questions about the ability of theitems drawn from the CASH material to discrimi-late between nursing units providing low qualityof care and those providing high quality.

The materials describing quality measurementwere foi the most part disappointing. Evidencesupporting validity and reliability of tools wasmissing in tnost studies. Rationales for the selec-tion of specific items for inclusion in a measuringdevice were not presented. The question of thesignificance of the content of an item, in termsof changes in the state of a patient and his re-covery, was enerally ignored. It seems clear that,in spite of t few heroic attempts at research onthis question, relatively little is known aboutquality. Continued attention must be given to thematter of describing and resolving problems ofquality measurement. Of prime importance is theidentification of what is to be delivered as thenursing product in the first place.

23

STATE OF KNOWLEDGE: NURSE STAFFING METHODOLOGY

Related

Some studies dealt with other variables. Thesedid not deal with patient classification (nursingworkload), measurement of nursing activity, orquality measurement. Rather they dealt with vari-ables related only indirectly to staffing and reflect-ing, instead. probable use of time, effectiveness ofservice, or organization of effect.

These related variables have been grouped intofive categories:

(a). architectual variables, such as design of units,implying travel distance, room placement,visibility of patients, and communication sys-tems:

(b) organizational variables, including charac-teristics of the hospital, the introduction ofnew roles, redesign of supporting systems, andrealignment of authority and responsibilityin nursing;

(c) variables relating to nursing staff background,preparation in service programs, turnover,and internal states such as satisfaction andmorale;

(d) hours of nursing care;(e) miscellaneous variables.Many writings discussed a combination of vari-

ables. Each work was read to learn whether itincluded specific findings regarding the relation-ship between these variables and nurse staffingprediction.

Organizational Changes

Twelve pieces of literature were specifically con-cerned with the introduction of the service unitmanager system or other changes in supportingsystems. The review of these writings leaves thereader with the impression that there is limitedknowledge about the impact of these highly im-portant variables.

One cannot, with confidence, say that an increasein supporting services will affect the amount ofstaff required; one can probably state that theamount of direct care to patients and standbytime of nursing personnel will be increased. Itcan also be said that administrative variables pro-bably have an effect on utilization and amount ofstaff required. Unfortunately, there is still no ade-quate basis for stating that under one style of

13

Variables

administration less staff will be required than underanother.

Architectural Variables

The literature search for writings concerned witharchitectual variables was not ehhaustive. Six majorreports were found that examined the effect ofdesign, travel time required of nursing staff, orboth.

The results of these si. studies, four of whichare major ones in the field (13,14,15,16), indicatethat nurse travel time does differ in types of nurs-ing units. Three studies indicate that in the cir-cular or radial units nurses spent more time indirect care. The general impression, although onecannot generalize because of the different swingsand the limitations of the studies, is that circularunits are favored by staff over the other two typesof units.

No study addresses itself to the question of c'if-ferences in staffing requirements. Yet a distinctimpression emerges that types of units differgreatly in travel distanCe and time. Whether or notthe distance and time required are sufficient torequire more staff in single corridor or rectangu-lar units is not known.

Nursing Staff Characteristics

A number of different variables relating to thenursing staff were found in the studies reviewed.One group of studies was concerned with nursingstaff satisfaction, turnover, and propensity of nursesto leave their positions. Nursing staff perceptionsof adequacy of services were considered in a num-ber of studies. These studies are few and, eventhough they include two fine research reports (3,17), are too limited to permit generalizations aboutthe relationships of nurse satisfaction to other vari-ables. They point to relevant questions for re-

search.

lnservice Educational Programs

The impact of inservice nursing education uponstaffing requirements was not examined in most of

24

14 RESEARCH ON NURSE. STAFFING IN HOSPITALS

tits studies. The rile( I of inservice educational pro-grams cart stall performance, either on the abilityto achieve gveater Workload or on the productionof higher quality G.re to patients, is still obscure.There is some evidence the programs eflectielyincreased patient contacts. However, the question,Does ail inserice educational program result ingreat e:liciency and economy of staff? is titian-met ed.

Nursing Hours per Patient Day

Only a limited number of research vudics pro-vide information on actual nursing hours per pa-tient day (N FWD) and the validity and reliabilityof the data used in those studies are in most in-stances questionable. Techniques are loosely de-scibed, samples arc small or ill der.ned, instrumenttoting is deficient, and rationales are weak. In mostinstances, information about the quality of care isderived solely from the judgments of the nurseinvolved. The studies vary greatly in scope andquality. Most apply only to specific hospitals. Ade-quacy of staff utilization was found to vary greatlyamong hospitals, and the hospitals varied greatlyin the average number of daily hours nurses wereavailable.

The confounding question of patient care qual-ity arises in the examination of these regtarchstudies. This kind of assessment was ignored, takenfor granted, or poorly handled. It is obvious thatNHPD is less titan satisfactory in staffing determinetion. The literature suggests that the use of guidesand procedures applied to a specific hospital situa-tion is a more logical approach.

Age and Care Requirements

Three studies examined the effect of age uponnursing care hour requirements. These three sup-port each other's findings that the over 65 yearsof age patient group exceeds the under 65 yearsof age group in hour requirements. Jacobs' study,(18) the most comiirehensive of the three, reportedthat the 65-74 year old group requires 21-31minutes per day more and the 75 year and oldergroup needs 55-79 minutes per day more. Hefound there was little difference between the timerequirements of adult medical and adult surgicalpatients, and that specialization efforts of thehospitals were not factors in the amount of carerequired.

Control of Costs

Most of the methodological studies of staffingwere highly concerned about the control of costs.However, none of the studies reviewed was suc-cessful in showing the relationship between thevariables of cost effectiveness and nursing staffing.The problems of defining cost, obtaining reliabledata, and isolating the dependent variable, quality,were recognized by .nany other authors. Twostudies presented highly interesting and relevantmaterial about the shortage of nurses, using dif-ferent economic models for analysis (19, 20). Bothauthors provided provocative pertinent data aboutthe dilemma confronting employers and especiallythe public.

Staffing Methodologies

Staffing methodologies in the literature fell intofour major classes:

(a) descriptive, using a numoer of variables, sur-vey methods, and the subjective judgments 'ofindividuals;

(b) industrial engineering;(c) management engineering (2'1);

(d) operations reseat ch.

These methodologies also represent a gradual in-crease in the application of logic to the problem

25

and the use of abstraction its resolution (seefigure 2).

To clarify the following disculsion, the terms"methodology" and "approach" and these fourclasses of methodologies will be defined. The term"methodology" is used to describe. a "logically andprocedurally organ:zed arrpagement of steps. Whenformally documented utilized, these steps mayconstitute a science of method for whatever dis-cipline uses them (22).


Figure 2.Diagram of evolvement of methodologies for the study of nursestaffing; application of logic and increases in abstractions.

Descriptive IndustrialEngineering

The first term, "methodology," is regarded as amuch larger concept than "approach." The lattlzterm is more limited in that it represents anorientation or a preliminary step and does not en-compass a full methodology..

The descriptive methodology makes use of anumber of data gathering devices about a largenumber of variables, she relationships of whichare uot clear. The final decision abo _4t amount ofstaff resides in the judgments of Individuals whohave a background of experience. The approachmay use simple ratios, formulas, and suggested pro-portions between types of personnel. However,there are no consistent strategies in the methodo-logy.

The industrial engineering methodology is di-rected at the study of nursing work on a nursingunit. It uses the techniques of work measurement,work distribution, task or function analysis, pro-cedure analysis, and the like. It is primarily directedat reorganization, reassignment, and redistribu-tion of work on specific nursing units.

The management engineering methodology,which claims to be based upon common sense,uses tools and techniques from industrial engineer-ing and from systems analysis, and builds uponfindings from operations research studies. Thismethodology must give evidence that the majorityof the component strategies are present, althoughcompleteness of each component may vary. Thecomponent strategies include:

a statement of performance objectives,an analysis of components and functions,distribution of functions,scheduling,training individuals for the use and testing ofthe system,

installation of system,quality control (23)

ManagementEngineering

OperationsResearch

15

The operations research methodology is directedtnward enabling one to make decisions for reallife situations. Mathematical models are built torepresent real life problems. The essential ele-ments are abstracted from the situation so that asearch can be made for a relevant solution to theproblem. The structures of solutions are exploredand procedures are established for obtaining them.The optimal solution and set of procedures arethen made available.

Since authors of the pieces reviewed often uti-lized the same techniques, such as counting, patientclassification, work measurement, and includedmany of the same variables, a set of arbitrary ruleswas drawn up for classifying methodologies. Therules established exclusive categories. The descrip-tive methodology is one in which there is noquantification of nursing work and no use of patientclassification system. The methodology drawn fromoperations research requires the construction ofa mathematical model.

The industrial management methodology em-ploys methods from industrial engineering andrestricts it to a study of the nursing unit. It doesnot include the full range of strategies seen inthe management engineering methodology and doesnot interface the work on the unit with systems ofthe hospital. It does not include p? ',nit classifica-tion. The management engineering methodologymust give evidence that the majority of componentstrategies are present, although completeness of eachcomponent may vary. It may use r thematical toolsand may draw upon the research conducted by per-sons using operations research methodology.

These four types of methodology for the deter-mination of nursing staff required in a hospitalare described in many of the research reports.manuals, pamphlet, and guides reviewed in theliterature. Linkar between the work of individ-ual arc discernis le, although the times at which

26

16 RESEARCH ON NURSE STAFFING IN HOSPITALS

Table 1.Comparison of staffing methodologies on selected features

Title of methodology

Feasibility of approach

Reliability of data Validity of dataReproducibility

of forms Setting requirements

Descriptive

Industrial engineering

Managementengineering

Operations research

Descriptive


No evidence

Depends on accuracyof reporting, trainingprograms, and clarityof forms. Checks pro.vided in some Leptis

Depends on accuracyof reportingroutinechecks are not built in,training of personnelfor data collection, andnecessary clarity offorms.

Evidence presented.Depends on accuracyof reporting.

Subjective judgmentabout value ofvariables

Studies and reportsfrom industry. As.samption made thatnursing is a compositeof tasks or activities

Face validity

Face validity

Face validity; relies onConnor's scheme andits modification.

Face validity; relies onConnor's classification.

Survey methods:counting, question.naires


27

Some excellentexamples in literature(C7, Gunn) (C14,Paetznick) (D29, Ludwig and Humphrey)

Excellent examples inliterature (A47, Mergen) (D4, Bartscht)(D20, Hansen)

Examples in literature(CI9, Scottish Health

Service, No. 9) (C13,

Operations ResearchUnit) (C4, CASH,Staff Utilization)(Cl2, Mass. HospitalAssociation)

Examples in literature(A72, Wolfe) (D46,

Wolfe, and Young,Part II)

Many. Examples:census data, hospitalstatistics, financial resources, training,nursing objectives,policies, clinical services, etc.

Personnel statistics,nursing care activity,and time require.ments, work layout,census statistics, workdistribution, and work.load factors.

No specific require.ments

Industrial engineeringconsultants or nursingstaff knowledgeableabout concepts andtools

Consultant help; nospecific hospital setting

Computer; consultanthelp

Broadly applicable,but relies upon experiences and knowledge of the director,the administrator, andthe staff

Various types of workmeasu:ement used.Nursing staff can makeapplication.

STATE OF KNOWLEDGE: NURSE STAFFING METHODOLOGY 17

Table 1Continued

Title of methodology Basis of rationale

Managementengineering

General approach used Variables considered Comments

Operations research

Research studies of theJohns Hopkins Opera-tions Group. Studiesand reports from busi-ness and industryusing industrial en.gineering and manage.ment methodology.Assumptions made:that nursing providedrepresents quality de.sired, that patientclassification is valid.that nursing sell icedemands are stochas-tically distributed, andthat standard times ofprocedures are validand reliable.

Studies and reportsfrom industry. Opera.dons research by Con.nor, Wolfe. and Flag le.Same assumption asabove.

Industrial engineeringmethods and systemsanalysis

Operations research:mathematical allocation and situation!maid, QueueingTheory. Industrialengineering and systerns analysis

(a) nursing workloadby task frequencyand levelstandard time fortasks or categoriesof patientsnumber of patientsin classes

(d) supporting services(e) I. innel levels

skill(f) administration

,licy and procedures

(g) attributes ofquality

(h) personnel dataand statistics

(i) age of patient(j) cost figures

(h)

(I) classes of patients(2) 16 task complexes(3) cost values

Programs are pur-portedly designed withthese components:(a) identification of

performance ob.jectives

(b) function analysis(c) scheduling of per.

sound(d) prediction of staff

allocation tables(e) training of per.

sonnet(f) quality control

program(g) management con.

trol informationreports

Application not widelyobserved in the literatare. Quality measure.ment recognized as notpresent.

they occurred are not known. Many of the writingsare undated and the investigator did not attempt,by personal contact with individuals, to learnprecisely when programs were initiated.

The majority of the manuals, pamphlets, andguides are poorly written. Little is included in themabout rationale or assumptions underlying themethodology being described. Protocol and pro-cedures are likewise sketchily outlined. The use ofthe manuals would not be possible without con-sultant help. Some reports of application of themethods were obtained that clarified the metho-dology consilerably.

A comp anon of the four. methodologies wasmade with respect to a Tecific set of character-iitics that included rationale, approach, variables,and feasibility (refer to table 1). In all four me-thodologies, subjective judgments are present,since nursing units selected for workload measure-ments are made either by individuals or by groupdiscussions. Measurements of care quality and per-formance standards are weak or totally lacking in

all four methodologies. Relatively few variables areconsidered. yielding narrow basic conceptual mod-els, and the impact of architectural variables, or-ganizational variables, and specialization efforts isalmost totally ignored. No attempt was made topredict the number of persons required for majorposition categories in the nursing care deliverysystem. Validity and reliability of the data arequestionable.

Except for a few reports applying the metho-dologies, tests for accuracy in reporting the dataand the training of observers are not proposed.Validity of the data resides in the judgment ofnurses selected to work with the methodologists.All four methodologies are applicable to a varietyof health settings. Two types of methodologiesrequire consultants in their use. Forms employedin recording tasks are incomplete, and standardtimes reported for task performance are not docu-mented. The classification schemes used in twoof the methodologies are drawn from Connor's(24) basic research.


Summary of Literature Search

The results of this literature search have beensobering. We can say that we know little. However,there are a few generalizations we can make ina summary. First, nursing activity can be measuredby standard methods drawn from the discipline ofengineering. The techniques are feasible and, ifappropriately applied, provide reliable data aboutwork that reflects nursing as it occurs in currentpractice. However, the tasks, activities, and cate-gories appearing in the literature do not, in myopinion, reflect the nature and full character ofnursing practice, but only a portion.

Second, most patient classification schemes aremodifications of that proposed by Connor who inhis dissertation provided an admirable rationale ofclassification. However, most of these classificationschemes have not been tested for reliability of thedata they produce. The evidence of validity pre-sented is weak. Although the variables used forclassifying patients are fairly universal, the numberand kind are limited and some requirements ofcare are either overlooked or ignored. Further,questions can be raised about the significance ofsome of the items included in the scales. Do theyreflect states of recovery and the objectives forwhich the patient is in the health care deliverysystem? The relationship between patient classi-fication schemes and nursing workload has beenbest handled in the research by Connor (24) andWolfe (//), especially the latter. however, theknowledge we have about the combinations of

nursing personnel and their ability to mount aworkload arising from a particular mix of cate-gories of patients is little.

Third, a large number of variables appears toinfluence the work of a nursing staff. Again, weknow little about the effect of the interaction ofthese. The impression one derives from the litera-ture is that the admission and discharge process ofpatients, the environmental structure of the unit,the supporting systems, the capabilities of the nurs-ing staff, and the clinical nursing leadership aremajor factors to be considered. Few pieces of theliterature reviewed referred to the impact of thephilosophy and practice of medical staff, their num-ber, their presence, or absence. Yet I predict thatthe physician and his behavior will prove to be aprime variable in the creation of nursing work.

Fourth, four types of staffing methodologies wereidentified. These all are feasible and are widelyused. The concept of variable staffing is adoptedin many institutions.

Fifth, quality measurement is poorly handled innearly all the studies. In large part this is due tothe failure to describe in operational terms theproduct to be produced.

Sixth, there exist in the literature no adequatedescriptions of models of the nursing care deliverysystem and nursing practice. The majority of stud-ies give little theoretical structure as a basis forinvestigation.

Problems and Issues

This review of the literature leads to the con-clusion that our knowledge about nurse staffingis very limited. The number of variables withwhich one mull deal is large. Furthermore, manyof them are almost incomprehensible and, for themost part, appear' to defy description and quanti-fication. There is little agreement in the literatureon the nature and characteristics of some obviouslyimportant variables: nursing practice, patientneeds, quality of care .control, complexity of care,intensity of care, and levels of practice.

Review and assessment of the literature notonly point up how little we know of the subject

49

but also force our attention on the central prob-lems and issues relating the nurse staffing metho-dology. Nowhere in the literature was found awell developed model of the nursing care deliverysystem for use in the development and examina-tion of staffing. The term "model" is used, in thissense, to refer to an analogy of the pattern ofrelationships observed in the nursing care deliverysystem.

The problems of model construction are exceed-ingly troublesome. First, there is the question ofsetting the boundaries of the model. Will the bound-aries be those of the nursing unit, as they are in


Jelinek's (25) and Wolfe's (II) writings, or willthey include parameters beyond the hospital, asimplied in the Illinois Commission Study (26,27)and Levine's (28) study of Federal and non-Fed-eral hospitals? Second, how many variables can beadequately explained and managed in the model?The problem of identifying relevant variables andeliminating irrelevant ones is extremely complex.Knowledge about the variables and their relation-ships is vague and measurement of some of thevariables is almost impossible.

Nurses who have been within the real workworld would undoubtedly identify three groupingsof variables:

(a) selected factors external to the hospital in-fluencing nursing delivery, such as nurse-power pool, financial resources of the clien-tele, and the health status of the personnel;

(b) components of the hospital system, such aspatients, medical staff, management, financialresources, supporting services, traffic patterns,nursing leadership, qualification of staff, staffproductivity, and staff stability;

(c) independent factors that can be used to de-termine the effectiveness of the system, suchas patient satisfaction, patient behavioralchanges, and efficiency of services.

Some good ideas are found in some writings. How-ever, thoughtfully conceived and carefully writtendescriptive models are greatly needed.

The second major problem confronting staffingmethodology development is the difficulty of iden-tifying in advance what the patient or group ofpatients is entitled to receive as the product,nursing rare, for which they are paying. The fun-damental reason for assembling a staff is to takecare of the patient.

Identifying the components of care a patientmust receive is essential to any staffing methodology.The arrival and discharge times, the hourly anddaily variation in the patient's requirements, themedical programs and scope of medical practice,the training programs, the problems of measure-ment of quality, and the lack of an operationaldefinition of care itself make the problem of meas-urement formidable. None of these items were ade-quately treated in the literature, and some weretotally ignored.

The assumption made by the present methodol-ogists is that, at the time of the survey, what is

30

19

provided for the patient is sufficient. Therefore,the present classification systems have examinedphysical elements of care or activities performedand have accepted them as requirements for care,without giving adequate thought to the question,What does the patient truly require? There is noempirical evidence to support the assumption thatwhat the patient is getting is all that he requiresor is necessary for his recovery. In fact, there issome evidence that what he is getting is given asa matter of ritual.

The majority of patient classification schemesreflect specific nursing tasks, a number of whicharise because of medical order and acuteness ofillness. The schemes do not reflect emotional needs,orientation of the patient, instruction needs, andcomfort, other than through the process of provid-ing physical care. Few have even attempted to in-clude these items. No scheme, for example, hasbeen built upon the nursing problems of thepatient, although several investigators began withthis intent. Yet, experienced practitioners in nursing can set a priority ranking of nursing prob-lems and combinations of problems that wouldreflect both amount of time and professional knowl-edge base and skill required for performing thecare to solve the problem. Only one study reportedthe reordering of priorities of nursing care as aresult of the admission of a critical patient.

If one assumed present patient classificationschemes valid, another problem exists. Except forone study, patient classification schemes have notbeen rigidly tested and have not been used in avariety of different settings in a way that wouldindicate how reliable they are. Connor's schemewas designed for a large medical center and onlyone formal test of it was found in the literature.Is it equally applicable to medical-surgical patientsin a small community hospital? What is the re-liability between raters?

?roperly used, the various methods for study-ing nursing work are feasible, reliable, and simple.The problems are those of sampling, of selectingmodel units for survey, of establishing categoriesthat represent major nursing functions, and of thelimitations inherent in the use of the method.What is most questionable is the assumption thatwhat is being observed is desirable. Further, thecategories used often do not reflect the planningand thinking time required if nursing is perceived

20

as having an intellectual base. The selection of amodel unit for study to make staffing predictionsis critical. The management engineering reportsindicate that the selection of the unit rests uponthe judgment made by an advisory group composedof nurses and hospital administrators. It appears,however, that this judgment is usually not basedon hard dam, reflecting the achievement of a spe-cific quality of care, evaluated by well defined cri-teria.

Another problem in nursing measurement is

grouping tasks or categories into complexes forfurther research on staffing methodology. It is ex-tremely difficult for sonic nurses to perceive thatpersonal services to patients may not be necessary.feasible, or economical. The use of expensive pro-fessional time for simple tasks, repetitive butpersonal for the patient, is overuse of skill andtalent. Ilia this statement is unacceptable to manynurses.

These nurses state that use of auxiliaries de-personaliies nursing care, that these personal tasksmay serve as vehicles for nursing action, and thatthe elimination of personal care from registerednurses' repertoire of duties is unwise. The cruxof the argument is not whether it is personal ornot, but what is the knowledge and skill requite-ment for its performance and how does this affectpatient recovery? It is uneconomical for registerednurses to spend time on giving nontherapeuticcare to patients. We should ask what registerednurses are doing for patients that no one else canprovide and what they are not providing.

The ultimate test of a staffing program, or anynursing activity, is the progress of the patient, notthe quality of performance of the nursing staff,the appearance of nursing units, or the complete-ness of records.'- The amount of time spent with

' Acknowledgment is given Dr. Marjorie Moore, AssociateDirector. the Universit of Iowa Hospitals and Clinics. forthe following paragraphs.


a patient, the skill with which a particular pro-cedure is performed, the number of times a patientreceives a procedure, the number of times thepatient has contact with the nursing stall, theflow of communications, the results of a nursingaudit, or even patient satisfaction are inadequatecriteria for the evaluation of the quality of nursingcare the patient has received until the relationshipbetween these factors and the progress of condi-.tion of the patient has been established. It hasgenerally been assumed that a positive relation-ship exists between certain accepted nursing prac-tices and the patient's condition. The evidence tosupport most of these assumptions is meager,particularly applied to a total nursing unitrather than a select group of patients.

The major reason the relationship betweennursing activities and patient condition has notteen generally determined may be the problemsinvolved in the development of patient conditioncriteria; i.e., identification of what one must ob-serve in, on, or around the patient to determinewhether the nursing care hr receiving is bene-ficial to him for his present and future status.It is time the question of the construction ofcriteria be squarely faced.

The measurement of nursing care quality andthe establishment of performance standards can-not be accomplished until nursing practitionersand leaders in nursing practice give attention tothe development of such criteria. The primacy ofthe problem is apparent. The questions to be askedare these: Where are the practitioners who candevelop stated criteria? How can the criteria bebuilt into an ongoing program of quality measure-ment? What modifications are required of the on-going program to build scales or indices whichcan be used at specific points in time for researchpurposes?

Future Research Direction

Although. the problems and issues reviewed arealmost overwhelming, I inn ccmfident progress canbe made on their tesolution. First of all, nursesknowledgeable of nursing practice and intimately

acquainted with the real work world of nursingcare delivery systems should be sought, encouraged,and supported in the development of conceptualmodels of the nursing care delivery system. This is


no easy task. Furtliermore, these individualsshould be encouraged to seek out and collaboratewith individuals from other disciplines.

Second, carefully designed research studies test-ing present classification schemes should be con-ducted. Since limited evidence appears in the lit-erature that these schemes are providing reliabledata carefully controlled tests should be madeof the schemes in a variety of hospitals. The ex-ternal factors that influence the character of thepatient population should be identified as fully aspossible.

Third, patient classification schemes built aroundnurse perceptions of care should be built and tested.These schemes should be based on reasonable carerequirements as nurses see them. The presentschemes do not reflect omissions of care that maybe serious; on the other hand, omissions may notbe important. However, the recommendation is

that proposed schemes for the classification of pa-tients indicating nursing loads be developed byknowledgeable practitioner nurses. These schemesshould be established around nursing problemsand should reflect nursing priorities, knowledge,and skill level. The scheme should lend itself tocomputerization. It should be based on data usedby the nurses in assessing patient tare requirements.

Foul th, since the effect of improved supportivesystems upon nursing staff requirements has nt.tbeen well documented, controlled studies of theeffect of improved delivery and retrieval systemsand placement of patient care items, equipment,and supplies should be conducted. Holdingqual:ty constant, what is the effect of improvedavailability of work materials and tools for pa-tient care upon the size and kind of nursing staffrequired? What is the effect upon cost?

Fifth, a concentrated attack on measurement ofquality must be made. From the review, the im-pression is that atiempts are too global. Exceptfor a few instruments, hot much is available toassist the director of nursing and the hospitaladministrator in making estimates of quality. Theevaluation must relate to the effect on the patient.

21

Therefore, different sets of criteria may be neededsince the patients in the hospital vary. The patientvariation is dependent upon t" purpose for hisbeing in the hospital, the pi gress he is making,and the course of his recovery. Development ofthese criteria will require the involvement of ex-pert nurse clinicians.

Sixth, guidelines for use of the four specificmethodologies should be developed. These guide-lines should state the rationale upon which eachis based, the variables studied, the limitations ofeach the appropriate use of the methodology, thecomponent strategies, and the procedures. It shouldinclude examples of well developed forms, reports,the guides for use of allocation tables, if included,and the like. The value of the methodology aswell as its limitations should be stated. The im-portance of concurrently studying supporting sys-tems, or even preceding the analysis of nursingwith a study of these, should be emphasized. Properutilization of supportive services and expecta-tions of adequate service from them should bemade clear.

Seventh, guidelines for new methodologiesshould be explored, greatly enlarging the variablesconsidered. The methodology should be based up-on a carefully described conceptual model of nurs-ing care delivery. It should provide guidelines thatwill take into account providing for omissions ofcare as well as eliminating repetitions of care, theorganization of the work of nursir personnel,and the sequence of its delivery.

These seven recommendations are specific pro-posals for research. My final comment is a generalone but one which I am convinced is of majorimportance. Research on these seven must be mul-tidisciplinarythe discipline of nursing must beinvolved. Mechanisms for sharing research find-

ings must be set up and knowledge systematicallybuilt. To achieve this end, we must address atten-tion to the structuring of theoretical frameworksfor the research on staffing of nurse delivery sys-tems.

,

22

References

(1) Scottish Health Service Studies, No. 9. Nurs-ing Workload per Patient as a Basis forStaffing. Scottish Home and Health Depart-ment, 1969.

(2) Mainguy, J., et al. The Reliability andValidity Testing of a Subjective Clussifica-tion System. Report of Canadian PublicResearch Grant No. 609-7-197, VancouverGeneral Hospital, Vancouver, British Co-lumbia, June 1970.

(3) Abdellah, Faye G., and Levine, Eugene.Effect of Nurse Staffiing on Satisfactionswith Nursing Care. Hospital MonographSeries No. 4. American Hospital Associa-tion, Chicago, Illinois, 1958.

(4) Aydelotte, Myrtle, and Tener, Marie. AnInvestigation of the Relation BetweenNursing Activity and Patient Welfare. FinalResearch Report (GN 4786 and N 8570),University of Iowa, Iowa City, Iowa, 1960.

(5) Illinois Study Commission on Nursing."Nurse Utilization in Thirty-one IllinoisHospitalsMethodological Report." Labo-ratory data presented to staff members ofthe Division of Nursing, National Institutesof Health, Washington, D.C., December1968.

(6) Ingmire. Alice E., and Taylor, Carl. TheEffectiveness of a Leadership Program inNursing. Western Interstate Commissionfor Higher Education, Boulder, Colorado,(NU 0078), 1967.

(7) Jaco, E. G. Evaluation of Nursing and Pa-tient Care in a Circular and RectangularHospital Unit. Final Report, prepared forthe Hill Family Foundation, St. Luke'sHospital, St. Paul, Minnesota, June 1967.

(8) Lewis, Charles. A Privileged Communica-tionThe Dynamics of Nursing in Ambula-twv Patient Care. Final Research Report,U.S. Public Health Service Grant (NU00145), (no date).

(9) Je linek. R. C.: Munson, Fred; and Smith,Robert. SUM (Semler Unit Management):An Organizational Approach to ImprovedPatient Care. A Study Report for the Kel-

33

RFSEARCH ON NURSE STAFFING IN HOSPITALS

logg Foundation, Battle Creek, Michigan,1971.

(10) Alfano, Genrose J., and Levine, HerbertS. Progress Report. Longterm Effects on anExperimental Nursing Process. (NU 00308-40). Montefiore Hospital and MedicalCenter, New York, New York, 1971.

(11) Wolfe, H. "A ;Multiple Assignment Modelfor Staffing Nursing Units." (GN-5537, HM-00279, and FR-0004), Unpublished DoctoralDissertation, Johns Hopkins University,Baltimore, Maryland, 1964.

(12) Wolfe, H.. and Breslin, Patricia. FactorsAffecting Quality of Nursing Care. ProgressReport, (NU 00192-0101), School of En-gineering, School of Nursing, University ofPittsburgh, Pittsburgh, Pennsylvania, 1968.

(13) Sturdavant, Madelyne. Comparisons of In-tensive Nursing Service and a RectangularUnit. Hospital Monograph Series No. 8,American Hospital Association, Chicago,Illinois, 1960.

(14) Trites, David K., et al. Final Report onResearch Project Entitled Influence of Nurs-ing Unit Design on the Activity and Sub-jective Feeling of Nursing Personnel. Re-search Office, Rochester Methodist lIospital,Rochester, Minnesota, (no date).

(15) Huseby, Robert. "Impact of Nursing Designon Patients' Opinion." Unpublished Master'sThesis, University of Minnesota, Minne-apolis, Minnesota, 1969.

(16) Lyons, Thomas F. Nursing Attitudes andTurnover: The Relation of Social-Psycholog-ical Variables to Turnover, Propensity toLeave, and Absenteeism Among HospitalStaff Nurses. Industrial Relations Center,Iowa State University, Ames, Iowa, 1965.

(17) Jacobs, Stanley. et al. American HospitalAssociation Nursing Activity Study ProjectReport. Public Health Service Contract No.PH 110-235, U.S. Department of Health,Education, and Welfare, Washington, D.C.(no date).

(18) Yett, Donald E. "The Chronic 'Shortage'of Nurses: A Public Policy Pile,ama."


Empirical Studies in Health Economics,Johns Hopkins University Pre.,s, Baltimore,Maryland, 1970,357-97.

(19) Bognanno, Mario Frank. "An EconomicStudy of the Hours of Labor Offered by theRegistered Nurse." Unpublished DoctoralThesis, University of Iowa, Iowa City, Iowa,June 1969.

(20) This category is used since it is a generalterm applied to the methodology describedin Dollar and Sense, a study report preparedfor the W. K. Kellogg Foundation by

Patric I. Ludwig, October 1971. Individualsarc referred to this monograph for informa-tion about the extent to which this metho-dology has been used.

(21) Optner, Stanford L. Systems Analysis forBusiness and Industrial Problem Solving.Englewood Cliffs, New Jersey: Prentice-HallInc.. 1965.

(22) Hi Frederick S., and Lieberman, Jeral 1.troduction to Operations Research. S n

Francisco: Holden-Day, Inc., 1967.

34

23

(23) Connor, R. J. "A Hospital Inpatient Classi-fication System." Unpublished I) IctoralDissertation, Industrial Engineering Depart-ment, Johns Hopkins University, Baltimore,Maryland. 1960.

(24) felinek, R. C. "Nursing: The Developmentof An Activity Model." Unpublished Doc-'oral Dissertation, University of Michigan,Ann Arbor, Michigan, 1964.

(25) Illinois Study Commission on Nursing. NurseUtilization: A Study in Thirty-one Hospi-tals. Vol. III. The Illinois League forNursing, Illinois Ntrses' As.ociation, Chi-cago, Illinois, 1966-70.

(26) Levine, Eugene. "A Comparative Analysisof the Administration of Nursing Servicesin a Federal and Non-Federal GeneralHospital." Unpublished Doctoral Disserta-tion, Nmerican University, Washington,D.C., 1960.

(27) Haussmann, R. K. "Waiting Time as anIndex of Quality of Nursing Care." HealthServices Research, vol. 5, no. 2 (Summer1970).

Factors Related to Nurse Staffingand Problems of Their Measurement

Charles D. Flagle, Dr. Eng.Professor of Operations Research

Department of Public Health AdministrationSchool of Hygiene and Public Health

The Johns Hopkins University

Introduction

Looked at in any way, the problem of staffinghospital nursing units is profoundly important andprofoundly difficult. In economic terms alone, weare dealing with several billion dollars in wageseach year in the United States, something on theorder of AO percent of all medical costs, and afourth of hospital costs.l Little wonder then thatnursing care is a focus of attention, not only inits formulation and distribution of tasks, but inthe criteria for admission and discharge to inpa-tient care. Throughout our discussion we shouldkeep in mind that the total cost of impatientnursing care to a community is the product ofthe cost of nurse hours per patient day timesthe total number of patient days. The total numberof patient days of care may be more subject tr, con-

M. Y. Pennell and David B. Hoover, in Health ManpowerSource Book Section 21, "Allied Health Manpower 1950-80,Public Health Service," (Publication No. 263, 1970) projectpresent total nursing and related sources from two millionemployeei at. present to three million by 1980, roughlyonefourth being registered. The Commission of HealthManpower, 1987. estimated 60 percent of the total areemployed in shortterm general hospitals.

25

35

trol in the health care delivery system than theresources used in a day of care.

In social terms, we have lived through a fore-shortened transition from a vocation expectingmuch personal self-sacrifice to one bearing con-siderable resemblance to the way of life and af-fluence of other skilled purveyors of human serv-ices. Changes in attitudes and images haveaccomim lied this transitionthe secularization, ifyou will, of patient care. The tangible correlatesof these changes have been such things as shortenedand more formalized working hours for nurses,higher wages, unionization, occasional strikes, andin the auxiliary positions an undeniable demandfor ethnic equality and upward mobility in jobs.

Interwoven with all the economic and socialchanges affecting the organization for patient care,there has been an unending series of technologicaldevelopmentsphysical technology and medical.Intensive care units and life islands, for example,have imposed demands on nurses for increasingspecialization.

In the face of all this change, heaped upon the

26

inherent difficulty of caring continuously and si-multaneously for 30 or so sick people in a nursingunit, it is little wonder that a great deal ofadministrative and research attention should becentered on the staffing problem. Determining whoshould be where, when, and doing what, is a prob-lem deserving all the attention it has had. Judgingfrom our accomplishments to date, one might.suggest that even more or broader effort shouldhave been expended through the years.

But I am concerned that, for all its importance,a focus on the mechanisms and decision rulesfor assigning personnel of various skills to groupsof nursing tasks will obscure larger issues. Sucha focus will certainly obscure the real emphasisof much of the research carried out during the1960's. If I may speak reminiscently of our ownexperience at Johns Hopkins, I can only capturethe whole story by saying that from the outset wewere hopeful of developing a rational system ofinpatient care, considering all resources, humanand physical, as potential elements of that sys-tem. If nurse staffing received special emphasis itwas because of the dominant aspect of nursingshortages and personnel turnover.

In those days there was an abundance of pa-tients, doctors, and beds, a growing army of ad-ministrators and ancillary specialists, a stream ofnew equipment and labor saving devices, and aprogressive shortage of nurses. It was to corrector compensate for that shortage that hospital ad-ministrations sought innovations. However, in theprocess, new strains were imposed on nursing per-sonnel, for although substantial assistance didcome from the centralization of such support serv-ices as sterile supply away from the nursing units,new problems of communication and coordinationwere created as a consequence.

Seeing the nurse burdened with a residue ofnonnursing tasks and simultaneously obliged toplay a large role in ensuring the logistical supportof her unit, research took a new direction. Seekingon one hand to introduce more new technology,more external support, researchers attempted torestore and improve upon the cybern'etic integrityof the nursing unit, which had existed in the claysof internal self-sufficiency by creating new mechan-isms of communication and coordination.

Think of a nursing unit in those terms. Itselements are its patients, set of doctors, teams of


nurses, physical setting, and supporting services.The goals of the nursing unit reflect the goalsof the larger institution of which it is a part.Care of its patients, training of its physicians,nurses, and other personnel, perhaps some re-searchall these are the stated objectives of theunit. To understand the behavior of the actorsin the hospital inpatient unit, one must also beaware that individuals may be guided by profes-sional values quite apart from the immediate de-mands of the work setting or the formal goals ofthe institution. -

To gain insight into the complex processes ofpatient care, various investigators have constructedmodels. Daniel Howland, for example, has pro-posed the concept of the nurse-patient-physiciantriad as a subsystem whose objective is the mainte-nance of homeostasis in the patient, and whoseillness is seen as the inability of his own sub-systems to maintain homeostasis. How well thephysician and nurse performthe quality of theircareis evidenced by the promptness of their de-cisions and actions in restoring physiological norms.

Given the diverse responsibilities of the physi-cian and nurse in the simultaneous care of manypatients, traditional roles have evolved for each.Though time has changed their roles somewhat,one can still identify formal functional relation-ships. The nurse has continuous custody of theinpatient and, through the physical design of thenursing unit, has the capacity for continuous ob-servation or awareness of the state of the patients.The physician sees patients intermittently but isresponsible for signaling actions to be taken on thepatient's behalf. Thus the activities of the nurseare responsive to two sets of signals, one fromher own communication and observation of thepatient's conditions, the other from the set ofphysicians' orders, which she may have had someinfluence in shaping.

An examination of the whole nursing unit as aset of triads of patients-doctors-nurses devoted tomaintaining homeostasis in individual patients doesnot reveal anything corresponding to homeostasisin the nursing unit itself. There is in fact noconstancy, no convenient norms to be maintained.Instead there is a constantly and widely changing

'See the series of papers by D. Howland and W. McDowellin Nursing Research, 1963 and 1964, and American Journalof Nursing, 1966.

FACTORS RELATED TO NURSE STAFFING AND PROBLEMS OF THEIR MEASUREMENT 27

pattern of aggregate patient needs. l'he cyberneticsystem problem is to formalize the process of ob-servation and respond in such a way that resourcesand skills can he called forth over the course oftime in amounts appropriate to the level of need.

In quantitative terms, this means matching twoerratic and not very controllable variables: theaggregate hours of direct patient care elicited(the direct rare workload), and the total numberof nurse hours available to render care. The natureof the problom can be seen in figure 1, which isa routine daily forecast of nurse hour requirementsfor a set of nursing units, a statement of scheduledhours for the day, and a gross estimate of the

3 See Fla*. C. D. "A Decade of Operations Research inHealth," in New Methods of Thought and Procedure, Zwicky.F.. and Wilson, A. C... editors. SpringerVerlag, 1967. Thehospital organisation chart is depicted as a set of phsician-patient relationships. with line delegation to nursing and theremainder of the hospital as staff support.

expected shortage or overage of personnel on theunits.4 It is usual for the workload index to varyfrom 15 to 45 nurse hours on 30-bed units, whilethe total nurse hours available range from 35 to37.

What causes these variations? To what extentare they controllable and whose decisions and. ac-tions are required to exert such control? How arethe vagaries of patient needs and available per-sonnel interrelated to arrive at nurse staffing prac-tices? Consider first the demand factors related topatients.

'Connor, R. "A Work Sampling Study of Variations inNursing Workload," Hospitals, May I, 1961. It would bedifficult to defend the accuracy of the fixed or variable nursehours (figure I) transferred to another service, 10 years later.Nevertheless. similar work studies elsewhere confirm theapproximate magnitudes, and the estimated gap betty !enprojected needs and resources is a useful signal for necessaryaction.

Patient Characteristics as a Factor in Staffing

On any given day, direct care activities in anursing unit are dependent on a number of aspectsof the patient population that day. These mayinclude the following:

(a) the total number of patients in the unit,(b) the physical and psychological dependency

of patients on nursing support,(c) the content and sequencing of physician or-

ders,(d) day of stay of patient,(e) legal requirements and hospital doctrine.

Total Number of Patients in the Unit

Although mere presence as a patient is not amajor determinant of workload, it does require aset of activities, a flow of visitors and inquiries,the maintenance of records, communications, andpreservation of the nurse-patient relationship.

Dependence on Nursing Support

Much of the practical outcome of the nursingresearch under review assumes that measures ofpatient dependency are adequate bases for pre-

dicting nursing needs. Dependencies have beenrelatively easy to define and observe objectivelyand such measures, or the patient classification as-sociated with them, do correlate empirically withpatient care as rendered, The weaknesses in usingdependency classification alone for staffing deter-minations are several. In particular they do notindicate time of day of required actions, and theyforce generalizations of past experiences to thepresent and future. Their virtue is that they arean easily obtained signal of abnormally light orheavy demand ahead.

37

The Content and Sequencing of Physicians'Orders

Much of what nurses will do on any given daymay be inferred from the set of physicians' ordersin effect for that day. The paper by Dr. Simborg(1) to be presented at this conference demon-strates the usefulness of orders, not only in estimat-ing total workload but in phasing nurse activitiesthroughout the day. The measurement problem inusing physicians' orders as a guide to staffing liesin the transformation of an order into a set of tasksfor which there are nurse times associated.

cc

Figure 1.Calculation of nursing staff needs.

Date

8/2,9/742B

uildingA

.)0Nt1141.

..-gE

H='1CD

c..)

Degree of care

required

44-

"Po-J

Hours

kscheduled

T2

1

-c'a)

2(.,(1)

ri=o-''.:

gcffs"0

...2

0.c)

a01

0 ...1-1"2, e)(9O

=c

in

Se';'=

......In--c'01r fa=

-e..2-E%

.- Cri

ti:'4,4_4.1)Z

-

0 '0:LI----is-is c0tu._

1311.a- i:(7)2)cu z

CI V

I

RE

MA

RK

S_

I

7-oI

..._

73rercsa_

.I

(7)cn

Pro

Aux

Total

4--X-0

4z24

94

1132

5Z/5

/833

-19-/

/8iliev

31.frqz

ooq,

43/8

52

//20

4095

/5:525

-152 5

43.61-1R

Al

822/

6.-2

/42/.5

4/.5/5

2/36

--56-iLig

-20.5:,#,-40,9- A

44-ivc7-7! 74"Szti

5323

72

1426.5

4.5/5

1833

- /3S4-2.5

1-1-

AIL

g4. /2

3/2

1434

2/IV

3q#i

1-3.5-1-1-5

,vo

10428

/3b2

JHH

1-007, 11/64t at

2:30 p.m.

f 9:00 a.m. - 6:00 p.m

.

ur d 00:9 - e 00:6 + "urd OVZ W t9/TI LOO-I HHr

gv II a tool,

5'£- 5v- *4- g/ a yg 5/ *I S Z bl 1 71

1,1i

ten. E .1-

aswegilvarrpor, Ar, ''''Y 4, 4 lar Na' 4,

siz- 5'81-

."14. 5 v 51- bb 1 8/ 1,5 17i 11 1 b a ig

(2°1 ria""/ '' eikv nr i -"lc 9- -k

roi-i- 5/- 5:0-1- 5- oi 1r b 5i 5/ OZ 0 Z zZ rg

(z-evadwai) /vat -ow.* - -xrpod-oiv,-

ivory -fin,' , - I

Z-f- 50- 5'ft li s'A 51Z -51Z YA 51 Z 0 LI Su

II 0 0 II

(E-ticy)(Y--i-c4147- iv-di-1.57 1

SW- 5V- El- 1E 51 It 6* hZ 8 .9 s / ze

S)IIVVV321

to c, a rt, ,cya

a 771-

0 ci. et o A) = ft

,74: g-ic

-,

---, v) -r 13)

New =-=- 0 0 X c

ft

--a_ c N

to 0 0 q a

rs.),-. 'cog a , -0 9,

- 0 AT

Z3 fa

= c a to n =- 2. c 11

__I 0 k

le101 xnV aid

r-- 0 Su

aL

-fr

s 0_ no x

--4-

(.1) CD

=i; .

13 A) P: As

I

-4 0 ID

1

2 in c in

73- g

painpaps .4'

smoH

paqnbai

al o Jo aathaa

. - ____ vneerio -Pie/nit/9M bulPlina

(Panul)uo3) I aln2!I

OL

30

Day of Stay of Patient

Evidence continues to accumulate to quantifythe difference between fi-st clay 'A stay in a hos-pital and subsequent days. Moores (2) reports ahalving of direct care after the sEcond day. A.corollary is the suwInent that nurse hours perpatient day vary inversely with a hospital's averagelength of stay, but this statement obscures the ex-

planation of significantly large transaction timesrelatal tc admission and discharge days.

RESEARCH ON NURSE STAFFING IN HOSF1 rAts

Legal Requirements and Hospital Doctrine

Society imputes certain characteristics to hos-pital patients vt d to those who serve them. As aresult some mandatory staffing practices have sig-nificant influence ()It staffing ratios. Registerednurse coverage during the night and preparationand administration of drugs by professional nurses(RN) are examples. Assessing the effects Jf suchrequirements does not pose severe measurementproblems. Work measurement studies of the ac-tivities and frequency of performance estimatessuffice to show the magnitude of these factors.

Organizational Response as a Factor in Staffing

Given the set of responsibilities resultant of pa-tient conditions, physicians' orders, legal require-ments. and nursing doctrine, we can address :.Leproblem of organizational response as it affectsstaffing. A number of measures may be used tocharacterize staffing: e.g., the familiar ratio ofnurse hours (or nurse wages) per patient day andbudgeted or total staff positions per bed.

Some of the rational determinants of these num-bers are as follows:

size of the nursing unit,mechanisms of communication and controltu)it,external supports of the nursing unit,architecture and technology,organizational factors,administrative policy.

Size of the Nursing Unit

In any service system subjected to randomlyvariable demands, a reserve of resources must beprovided to accommodate periods of peak load.The statistical law of large numbers tells us thatfor any specified probability of available servicethe necessary reserve capacity is proportional to thesquare loot of the capacity required for averagedemand. By way of example, the reserve for peakdemand in a I6-bed intensive care unit may behalf again as large as the base staff. This arguesfor larger nursing units to achieve economies ofscale, although there are other ways to handle the

40

problem of random variation in workload. Moores(2) reports a smooth drop in staffing ratio fromover four nurse hours/patient day in units of 10beds or fewer to 2.5 or less for units of 30 to 37beds.

Mechanisms of Communication and Control

From an administrative point of view, thesimplest way to determine staffing requirementsfor a nursing unit of given size and service is toohserve its pattern of patient demand over a pe-riod of time and assign staff to meet observedpeak loads. This is an easy but costly processleading to large amounts of idle time. Given somemechanism fOr short term prediction of patientneeds, say a clay in advance, several arrangementsare possible each of which in its own way permitsaccommodation of peak loads without a large re-serve staff.

Progressive patient care.By grouping patientsaccording to level of intensity of illness or de-pendency, the variations in demand within eachgroup are relatively less than the variation of anursing unit in which all levels of intensity arerandomly mixed (3).

Variable staffing.It has been observed that themajor source of variation in aggregate patient de-mands on any day is the number of intensive carepatients in the unit, followed closely by the num-ber of admissions. Given short term forecasts ofeach of, these numbers, it id possible R. predictroughly the number of direct care hours needed. It

FACTORS RELATED TO NURSE STAFFING AND PROIII.EMS OF THEIR MEASUREMENT

has also been observed that workloads in the manynursing care units of a hospital are not correlated,except for seasonal variations or epidemics. Again,the law of large numbers suggests that the abilityto share personnel among nursing units can reducestaff requirements. or conversely, can increase theprobability of local adequacy of staff given a fixedtotal hospital staff (1).

Controlled admissions. It was noted earlier thatthe first day of patim stay, particularly for non-elective admissions, imposes more than twice thedemand on nurse staff than do subsequent days.Given some choice in placing a new admission plusknowledge of the number of intensive and firstday patients on each unit, there is a potential forstabilizing workloads by selective admission tounits. Theoretically, controlled admission can havethe same effect on staffing requirements as pro-gressive care or variable staffing. In practice it hashad some drawbacks:

In teaching hospitals, the criteria for patientplacement by service and type of care narrowthe options for localizing any particular pa-tient.Even with controlled admission there are stillsubstantial variations in workload, calling forvariable staffing as well to match aggregate de-mands to nursing resources.

Variably work routine.--In searching for a poolof nursing time to draw upon for peak patient de-mands, one should keep in mind that direct careconstitutes only about one-third. of total nursestaff time. A priority ranking of activities of theremaining two-thirds may give clues for combiningdaily forecasts of patient needs with other tasks tocreate daily routines in some detail. The detail involvecl is immense, but it can be aided hopefullyby computer based information handling proce-dures.

External Support of .,ie Nursing Unit

There have been numerous studies and admin-istrative moves to transfer some tasks traditionallyperformed by nurses to administrative depart-ments in the hospital or to externally purchasedsupplies or services. The same arguments for eron-nnies of scale that support the notion of largenursing units also point to efficiencies in managinghousekeeping, dietary, escorting, and sterile supply

functions. It is simple enough to measure the bitsof nursing time released by each elemental tasksupported or removed, but it is very difficult tofind an influence on the staffing ratio. In part thisis due to the fragmented nature of the time re-leased. In part it reflects that the aggregate of suchtime savings may be outweighed by the admissionof one or two intensive care patients.

Architecture and Technology

The comments above apply equally well toarchitecture and labor saving technology on thenursing unit. Specific elements of time saving canbe demonstrated, but it is difficult to trace theelements to a justifiable reduction in staffing ratio.A primary effect of architectural layout is ontransit time of nurses in the course of their duties.The measurement technique called "link analysis"is quite useful for observing the intensity of trafficbetween important nodes in the unit. Havingidentified the busy links it is often possible toshorten them by making the nodal points con-tiguous.

Various studies have shown transit time reduc-tions to be effected by placing medication roomsand critical patient rooms close to the nursingstations. A work corridor has been shown to halvetransit times if nursing teams remain in the cor-ridor and near their patients.

Technology, in its accustomed role of lifting theburden from man's back, has had little to offernursing. It is most effective in taking over routine,repetitive tasks, and these are generally scarce inhuman services. The promise of technology, as yetunexploited, lies in its potential role as part ofthe communication and control process. Indeed theonly service areas where technology has made contributions to productivity comparable to those inindustry and agriculture are those in which alarge amount of paperwork is involved. We knowthat formal communications on a nursing unitaccount for as much time as patient care itself.Serious efforts are under way to recover some ofthis time through computer based systems, butthese are in difficult development stages.

41.

Organizational Factors

In its heyday the high productivity of American

32

labor was attributed to techniques of specializa-tion and division of labor. The notion of dividing atotal task into elements and assigning each elementto a well trained well equipped worker underliesmost of our mass production techniques. If taskscan be balanced and machine pace,: on a preciseschedule, such a system can be nearly 100 percentefficient in utilizing its man days in the short run.It is, however, vulnerable to interruptions andbecomes costly when used to less than its capacity.

In a way, specialization and division of labor areconsistent with the guild system, in which minortasks were handed to apprentices with major tasksreserved for masters. If the apprentice wearied ofthe menial tasks assigned him, he could comforthimself with the thought that the wry process as-sured that one day he would be a master himself.It has been this way with bricklayers and plumbers,lawyers and doctors, but not with the nursingstaff. In the absence of both upward mobility ofauxiliary workers and the tightly controlled set ofoperations on a nursing unit, one wonders whethera hierarchical distribution of tasks can be efficientor effective in the long run.

One alternative to the hierarchical distributionof tasks is to require versatile performance by mostmembers of the team. Indeed in two experimentalunits in the San Francisco Bay area, the KaiserFoundation Hospitals have demonstrated a reduction in staffing ratio from 6 to 4 nurse hoursper patient day by assigning full care responsibil-ities of groups of eight patients to each staff nurseassisted by nursing aides. An assessment of thehierarchical versus versatile approaches to nursingcare poses difficult measurement problems forwhat must be captured lies in the subtle dynamicsof performance of a sequence of tasks by morethan one person. Nevertheless the most significantinfluences evident on nurse staffing are in the areaof organization of tasks, where the issue is speciali-zation versus versatility.

Administrative Policy

The variables discussed so far may be rational


but they are not powerful. All can be thwarted bynoncompliance on the scene of patient care or byadministrative fiat. An administrative decision tobase nursing oudget on a rule of thumb, say 3.2nurse hours per bed day, assures that the onlyvariable having an influence on staffing ratios isoccupancy rate. Under such circumstances thetotal cost of nursing to a community would bestrongly related to the number of hospital bedsconstructed.

Empirical Evidence of Factors AffectingNurse Staffing

fx an-attempt to Astra the influence of many ofthe factors discussed here on nurse staffing, a sam-ple of 31 Illinois hospitals was studied (5). Meanswere devised to measure such variables as hos-pital size, nursing 'mit size, architectural configu-ration, degree of external support, managementfactors, state of technology, patient satisfaction,staff 4nd administrative characteristics. No effortswere made to control these variables, but correla-tions were sought between their natural rangingand the observed range of 3.78 to 5.50 in thedependent variable, nurse hours per patient day.The strongest relationships were found in thosevariables related to physical plant. The largernursing units and the larger hospitals operatedwith smaller staffs in terms of nurse hours perpatient day. Also, those nursing unit designs thatrequired less nursing travel to perform duties op-erated with fewer nurse hours. Patient satisfactionalso appeared more related to environment (noise,ventilation, comfort) than to nursing care.

Variables of a day-to-day nature, such as mix ofintensity of patient dependency, even with the at-tempts to accommodate this variable by provisionof an intensive care unit, did not correlate withnurse staffing ratios. Whether there is an inherentweakness in many of the variables we have pos.tulated or whether the administrative process isinadequate to take advantage of them is not revealed by the Illinois study, but the questionsraised demand further inquiry.

FACTORS RELATED TO NURSE STAFFING AND PROBLEMS OF THEIR MEASUREMENT 33

Another View of the Nurse Staffing Problem

One emerges from the foregoing considerationsoverwhelmed by the weight of it all. So much ef-fort has gone into research related to nurse staff-ing, so many ideas have been espoused, and solittle can be prescribed as a result.

If the problem of nurse staffing in hospitals werenot such an old one, we might have begun onanother tack, asking fundamental questions aboutthe objectives of the health care system and thepossible roles of nursing in achieving those objec-tives. We may find an optimal balance amongpreventive care, ambulatory, and inpatient carethat would diminish the total number of inpatientdays of care. One signal of success in achievingsuch a balance in the large system would likely bean increase in nurse staffing ratios and per diemhospital cost, as outpatient care is substituted forsome hospitalization.

We have the opportunity to ask in some detailwhat the beliefs, aspirations, and goals are of vari-ous segments of society relative to health care andhospitals. If we wish to confine our considerationsto hospital care, we can do so after visualizing thehospital's goal in a larger system and forecastingthe characteristics of the inpatient census. Surelythere will be inpatients, and we can begin byasking their wants and objectives. What are theobjectives of administrators, physicians, and nursesthemselves? Are some of the objectives conflicting?If so, how is the conflict to be resolved?

A well stated objective should imply measures

by which one can test the degree to which theobjective is achieved. We have dealt with staffingratio as a measure of effectiveness in nursing care.What objectives does it reflect, or more important,what objectives does it fail to reflect?

If it is possible to list a coherent set of objectivesand their associated measures of effectiveness, anext step in systems analysis is to ask what varia-bles affect these measures of effectiveness that canbe regarded as outcome variable.

We seek a definitive list of two kinds of inputvariables: controllable and uncontrollable. Wemust then estimate the influence of tacit of theinput variables on the outputs. This is essentiallywhat we have been doing in a limited way in thispaper and in the research it describes. If as aresult we had, over the years, produced an ade-quate supply of nurses who were happy in theirjobs and who, along with physicians, administra-tors, and patients themselves, were satisfied withthe care rendered, then no more need be said ofobjectives and variables other than those we hadbeen considering explicitly.. But the fact is thatmany objectives are still to be achieved. We stillare not able to specify an ideal nurse staffing ratio.The factors we have considered have not beenshown by themselves to explain or control thestaffing ratios experienced. We must continue ourinquiries and experiments to bring to bear newknowledge on the old and deeply human problemof nursing care.

43

34

References

(1) Simborg, D. "Impact of Computerized In-formation Systems in Hospitals on NurseStaffing."

(2) Moores, Brian. "Reallocation of StudentNurses to Match Patient Requirements inDifferent Hospitals on the Assumption thata Shortage of Staff Leads to Increased Wast-age." Unpublished Doctoral Dissertation,Johns Hopkins University, Baltimore, Mary-land, 1971.

(3) Weeks, L., and Griffith, J. "Progressive Pa-tient CareAn Anthology." University of


Michigan Graduate School of Business Ad-ministration, Bureau of Hospitals Adminis-tration, Ann Arbor, Michigan, 1964.

(4) Flag le, C.D. "The Problem of Organiza-tion for Hospital Input Care." ManagementSciences: Models and Techniques, NewYork: Pergamon Press, 1960.

(5) Illinois Study Commission on Nursing,Nurse Utilization, Illinois: A Study inThirty-one Hospitals. The Illinois Leaguefor Nursing and the Illinois Nurses' Associ-ation, Chicago, Illinois, 1966-70.

DISCUSSION OF DR. FLAGLE'S PAPER

Discussion

Dr. Alexander BarrChief Records Officer and Statistician

Oxford Regional Hospital BoardOxford, England

Leaders

Dr. Richard C. JelinekVice President for Research and Development

Medicus Systems Corporatim.Chicago, Illinois

Dr. Barr

Good morning, Madam Chairman, Ladies, andGentlemen. It is always a great pleasure for me tolisten to Dr. Flag le, and I am happy to commenton his paper. If I may use the blackboard, I shalluse my 5 minutes this morning to draw you adiagram (see figure 1) illustrating a researchframework into patient care and nurse staffing inhospitals. This gives me an opportunity of indi-cating some of the things we have been doing inOxford, and I hope it also illustrates that thereisn't a very wide divergence between our thinkingin England and the thinking here today.

Our basic prcblem, or the one we think wehavelet's start that wayis to define hospitalstaff complement. As far as we are concerned inthe United Kingdom, we have tended to look atthis in two separate ways. First of all there isdeployment of staff and, in particular, deploymentof nursing staff. This can be divided into twoseparate areas of research. One of these istrainingtraining of nurses or training of anystaff, and the other is allocation of staff to the jobs.

On the other side we must look at the serviceelement, the actual running of the hospital and thedelivery of care to the patients. This also has twobranches. One of these is the quality of care thatwe want to give. The other is the quantity. Wehave already heard something about quantity andquality and been given to understand that thebridge between them is a bit flimsy at the mo-ment.

35

45

Research is going on in each of these areas, butthere is a certain amount of speculation as towhere we go in the future. Hopefully, these vari-ous elements, research components, will somehowwrite them into a special box that we will have tolook on at the moment as a black box of totalpatient care. If we could define this black box, theresearch framework could be drawn with greaterconfidence. The real danger is that in our strivingfor efficient organization the real business of pa-tient care will get overlooked or given

14i. low, pri-

ority,Let me tell you a story about an American

general, which may help to illustrate the sort ofrelationship I think we have gotten ourselves into.An American general was supposed to have had aproblem in Vietnam. As is usual nowadays, heasked a large computer in the States to solve theproblem for him. We don't do our own thinkingnowadays; we ask a computer to do it for us. Thegeneral cabled the computer and said, "I'm sur-rounded on the north, south, east and west. Whatshould I do?" And the computer cabled back andgave him the very terse reply, "Yes." Nothingelse, just, "Yes." And the general became a bitirritated at this and he cabled back to the com-puter, "Yes, what?" And the computer returnedthe reply, "Yes, Sirs"

How can the problem be defined in a way thatit can be answered intelligently? I rather suspectthat we have done much of the research because it

-if


Figure 1.Interrelationship of staffing and patient care.

Deployment

Trainingof staff

I

Hospital nursingstaff complement

Al locationof staff

Service

Qualityof care

Total patient care

Quantityof care

DISCUSSION OF DR. RADIX'S PAPER

is relatively easy to do. Topics are selected becausewe can, in fact, get solutions to them, and it alwayslooks better to have some sort of a solution inhand. However, as I have already pointed out,perhaps we should have been asking a differenttype of question. What do we want to give thepatient? What is total patient care? And follow-ing this, what does the nurse do about the deliveryof this care?

Sooner or later hospital care will have to be re-lated to the community services because obviouslythe sort of care that is given in the hospital hasits counterpart in the community itself. Early dis-charge exacerbates the load on the communityhealth staff. If patients are not discharged veryearly, you obvichisty live It --ii iitttent pattern ofcare within the hospital, perhaps relieving thecommunity of a lot of its present problems. Thepoint worth emphasising is that these are inter-dependent and not independent services.

Just before I finish, may I draw your attentionto one or two specific things we have been doingin this field? On the training of staff we havefound some conflict between training and serv-ice. What the Training Council requiresTheGeneral Nursing Council in Englandsometimesis at variance with the actual hospital needs asfar as services are concerned.

If the training model is unbalanced (for ex-ample, the time spent by students in a specialtyis too long), it has repercussions in the staff com-plement. You can, in fact, have increased staffcomplements just because training is scheduled in-efficiently. On quantity of nursing care, the nurs-ing dependency based on the work Dr. Flagle hasdone in Johns Hopkins is very useful for measuringthe quantity of nursing care actually given to the

37

patient. Most of us recognize there are limitationsto the actual method, but in fact it has some veryuseful practical elements attached to it.

On the quality of care, I think most peopleagree this is a very difficult area to work in. As apreliminary we have looked at what the staff thinkabout the care given. We have asked professionalnursing staff of all grades along with 'doctors andparamedical staff to classify wards into good orpoor care, into strict and slack discipline, and intohappy and unhappy wards. This has been very,useful because there is a fair degree of conformityof thought among the staff about wards in general.Dr. Revans probably has a lot of information onthis as well.

We also extended that to ask patients what theythink about the l-lspital service provided, andagain there is a high degree of correlation betweenwhat the patients think and what the staff think.Perhaps we haven't made enough of the con-sumer reaction in hospital affairs. We should bemore attuned to what the patient thinks about theservice.

Then, on the black box of total patient care,which is probably the most difficult area for all,we have recently been looking at the possibility ofusing videotape records of the actual service pro-vided in the ward. This is a very promising tech-nique. It enables us to capture data and to useand assess it over and over again. The trouble, ofcourse, is in reducing this data to some sort ofmanageable indices, but again I don't think thatthis in itself should deter us. If we did capture thiskind of data and showed it to the relative pro-fessional staff, we might through experiment anddiscussion arrive at a stage where we could sayw:tat total patient care is all about.

Dr. Jelinek

I am probably the wrong person to critique Dr.Flagle. As you all know, Dr. Flagle is consideredthe "father" of operations research applications inthe health care field and in a way is like God. Tocritique him would be the same as to ask you tocritique- God and, therefore, I have to agree witheverything Dr. Flagle said.

However, I'd like to make a few points, most ofwhich directly support some of the statements Dr.

. . 47

Flagle made. I have been for 6 years in the aca-demic environment involved in health care re-search. For the last 3 years I have been associatedwith an organization attempting to implementsome of the research ideas. I would like to use ex-periences stemming from both the academic andthe operational settings to comment on Dr. Flagle'spa per.

The problem, in my opinion, in the area of

38

nurse staffing is that we do not have any objectivemeans of relating staffing to the level of service.This can best be illustrated by suggesting that inmost hospitals it is not possible to observe anysignificant impact on the level of service or qualityof care resulting from a change in staff size. Mostcertainly sur.h impact is difficult if not impossibleto quantify. Patient care will not go up a great dealif you double the staff, nor will it decrease signif-icantly if you reduce the staff by 50 percent.

Of course, this is a tremendous problem. Howdo you staff if you cannot evaluate a situationsuch as I have described? This, in fact, is whatDr. Flag le pointed out through some of his illus-trations. As his presentation has shown, on a par-ticular patient unit the staff may he twice as highon a per patient day basis one week as it is anotherweek. The impact of this variance on patient careis not understood, nor is it measured. We also seetremendous differences in staff site and in cost ofrunning a patient unit from one geographic areaof the country to another. It is not clear that qual-ity of care is related to these regional staffing andcost differences. Of course, the main probletn is

that we do not have any clearcut measure.; of qual-ity. Such measures will have to be derived beforethe staffing problems can be resolved.

For purposes of discussion I would like to raisethe following question. If in fact there exist suchlarge differences in staff on a per patient day basisbetween institutions without comparable differ-ences in quality of patient care, why not immedi-ately cut staff size in those hospitals that show rela-tive overstaffing? Other than major reduction incost, what would be the impact of such action? Iam just asking the question, not making a recom-mendation.

I would like to make some general commentsregarding another dimension of this area of re-

search. A great deal of effort has been devoted tothis subject and many conceptual ideas have beenadvocated, including patient classification scheLies,variable staffing procedures, selective admissionsprocedures, and reporting systems. However, veryfew of these ideas have been implemented in a

broad scope. For example, we see many hospitalshaving variable staffing, but when we look closelyat their system it imposes little control over match-ing staff size to workload. The implementation

RESEARCH ON NURSE STAFFING IN HOSPITAI.S

therefore is often found more in works than inactually achieving the established objectives.

To conclude, I would like to make a number ofobservations and raise a number of questions related to this area of research.

First, the problem that faces all of us is theneed to establish objectives for the nursing pro-fession as a whole. What is nursing? For us asengineers working with nursing it is hard to doanything until we understand what it is that nurs-ing is trying to achieve.

Second, there exists a problem of explicitly de-fining certain terms used by nurses in identifyingthemselves. For example, what is clinical nursing?We keep hearing the difference between task ori-ented nursing and clinical nursing. We have diffi-culty getting a precise definition of what is meantby clinical nursing, clinical nurse specialists, clin-ical leadership, etc. Many terms have been used todefine the nurse and her role.

What is the psychosocial element of care? Anyengineering study that comes out in terms of ob-jective measures k generally criticized in terms oflacking an understanding of the psychosocial as-pect of care. What exactly is it? We know it inwords. We have difficulty quantifying it.

What is meant by continuity of care? We knowthere are advocates of continuity of rare. One waythis is expressed is to propose that the same nurseprovide continuous care to the same patient. Butwe also know this is impossible to achieve totally,because the nurse has to sleep and take her daysoff. What then is continuity of care?

Next is the question of relating research to op-erations. Are there enough ties? We see nursesdoing research in the academic environment. Arethey doing this research for the sake of publishingand being recognized among their peers or arethey doing it to improve operations? 1 believeresearch can be clone for the purpose of beingrecognized in the academic community, but thequestion I'm posing is, Do we have enough of atie between research and operations?

Even more important is the tie between educa-tion, and research and operation. I think this is anarea where the Division of Nursing is playing animportant role in trying to establish a better tie,but a significant amount of additional work isneeded.

DISCUSSION OF DR. PLAGUE'S PAPER

I have a general comment regarding the alter-native forms of organizing patient care operations.As Dr. Aydelotte pointed out, we now see manyorganizational systems in operation. For example,there are many variations of unit managementsystems. We have many variations of the use ofclinical nurse specialists. All of these systems arefeasible in that they get the job done, yet we are

39

not ?hie to establish what the effect of each ofthese systems is on quality of care and cost.

My observations boil down to stressing the tre-mendous need for establishing objectives for nurs-ing and for devising methods for measuring leveland quality of care. I believe that if this conferencecould somehow locus or direct us as to how weshould handle these problems we could make acontribution to the field,

49

Measurement of Patients' Requirementsfor Nursing Services

Mrs. Phyllis GiovannettiProject Director, Nursing Research

Hospital Systems Study GroupUniversity of Saskatchewan

Saskatoon, Saskatchewan, Canada

IVe dance round in a ring and supposeBut the secret sits in the middle and knows.

Robert Frost

I am pleased to share with you the results ofthree research studies conducted in Canada as wellas many of my own views on the subject of pa-tients' requirements for nursing care.

Although none of the research investigations setout principally to measure patients' requirementsfor nursing services, they did come face to facewith this problem at an early stage. They focusedon the identification of patients' nursing careneeds, and where attempts were made to quantifythese needs, the researchers had to conduct theirinvestigations within the realm of actual caregiven. We are still faced with the prohlem ofevaluating whether in fact the care given is thatwhich is required.

The first study developed an index of care, adocumentation of the types of nursing activitiescarried out in Department of Veterans' Affairs

41

(DVA) Hospitals. The second investigated the re-liability and validity of a completely subjectiveclassification tool for determining patients' re-

quirements for nursing care.The third study, somewhat more comprehen-

sive than the others, began with the developmentof an objective classification tool that was subse-quently quantified to produce an index of work-load. In addition, this study developed and tested anew concept of ward organization, the Unit As-signment System. All of the studies were supportedin whole or in part by Canadian National HealthResearch Grants.

The underlying theme has been predominant inthe vast majority of related studies of matchingnursing resources to the nursing care requirementsof inhospital patients. The keynote was effectiveutilization.

Filially, some comments regarding guidelines forfuture research are presented. A descriptive modelfor the study of nursing as an interdependent vari-able in the health care system is discussed.

42

An Index of Care

The first investigation (1) was carried out atDVA Deer Lodge Hospital, Winnipeg, Manitoba,in 1965. It arose out of an interest in modifyingmorbidity and mortality data to show the type ofpatients being cared for in DVA hospitals acrossCanada. Prior to the study the statistical data col-lected under the Veterans Treatment Regulationsprovided little information to indicate the type ofcare the patients were receiving in the hospitals.A classification of patients according to the typeof care provided to them was required.

A questionnaire was devised utilizing a selectionof procedures that appeared to best indicate nurs-ing efforts. This questionnaire became known asthe Level of Care Assessment Return, a checkoffsheet of some 55 items to describe a patient. The55 items were grouped under the following head-ings: Clinical Monitoring, Technical Nursing,Nontechnical (basic) Nursing, Physical Medicine,and Organized Psychiatric Unit.

At the time that the Level of Care AssessmentReturn was being completed on every patient inthe hospital, the patients were all classified accord-/mg to seven predefined care categories rangingfrom intensive to residential.

Analysis of the results of the survey showed thatalthough the amount of direct nursing care rangedfrom a large amount to a small amount, only threelevels of care could be identified: Level Iminimalamount of direct nursing care, Level IImoderateamount of direct care, Level IIImaximum amountof direct nursing care. In fact, the residential, in-vestigative, and convalescent care categories weregrouped into Level I (minimal care), the ex-tended treatment and intermediate care categoriesgrouped into Level II (moderate cal.!), and theacute and intensive care categories into Level HI(maximum care). Analysis of the completed Levelof Care Assessment Returns matched with eachlevel revealed the procedures of direct patient carethat characterized each care category.

A numerical value was assigned to each of the55 items of the Level of Care Assessment Return,bearing in mind the amount of effort involved inthe particular procedure. This step made it possi-ble to produce a classification of patients accordingto the levels of care by data processing methods.

Thus a method of classification of patients ac-


cording to the type of care provided to them hadbeen developed. It required only that the chargenurse complete a Level of Care Assessment Returnfor all patients. Following data processing, totalpatient census was divided into three levels of care.

Further validation was established in a numberof hospitals through the timing of the procedureslisted in the Level of Care Assessment Return (2).Both the direct and indirect care times associatedwith each procedure were obtained for each pa-tient level. Through continuous observation for 24hours the total amount of direct nursing care received by each observed patient was determined.The timing studies demonstrated that the Level ofCare Classification of a patient, correlated with theamount of care he received and measured in time,is a valid method of categorizing the severity of thecase.

The research resulted in a method of classifyingpatients, independent of observer judgment, intolevels of care and a system for estimating thenursing workload associated with each level. Inaddition, the analysis contributed a great deal toour knowledge about the variables influencingstaffing.

There has been, however, a limited amount ofapplication to non-DVA hospitals. I would proposethree major reasons for this. First, the classificationsystem is based totally on the physiological needsand care of the patient. Instructional and emo-tional needs, independent of the nursing proceduresstudied, do not appear to have been considered. Ac-cording to Dr. MacDonell, the assessment providedlittle evidence of emotional or psychosocial needsbeing met apart from that accompanying regulartreatments. Second, the nursing activities and proce-dures identified and their determined times maybe peculiar to the types of hospitals studied. Third,the patient mix in terms of kinds of care neces-sary, coupled with the treatment modalities, maynot be similar in other institutions.

Regardless of the limitations, real or perceived,the system does provide a more accurate accountof patient care than the usual census or diagnos-tic category estimate. For the DVA hospitals acrossCanada precise information regarding type of careprovided to patients at any given time is available

51

MEASUREMENT OF PATIENT'S REQUIREMENTS FOR NURSING SERVICES

and allows a more realistic judgment to be exer-cised concerning immediate needs of the hospitals.Trends in the types of patients for whom care isprovided may be identified on the basis o;accumulated returns, making it possible for pre-

43

(fictions and projections concerning future require-ments for facilities and personnel. Finally, theinformation can be utilized for staffing patternsindicating types and numbers of nursing staffwithin the hospitals studied.

A Subjective Classification Analysis

The second investigation (3) was conducted atthe Vancouver General Hospital, Vancouver, Brit-ish Columbia, during 1970. The study was initiatedin the interest of testing the reliability and valid-ity of a completely subjective patient classificationtool. A comprehensive review of classificationmethodologies had led to the conclusion thatexisting tools paid little attention to the patients'requirements for emotional support and instruc-tional needs and were too restrictive in terms ofthe predetermined criteria selected. In addition itwas felt that the more commou three and fourcategory classification systems were not sensitiveenough to the diverse needs of patients in acutecare hospitals.

A critical indicator, independent of observerjudgment, to encompass the diverse psychosocialneeds of patients was not apparent. Thereforethe notion of an objective checklist was discardedin favor of a completely subjective tool. A classifi-cation tool was developed with five categoriesranging from minimal to intensive care. It wasthis completely subjective tool that was selectedfor the rigid reliability and validity testing.

Five acute care hospitals provided the studysetting, using one adult medical and/or surgicalnursing unit of approximately 35 beds from each.All patients in the selected units were classifiedon five different occasions independently by fourcategories o. nurses. The head nurse classified allpatients. Registered nurses on the unit classifiedthose patients assigned to them. One nurse bornanother nursing unit of the same hospital classifiedall the patients as did a nurse from one otherhospital.

This procedure allowed researchers to test twomajor hypotheses of reliability. The first was totest for differences in the way various categoriesof personnel in a hospital classify patients: i.e.,

head nurse, float nurse, ward nurse. The second

52

was to test for differences in the way variousnursing personnel from different hospitals classifypatients.

It was found that there were significant differ-ences in the way different categories of nursingpersonnel classified patients, both within a hos-pital and between hospitals. Thus, using discreteanalysis, the subjective claisification tool as usedby the nurses could not be considered a reliablemethod of determining patients' needs.

There was a tendency, however, for the nursesto become more discrete in their assessment ofpatients' needs as their familiarity with the classi-fica on tool increased. An alternative form ofanalysis was presented that allowed one degree ofscale difference between nurse classifiers. Using thealternative analysi, the tool was found to be re-liable.

While the nursing personnel were classifyingthe patients using the designed tool, a nurse re-searcher obtained additional information on eachpatient. This information was comparable to theobjective classification tool developed at JohnsHopkins Hospital and to the procedures docu-mented in the CASH studies on staff utilization(4,5). The validity of the five category subjectiveclassification tool was tested and confirmed againstthe Johns Hopkins system. The patient character-istics data fashioned after the CASH studies werecollected primarily for their possible contributionto a measurement of workload related to the fivecategories. These data, however, were not analyzedat the conclusion of the report.

The study contributed a great deal to solvingthe problems encountered in establishing reliabil-ity in classification tools. In addition it identifiedsome of the problems to be reckoned with if weare to use classification tools to provide compara-tive information on more than one hospital. Evenin the face of the reliability restrictions the tool

44

can represent a gEoss indit atm of the nursing careneeds of the patients an indicator that is five


times more informative than the presently usedcriteria census.

A 5 -Year Study of Nursing Problems

I would like now to report on a third studyjust recently completed after 5 years of investi-gation. The study was conducted lr the HospitalSystems Study Group at the t iniversity of Saskatche-wan in Saskatoon, Canada, with most of the datacollected in the University Hospital in Saskatoon.Five major reports dealing with different phasesor aspects of the study were published during theperiod of investigation.

Phase I began in 1966 with the objective offinding an operational solution to arc problems ofreorienting hospital organizational structure intodirect contact with the patient and his needs (6).It is not surprising that the first step was to identifythe needs of the patients.

A four category classification system was de-signed to objectively categorize patients accord-ing to their needs for nursing care (see figure I).Font components of care were isolated as thecritical indicators of the patients' overall require-ments of nursing care. These componentspersonal care, feeding, observation, andambulation--were further broken down into in-dividual determiners that represented a patient'sdegree of dependence upon nursing. Two addi-tional determiners were found to be necessary:one concerning incontinence and other regardingsurgery.

The classification tool is in the form of a check.off list comprised of the 13 determiners, each des-ignated according to the amount of nursing in-volvement they represent: Aminirnal, 8average,Cintense.

Although the tool has the appearance of anobjective classification method, there is in factsome subjectivity involved, This is mote clearlyevident in the guidelines developed to assist thenurses in selecting the correct determiners of care.The guidelines, presented in appendix II, set downthe parameters for all of the critical indicators.Of particular importance is the observation com-ponent. This component provides for the inclusionof emotional and instructional needs of the pa-tients.

53

Identification of the characteristic needs of eachcategory of patient in terms of the amount ofnursing care time they receive, the type of staffwho care for them, and the type of functions per-formed was the second step. This was achievedby conducting direct care (inroom) activity stud-ies on both general medical and surgical nursingunits.

The results of the study documented a widevariety of problems:

There was wide variation in nursing caretime spent with patients of different categories.There were extreme fluctuations in workloadwhen one or more intense care patients wereremoved or added to the ward population.There was significant difference in allocationof functions among different levels of staff de-pending on the care category of patients served.There was variation in magnitude and incharacter of the workload to which staff mustadjust.There was lack of clear role definitions forthe various levels of nursing personnel on theward.Staff appeared to enter the toom to performa single function and then leave.Numerous people participated in the patients'daily bedside care, often for only short periods

time. Intercommunication and supervisionand checking of subordinates appeared to in-volve a considerable amount of time, effort andconfusion.Continuity of assignment was difficult or im-possible. resulting in limited opportunity fornurses to know their patients as individualsand thus to provide personalized patient care.Considerable amount of time was spent bynurses "in flight" between the center of com-munication (central nursing station), supplyarea (service room), and patient's bedside.The unpredictable nature of the workloadmade it impossible to staff in a manner thatwould ensure a consistent level of nursing carematched to patient's need. Periods of signif-

4EASUREMENT OF PATIENT'S REqUIREhtENTS FOR NURSING SERVICES

Figure 1.Patient classification tool.

WARDPERS)UALCARE

FEED-ING

OBSERVA-TION

AMBULA-TION

DATE

PATIENTS

c.,

4,-.

v0AOP0kZ44X40.04,0012oc.)

cov

OPOVI*elX 4+

04.m

ISi 03 01

I

4 cv

.00

0

0,4

RI01

....

c do:v

40

V

,..,

,.....

a.pi

A0.40

..4Pk0b

04

,-,um...4

4pi

'0

PX0P0

00

.--

v

A0'

i

N0'

......

ccv

00Vk

0-I

11

0

.c/0'

,-.um<v

P00

0 0.444 fi

0cos o

V 400000

0V pi0 VA

w0 04-4 0.

.-.

v

0

00 044+ 10

4,n4

al

0 4+

O4

AP '41As A

Or 1*= 0=

v

4+06+00'00IS

0.

.-.

0...4

0

el44000.01-4

0.-.

L.1v

0.01

Vk

A.

_

0X40.4Z<A

V

XkV*

B

000VP...10

00*RI

X*0AIC

C

40cmV1.,mu Comments

_

Brown, Mr. A.,

M A

.

X X, X 3 0 u 1 Minimal care

Wacker Mrs. W. X X X X 0 4 0 2 Avera e care

Smith, Mr. G. X X X X 0 1 3 3

Aoveaverage care

White, Miss G. X X X X 0 0 5 4 Intensive care-

... %.'---- ___../---s-----.-"---'---

45


icant overstaffing and gross understaffing re-sulted.

These problems concurred with those expoundedby the nursing staff during a period characterizedby high turnover rates and low morale. Duringthe course of the research, every effort was madeto find operational solutions to these problems.Since many of the problems were more clearlyexpressed as a result of the categorization of pa-tients, the next step was to validate the classifica-tion tool.

Validation was carried out by comparing theresults of the classification tool with the headnurse's subjective opinion and with the Deer LodgeLevel of Care Assessment Return (7). The headnurse was instructed to evaluate the patient's needfor care on the basis of her knowledge and exper-ience and to assign each patient to one of fourlevels of care ranging from minimal to intense.A project nurse classified the same patients usingthe Phase I tool and the Deer Lodge Level ofCare Assessment Return.

There was no .significant difference among thethree systems when the chi-squares test of good-ness of fit was Applied to the data. At the sametime, the Phase I tool was found to be easier andless time consuming to use than the Level of CareAssessment Return. A general description of thefour categories is presented in Appendix I. .

Workload Index

As a result of the extensive activity studiesconducted, the relationship among the four cate-gories of care, in terms of the average amount ofdirect nursing care patients received, was identi-fied. It was seen that on the day shift averagepatients received twice as much, above averagecare patient:, six times as much, and intense carepatients 12 times as much direct nursing care asminimal care patients. The ratio among the cate-gories on the day shift became identifiedas 1:2:6:12. Subsequently, the ratio of the averagedirect nursing care times on evening shift wasfound to be 1:2:7:14, and on the night shift,1:2:7:25.

Further analysis revealed that, on the average,patients received four times more direct nursingcare on the day shift than on the night shift andtwice as much care on the evening shift as on the

night shift. Thus the ratio of the amount ofnursing care received on clay, evening, and nightshifts was 4:2:1.

By multiplication of the day shift figures by fourand the evening shift figures by two, the result-ant numbers become the actual index and reflectnot only the relationship among categories on ashift but the relationship among shifts for eachcategory as well. Table 1 shows these relationships.

These ratios may in turn be used as weightingfactors for the corresponding categories on eachshift to arrive at an index of patient load on thenursing unit for any shift.

The procedure is very simple. If, for example,the ward census is composed of 10 minimal, 20average, 2 above average, and 1 intense care pa-tients the workload index for the day shift is de-termined by multiplying the weighting factor foreach category for the day shift by the number ofpatients in each category and summing the prod-ucts. That is, (10 X 4) + (20 X 8) +(2 x 24) + (1 x 48) an 296. The same pro-cedure would he used for the evening and nightshifts and, assuming the same patient mix, yieldworkload indices of 156 and 89 respectively.

The impact of the workload index when calcu-lated over a period of time on a nursing unitcan be very informative and the relationship be-tween census and workload most revealing. Table 2shows the patient load index for all three shiftsduring 1 month on one surgical nursing unit. Itcan be seen that the census ranged from a low of28 patients on the 21st day to a high of 38 patientson the 28th day. However, the day load indexon the 21st day was quite high at 492, but on the28th, with 10 more patients, the index was muchlower at 396. Similarly, on the 5th and the 12thof the month the census was equal at 35 patients,but the day load index was 512 and 272 respec-tively for the 2 days.

Table 1.Ratios of direct nursing care timereceived per shift and patient rattgory

Shift

Category

AboveMinimal Average average Intense

DaysEveningsNights

4

2

1

8

4

2

24

14

7

482825

MEASUREMENT OF PATIENT'S R EQC MEM ENTS FOR N"RSING SERVICES

BEST COPY AVAILABLETable 2.Daily patient load index

47

Patients per category Load index

Date Minimal AverageAbove

aierage Intense Census Day Evening Night

1 4 19 7 1 31 384 210 116

4 16 7 1 28 360 198 110

3__. 8 17 5 3 33 432 238 152

7 19 3 4 33 444 244 166

5 _ _ 8 15 9 3 35 512 286 206

6 . 6 17 10 - 33 400 220 110

7_ _____ . 6 IS 8 2 34 456 252 148

8_ ._ _ 6 21 5 - 32 312 166 83

9 5 23 3 1 32 324 172 97

10_ 9 20 - 1 30 244 126 74

11. 10 20 2 1 33 296 156 89

12 12 22 1 35 272 140 81

13 14 16 3 3 36 400 218 142

14 . 10 23 - 1 34 272 140 81

15 .. ... 9 16 3 4 32 428 236 162

16. 12 15 2 2 31 312 168 106

17 10 20 1 1 32 272 142 82

18 7 18 6 2 33 412 226 135

19 7 15 7 3 32 460 256 161

20 6 16 4 4 30 440 244 166

21 7 13 1 7 28 492 276 215

22 . 4 18NW

4 5 31 496 276 193

21 6 4 33 512 284 186

24 _. . 4 '440 8 3 34 504 280 173

25 . _- 2 21 10 1 34 464 256 139

26 6 21 6 3 36 480 264 165

27. 5 21 7 2 35 452 248 146

28 .. 9 24 3 2 38 396 212 128

29 27 3 3 36 444 240 153

30 6 24 2 3 35 408 220 143

31

Table 3.Conversion of index to staff required

Derivation of conversion method Example

Minutes of direct nursing care required on the Kard

Hours of direct nursing care required on the ward

Hours of direct nursing care provided per nurse per fl hour shift .

Total nursing staff needtd to meet all care requirements for the8 hour shift

Simplified expreision

INDEX x M 200 x 5 + 1,000 min.

INDEX x NI 200 x 560 60

16.67 hrs.

I' x 8 hrs. .42 x S = 3.36 hrs./nurse

INDEX x 111 16.67 hrs. req'd. 5

60 3.36 hrs./nurse nursesp 8

INDEX x 51 200 x 5 200 5 nursesP x 480 .42 x 480 40

"Ntirse" represents ativ nursing staff member, including registered nurses, certified nursing assistants, and orderlies, whoseprimary role is to providt. d,rcct patient care.Mis the average number of minutes of direct nursing care received by a tnittimal care patient on night shift (e.g., M = 5

minutes).Pk the percentage of nursing staff time spent on direct nursing care per shift (e.g., P = 42% = .42).Indexk the workload index for one shift (e.g., Index = 200).

48

Figure 2 represents a graph by day of the monthof the patient load indices given in table 1. Therange on the day shift extends from a high loadindex of 512 to a low of 244. The daily fluctua-don in workload is evident through the monthand particularly so when one notes the decreasein workload from the 7th to the Sth and the sud-den increase front the 14th to the 15th.

The fact that census may be most misleadingwhen used as the sole basis for judging workloadis obvious. The problems related to providing a

consistently high standard of patient care in theface of dramatic daily fluctuations in patients'nursing care needs are also obvious.

When several months of graphed informationare compared (see figure 2), trends in the pa-tient care load over a period of time become evi-dent. They provide an objective basis on which tocompare patient load from dad to day or monthto month on any given nursing unit or to comparebetween nursing units and to substantiate, wherecensus cannot, real changes in workload.

Workload Index as a Staffing Guide

Relating the workload index to the requirednumber of nursing staff was the next step in theproject and necessitated that the number of hoursof nursing care that the index represented 1.7.

known (8). This information was in fact easilyidentified as a result of the activity stueies. It wasobserved that, on the average, the minimal carepatients on nights received 5 minutes of care.' Sincethe ratios of categories to each other and betweenshifts are all expressed as multiples of the amountof care a minimal care patient receives on thenight shift, the number of hours of direct nursingcare required could he obtained by multiplyingthe calculated load index for any shift by 5/60.The activity studies also show that on the averagethe nursing staff spent 42 percent of their 8 hourday (3.36 hours) giving direct care to patients.

By dividing the number of hours of direct carerequired by the amount of direct care providedby one nursing staff member, an estimate of therepresent the upper and tower limits for each shiftthat may reasonably be expected by the budgetednumber of nursing staff required to provide

'Further studies revealed that 3 minutes of care was morelikely to be the average for minimal care patients on nights.

RESEARCH ON NURSE STAFFING iN HOSPITALS

the total direct and indirect care requirements ofthe nursing unit was obtained. Table 3 representsthe derivation of the conversion method of work-load index to required nursing staff.

Figure 3 graphs the daily workload index pershift over a 1-month period. The required numberof nursing staff as determined by the conversionmethod is also shown. Thus, for a specified work-load index, the number of staff required may beestimated. The broad horizontal bars in this graphnumber of nursing positions. Thus, as indicatedon the clay shift, the level of activity for the 10budgeted positions should not be expected to ex-tend beyond the range of that which 91/2 to 101/2people could manage. It should be noted that thehead nurse and assistant head nurse are not con-sidered part of the budgeted nursing staff, as onlythose staff members whose primary role is to pro-vide direct nursing care are included.

The utility of such graphs extends far beyondthe realm of academia. Indices that peak far abovethe bounds or baseline provide documented evi-dence that additional nursing staff are temporar-ily needed. When the indices of more than onenursing unit are peaked above the baseline, thegraphs become useful in aiding administration toset priorities in the allocation of the nursing staffavailable. Indices that dip below the baseline mayindicate that staff members should be transferredtemporarily to another nursing unit.

Finally, the need for permanent reallocation ofnursing staff becomes evident in the face of con-sistently high peaks above or dips below the base-line. The daily calculation of the workload indexcoupled with a flexible staffing policy reduces to alarge extent the great variability in workload towhich a given number of nursing staff must adjust.Closely matching the staffing on each shift to thenursing care needs of patients can do much topromote a consistently high standard of patientcare.

The Unit Assignment System

The development of a simple objective tool forclassifying patients according to their requirementsfor nursing care afforded the opportunity to in-vestigate more clearly the different nursing activi-ties required for each patient category. This in-vestigation resulted in a new concept of ward

MEASUREMENT OF PATIENT'S REQUIREMENTS FOR NURSING SERVICES

BEST COPY AVAILABLE

.

e ... .

141. I4! .,

____t._._I

)- -4 1---4----t---f f.-I

_ _ _ _ ___i .

..-

1 ; / I--*%:-.

14% 1I ; I

!

1 i

iI

44- ---f , ........._ all.!. 4

__1.--4,..., - . -------# 11-;;._ +---"" 1 iii, Ni ....

-----ii . ... SI,- -,-- t wo II 0

Imo I,

!.1I

4 44:1-i

17 , a

I ! , 1 ,4.

1 ,1 . ago.! aria t ,60. M

*/ -, ---Ilesrl- i - 8 4 --- --t----.- I/.

. ..

1.. N 1111 I. , .

1

. .5i

f 1 i Mai Nal 0_..... __i--, 1 :

I 1.

00. 40. di.i

al.

kr...it 1-1143.1"- (4.

gill .

-1--- --I

.i." ow de 1 I iggeggiiiiau. I 1 I

elar! -1 1- ------1--- I,,§4.1----j- 1"--114INI--1------i-----.---i---

eel

1111111112M14%I

; ibiTilliI

-....

8kg)

00

Or

00N

g 8 8 0

In

r-N

te

er

N

N

2

49

0[/©S

OH

OI

MO

O'

El M

C

,.

.

mm

..

.

Mir

MU

MM

NM

INN

MIN

IM

il1111

II

I Ol IM

IMN

IEM

I!M

IR

ilIM

Ilill1R11111=

NO

MM

EN

''

.

'CE

NW

OR

D:'4/4/ 4 .r./

glEIN

WA

SO

ZV

IS772r

OW

"A

wl

7A

zwv )24111"M

Ar 4E

' Mr r

AN

.PF

7 EM

"E

r4fA

.A

nwar A

mp' ,w

zwipA

rastwom

.APiii'A

mpE

zrrz40' m

arIhA

rP

A A

' .r Ar." z*E

..;11.&

f rA.F

.1'i'

0' AP

'A111111119110111

iniligiliW

illtrAiiiiN

aIN

NIN

IIIIS 13111111,11

11111111141'

1311111011 01A

EI IN

IMN

E V

ia SIN

EW

151111"N

IE

MI

MIE

NIM

ME

N11111111111111111111111W

111111W

IEN

!A

kir :EN

'OPA

W.08/4 M

V'

AN

EW

.J.'a' AN

'.W.A

" _Nun. m

v.AvA

r ,ramor

,Paf,m

9AIll

mIIIIIII1

11111U

SE

R.

EU

MS

' 11 ME

EN

'

WE

NN

EM

=as

.... ....

...

Ie. .111111111111511

NM

I IMO

NI

MEASUREMENT OF PATIENT'S REQUIREMENTS FOR NURSING SERVICES 51

organizationan organization that would focusthe attention of the health care team on thepatients' needs and care. The activity studieshad reveuled that the central nursing station, tin-der a modified team nursing method of assign-ment, was the principal focal point on the ward.In addition, many of the activities of the nursingstaff related more to the organization of the wardand the method of assignment used than to theneeds of the patients. In summary, the patientwas not the principal focus; the attentions of thehealth care team were not centered around hisneeds or his care.

The system entails decentralization of the wardstructure by dividing it into units of care classified

as intense, above average, average, and minimalcare units. A unit is defined as that number ofp; ants that can be effectively cared for by arep, ...erect nurse given adequate nursing assistancemid supply services. A standardized portable sup-ply and communication station is centered in eachunit and is stocked with patients' charts, drugs,trays, and a phone intercommunication system.Each unit is situated close to a group of patients,making it possible for the nurse's attention toremain on the patients.

Guidelines for Future Research

Our personal endeavors have frequently cen-tered around a vision of a system of patient clas-sification accompanied by some form of workloadindex applicable to the vast majority of generalor acute care hospitals. We have been influencedby their potential impact not only on the internaloperations of a hospital but also on the entirehealth care system. We have envisioned a classifi-cation system based on patients' needs accompan-ied by an index of the workload required to meetthose needs that %%odd be applicable to all hos-pitals.

It would appear that at the present time thedeveloped tools fall short of our aspirations. As

internal tools, particularly within the institutionswhere the basic research was conducted, they have

a great deal to offer in terms of understandingthe patient care process. Our experience has been,as previously mentioned, that patient classifica-tion systems alone provide a much more reliableaccount of the nursing care needs of patients thanthe usual patient census or diagnostic categoryestimate.

It is my impression, however, that in terms of

our present data base, their application as toolsby which comparisons between hospitals may be

made is limited. Differences in treatment modali-

ties, physical structure and design, educationalcommitments, and policies of out hospitals arejust a few of the reasons why difficulties are en-countered.

I would suggest that there are at least threereasons why our tools have not been able to ac-

GO

complish much beyond their present value ofinternal information sources. First, classificationtools that include the psychosocial needs of thepatients frequently do so at the expense of forcingsubjective opinions to enter into the assessments.This compounds the problems of reliability amongthose using the tools. Even when tools are designedto be completely objective assessments, rater re-liability should not be considered inherent. Second,the problems of validating the tools have not been

satisfactorily reconciled, and the attempts thathave been made are not necessarily transferableto other institutions. Finally, workload indiceshave primarily been developed through investi-gations of present operating systems and againare not necessarily transferable.

Study Model

As we approach a path for further research, Ibelieve our greatest commitment should be toapproach it within the realm of the health caresystem. This single fact alone should urge usto identify a model to permit a total systems ap-proach. Figure 4 is one such model. It identifiesnursing as an interdependent variable in the healthcare system.

The components are easily identified in termsof the system being studied. Patients may be

viewed in terms of care given, attitudes, expecta-tions, or post hospital experience, etc. Stafi

may be viewed in terms of volume, competence or

52BEST COPY AVAILABLE


Figure 4.Nursing: an interdependent variable in the health care system.

Patients

:ENO

Staff

InternalLinkages J

External I--I Linkages

Adapted -by Shirley M. Stinson, Procevsor, School of Nursing and Division of Health ServicesAdministration, University of Alberta, Edmonton, Alberta, September 1972. Revised April 1972(mimeographed) from a Paradigm, "System Properties of Educational Units," designed by SLOANR. WAYLAND, Professor of Sociology and Education, Teachers College, Columbia University, NewYork 1966.

61

MEASUREMENT OF PATIENT'S REQUIREMENTs FOR NURSING SERVICES 53

attitudes, etc. Internal linkage., may refer to a spe-cific department or nursing station and externallinkages may refer to a coordinated home careprogram, a communication system, or a visitingstudent. In the same way, all components may bedescribed and expanded to relate to the area ofscale.

This approach, although not new to health careevaluation, provides its with at least three distinctadvantages. First, it prevents us from omitting thepatient as a central focus. Second, it provides theframework for the difficult task of defining vari-ables. Third, it assists us not only to recognize butto study the relationship of interdependent vari-ables. Once the system is defined, the framework

for alterations becomes visible and the model canact as a manual simulation tool.

We arc still faced with very serious economicproblems. Undoubtedly, much of our future re-search will be directed toward achieving an op-erational solution to the effective matching ofhealth care resources to health care needs. Wemust not, however, neglect to conduct our researchwith equal attention to economics: the effectivematching of research resources to research needs.

I propose that if we conduct our studies withinthe realm of a model such as presented, we willhave conducted it in an efficient manner and itsapplication will have the potential for implemen-tation in a wider variety of settings.

62

54

MINIMAL CARE

RESEARCH ON NURSE. STAFFING IN noserrms

Appendix I

Definition of Categories

Patient is physically capable of caring for him-self but requires mininol support plus treatmentsor monitoring (B P., T.P.R., observation, etc.) bythe nurse q411 nl ess.

AVERAGE CARE

Patient requires an average amount of nursingcare and medical support, e.g..

past the acute stage of his disease or surgery,3-4 clays post-op cholecystectomy,diabetic for reassessment,conditions requiring extensive investigativeprocedures.

ABOVE AVERAGE CARE

Patient requires a greater than average amount

of nursing care, medical support, and use of spe-cial equipment, e.g.,

after acute phase of CVA (residual paralysis),advanced Parkinson's syndrome,radical mastectomy of cholecystectomy,2nd day post-op,a debilitated, dependent elderly ,erson.

INTENSE CARE

Patient requires very frequent to continuousintensive nursing care and close supervision bymedical personnel with support from technicalequipment, e.g.,

multiple sclerosis on a rocking bed,sevf e burns,1st clay retropubic with continuous irration,comatose patient,severe toxemia of pregnancy.

Appendix II

Guidelines to Assist in Selecting the Correct Determiners in the Classification System

PERSONAL CARE

Complete BathPatient is dependent upon the nurse for

his complete bath. It could be that the bedbath is given by the nurse or that the patientrequires continual coaching by the nurse duringthe bath, as in rehabilitation. It could also he atub bath or sling bath if the patient requirescontintring assistance or if more than one per-son is needed for assistance; e.g., burn bath.One must consider age, general condition, tubes,appliances, etc., which inhibit the patient's abil-ity to be independent.

Basin or Tub with AssistancePatient requires assistance with washing back

and legs. Patient needs bath equipment set up,partial assistance and periodic supervision and

63

coaching. Patient may only require preparationof the bathtub and assistance in and out of thebathtub.

Basin or TubPatient takes bath independently; back may

be washed by nurse.

NUTRITIONFed or N.P.O.

Fedtotal meal fed by nurse and continuoussupervision during the meal.

N.P.O. -- implies tube feedings, gastrostomyfeeding, etc. Exclude diapostit procedures.

Partial HelpMeal tray set up by nurse, followed by partial

assistance but not continuous supervision or fre-quent encouragement and teaching.

MEASUREMENT OF PATIENT'S REQUIREMENTS FOR NURSING SERVICES 55

OBSERVATiON

Note: One must recognize the difference betweenintensive medical therapy and the patient's needfor intense care when interpreting this component.This component shoidd be interpreted verybroadly.

It implies observation of the patient's ,,ate ofwell being and psychological support to patients,relatives, and friends. Include the patient's needfor explanation and teaching; e.g., new mothers,colostomy, diabetic. Consider procedural activi-ties in which nursing observation is the basiccomponent; e.g., vital signs, dressings, I.V. in-fusions, deep breathing and coughing, medica-tions, compresses, intake and output, care ofdrainage tubes.

Consider diagnostic tests; e.g., gastric analy-sis, lumbar puncture, Hollander test, 24-hoururine collection.

It could include a confused patient who is upin chair or up walking and requires surveillanceat regular intervals. The frequency of the obser-vation will guide you in selecting the appropri-ate determiner:

CONSTANT to QIHAbove Average or In-tense.

Q2H to Q4HAverageQ4H or lessMinimal

ACTIVITY

In Bed or Chair with Position and SupportBed patient who requires exercise, position.

ing, and support with pillows or sandbags atleast Q211; e.g., a helpless patient followingC.V.N. or an immediately post-operative pa-tient. Consider as well the patient who requires

the above while in bed but is also lifted out ofbed and positioned in a chair BID or morefrequently; e.g., quadriplegic, chronically de-bilitated dependent patient.

Bed Rest with B.R.P. or up with Assistance.Patient who requires bed rest but changes

his own position as necessary in bed. This pa-tient may require assistance in getting out ofbed as well as to and from the B.R. Patientsmay require assistance with activity because ofthe presence of Levine tubes, I.V.'s, catheters,etc., or because of their general condition.

Up and AboutPatient who does not require assistance with

ambulation. He is up as tolerated or independentwith the wheelchair. He has learned to managenonrestricting tubes such as catheters, T-tubes,or I.V.'s. A rehabilitated paraplegic could beconsidered in this group.

Note: The two following determiners are abovea..crage or intense care determiners because theyinclude psychological support and teaching as wellas the procedural activities involved.

IncontinencePatient who does not have voluntary control

over the excretion of bodily wastes; e.g., feces,urine fistula, or sinus drainage or patients withcolostomies, ileostomies or ureterostomies

Pre-opThis column is checked the morning of cur-

goy for all patients going to the OperatingRoom. Also, it includes any diagnostic proce-dures requiring extensive workup, preparationof the patient, and frequent observation andfollowup; e.g., cardiac catheterization, renal ar-teriogram, pneumogram.

64


(1)

References

MacDonell, J.A.K., and Murray, G.B. "AnIndex of Care." Medical Services Journal(Canada), Reprint XXI:8:499-517 Sep-tember 1965.

(2) MacDonell, J.A.K., et al. Timing Studiesof Nursing Care in Relation to Categoriesof Hospital Patients. Deer Lodge Hospital.Winnipeg, Manitoba, 1969.

Giovannetti, P., et al. The Reliability andValidity Testing of a Subjective PatientClassification System. Vancouver GeneralHospital, Vancouver, British Columbia,June 1970.

(4) Connor, Robert J. A Hospital InpatientClassification System. Johns Hopkins Uni-versity School of Engineering. UnpublishedDoctoral Dissertation, 1960.

Commission for Administrative Services inHospitals (California). Staff Utilization andControl Program. Nursing Service Orienta-tion ReportMedical/Surgical 1967, 32 pp.

(6) Holm lund, B.A. Nursing StudyPhase I.University Hospital, Hospital Systems StudyGroup, University of Saskatchewan, Sas-

katoon, Saskatchewan, September 1967.Sjoberg, K., and Bicknell, P. Patient Clas-sification Study. Hospital Systems StudyGroup, University of Saskatchewan, Saska-toon. Saskatchewan, September 1968.Bicknell, P. Staffing by Patient Care Work-load. Hospital Systems Study Group, Uni-versity of Saskatchewan, Saskatoon, Sakatch-ewan, December 1970.

(9) Abdcllah, Faye G. "Overview of NursingResearch 1955-1968, PARTS I, II, 111."

Nursing Research, 19 (JanuaryFebruary

(3)

(5)

(7)

(8)

65

1970) 6-17: 19 (MarchApril 1970) 151-162: 19 (MayJune 1970) 239-252.

(10) Blom, David 0. A Study of a PatientClassification System as a Measure of theUtilization of Selected Supportive Services.M.H.A. Thesis, University of Minnesota,Vancouver General Hospital, UnpublishedMarch 1972.

(11) Sjoberg, K. "Unit AssignmentA New Con-cept." Canadian Nurse, July 1968.

(12) Sjoberg, K., and Bicknell, P. NursingStudyPhase IIA Pilot Study to Imp le-ment and Evaluate the Unit AssignmentSystem. Hospital Systems Study Group, Un-iversity of Saskatchewan, Saskatoon, Saskatch-ewan, December 1969.

(13) . Nutting StudyPhase 111 TheAssessment of Unit Assignment in a Multi-Ward Setting. Hospital Systems StudyGroup, University of Saskatchewan, Saska-toon, Saskatchewan, August 1971.

(14) __. The Development of a ResearchTool to Evaluate and Compare the Stand-ard of Patient Care. Hospital Systems StudyGroup, University of Saskatchewan, Saska-toon, Saskatchewan, Unpublished April1971.

(15) Sjoberg, K.; Heiren, E.; and Jackson, M.R."Unit Assignment, A Patient Centered Sys-tem." Nursing Clinics of North America,6:2 June 1971.

(16) Stimson, R.H., and Stimson, D.H. "Opera-tions Research and the Nurse Staffing Prob-lem." Hospital Administration, 17:61-69Winter 1972.

DISCUSSION OF

MRS. GIOVANNETTI'S PAPER

Discussion Leader

Dr. Harvey WolfeProfessor, School of Engineering

University of PittsburghPittsburgh, Pennsylvania

Dr. Wolfe

I have been trying to figure out a 1,ay to or-ganize my comments and I have come to the con-clusion that there is no possible way to do it.So I'll just make. some general comments and crit-icisms relative to the staffing situation as discussed.

First, I disagree that this kind of classificationmethodology is not transferable. I was not at allsurprised to find that three categories were finallydecided upon. i have seen this happen over andover again. The VA, in investigating its four cate-gories, found the last two classifications indiscern-ible. I think four classifications were found in onesituation because many hospitals have intensivecare units and we, to some extent, are combiningprogressive patient care and controlled variablestaffing to take care of the four categories thatprobably do exist rather than the three.

I am not sure why we are still doing studies onclassification. It appears that no one wiFties tobelieve anybody else's results without some kindof reevaluation for their own institution. I don'treally think hospitals are as different as we claimthey are. They have many commonalities andmany similarities that can be specified if we takea very close look.

I think this lack of acceptance results from thefact that we nursing people are not willing toaccept the validity related to patient classification,nor have we done any studies to really validate

57

the effect of classification on workload. That'swhat we really have to validate, not whether theclassifications are valid and reliable or whateverelse you like but whether the whole concept ofattaching classification to workload is valid.

We can assume that nurses are humans and havehuman capacities, and we have all observed thatin a nursing unit the ability of the nurses to ac-commodate to the workload is terrific. They anttake shortcuts. They can skip steps. They do whatthey must to perfurm the minimum tasks, and ifthey can do more tasks, fine, but if they can't theyat least get the minimum tasks done.

Given the fact that nurses adapt to heavier orlower workloaus, what we need to do is to evaluatewhether the minutes of care given are a real vali-dation of workload. Real validation requires ex-tensive study of how the number of tasks and thequality, and by "quality" I mean comprehensive-ness of tasks performed, for the whole nursingunit varies as these workload indices vary. Thatis what we have to demonstrate. Patient satisfac-tion and dissatisfaction are not the only measures,but they are certainly measures we can use to de-termine whether any change in these factor:, takesplace relative to workload.

My second comment is that the expectation wehave of staffing studies is too great. We typicallyget estimates for times and so forth that are much

58

more precise than we have any need for. We onlyhave the ability to manipulate staff in terms ofwhole people or sometimes half peoplein rarecases, 2 hours' worth of personand the fact thata workload differs by a couple of points withinrelatively wide ranges doesn't make any differenceat all.

What we'd like to do is to have a classificationscheme that is reliable within these limits, andthat's why I think that tie subjective classificationscheme just discussed is perfectly reliable. If it isoff by one category, that wouldn't bother me. Iwould consider the reliability criterion to bewhether it would change the number of personnelyou need, not whether it correlates with itself.


If blue dots on the forehead were enough to dis-criminate various patients then that's all we'dneed to use, nothing more complex than that.

Finally, classification is only a guide and itshould he used as such. Nothing more should begleaned from it than it is capable of giving. Oneof the things we haven't discussed is the use ofthe case mix, and this is an important considera-tion in any staffing studycase mix and staff mix.

It doesn't always make sense to be perfectlyefficient, and we only have to be efficient withinrelatively wide spectrums. We need a proper bal-ance of the specialization and the economics as-sociated with it.

Nursing Staffing Patterns and HospitalUnit Design: An Experimental Analysis

Dr. E. Gartly JacoP1 of essor, Department of Sociology

University of CaliforniaRiverside, California

Intvoduction

This is a report of part of a larger study (1)conducted to evaluate potential differences in pa-tient care in two basic types of hospital nursingunit design; namely, radial and angular shapedunits. Very little research has been conducted, par-ticularly columned experimental studies, on nursestaffing patterns. Espec:aily neglected has been theeffect of physical design and spatial geometry ofhospital nursing units on patient care (2). Whileseveral studies have investigated the potential im-pact on nursing care of a few differently shapednursing units their relationship to specific nursestaffing patterns has been ignored (3,1,5).

Man's use of space has long been a topic ofscientific intezest, particularly in the social andbehavioral sciences (6,7). These and more recentstudies and conceptualizations have provided a

theoretical as well as a conceptual foundation uponwhich furthe- and more sophisticated research onthe potential linkages between human uehavior.'nd physical settings can be based, particularly

59

the writings of Hall (8), Sommer (9), Barker (10),and Osmond (11). Conceptualizations of macro-spare deal with the larger physical environmentin such terms as "territory," "community," "local-ity," "natural area," "ecological area," and havetraditionally been essential aspects of the sullfield ofhuman ecology. Behavioral scientists have also beeninterested in human ritilizatinn of m.icrospae,-, in-cluding perceptions of and reactions to such smaller,more immediate physical surroundings of in-dividuals, presenting such useful concepts as Hall's"proxemics," (12) Osmond's "sociopetal space"and "sociofugal space," (11) and Barker's "un-dermanned" and "overmanned" behavinral set-

tings (10).Thus a rich and varied literature dealing with

many facets of t physical environment, usefulfor a better unuerstanding of the relat:on ofphysical design of the hospital lo nursing andpatient care, prevails in the social and behavioralsciences. Despite their availability, little use has

60

been made of these theoretical and empirical worksin extant studies of hospital design and patientcare and particularly with respect to nurse staffingpatterns. Indeed, the author's own research onthese topics did not benefit from these theoreticalconceptualizations until after such studies wereunder way.

The Research Project

Nurse staffing patterns were a major factor in alarge research project on nursing and patient carein circular and angular hospital nursing units,since the types of nursing personnel could affectpatient care on a nursing unit of a given shape.The major independent or "causal" variables ofthe study were the radial and angular shapednursing units. The dependent or outcome vari-ables were major types of nursing care, level andamount of nursing care, and nurse utilization ofthe unit. Nurse staffing was a key interveningvariable along vith occupancy levels and certaincharacteristics of the patients in the unit. Theseinclude primary diagnosis, room accommodation,and socioeconomic and demographic traits.

The overall crudy was conducted in a voluntary,acute. nonp--, 350-bed general hospital in St.Paul, Minnesota. In the early 1950's this hospitalconstrue:cc! a number of circular or radial nursingunits as well as the traditional angular or L-shapedunits. It had and still has a progressive patientcare program, including intensive care, intermedi-ate care, and self-care or minimal care. The pres-ence of both radial and angular units afforded

. an opportunity to do a comparative analysis ofpatient care given on the two types or shapesof units. Both were almost new when the studywas performed and were on the same floor of thehospital.

The nursing units for the study comprised 22.beds: two private or singe rooms, six semiprivateor 2-bed rooms, and two 4-bed wards. The radialunit had a total area of 5,180 square feet withcorridors of 900 square feet with 133 running feet.Its nurses' station in the center of the unit was18 feet from the farthest patient bed and com-prised 8.2 pe.rent of the total area The angularunit, resembling an L shape with its nurses' stationlocated at the vortex of corridors. totaled 5,910square feet. It had corridors of 1,258 square feet


with 145 running feet. The nurses' station, 70 feetFrom the farthest bed, totaled 5.3 pe;cent of totalarea (see figure 1). Both units were of new con-struction, similar in interior decor, air conditioned,and in close proximity to each other on the thirdfloor of the facility.

Patient care at the three levels of intensive, in-termediate, and self-care were studied in the totalproject. However, only intermediate care will bereported here, since more variations in nurse staff -ing patterns were observed at this level of carethan at other levels. Also it is the more typicallevel of patient care provided in such hospitals.An effort was deliberately made to keep the kindsof patients reasonably similar in both study unitsduring the course of the project. Admission to bothunits was restricted to general medical and surgicalpatients needing intermediate care, who were notemployees or relatives of hospital employees, hadnever been admitted to a radial unit, and were notemergencies.

To control the intake of patient characteristicsduring the study an unusual admitting procedurewas set up in the Admitting Office. A pair of dicewas used to select patients on a random basis. Thiswas after all the previously mentioned criteriahad been applied. (We then, you might say, hada multitude of patient characteristics, primarydiagnosis being one. We were concerned with theage, sex, and social class of the patient, and wetried to hold the number of factors reasonablyconstant by randomizing between thci two unitsduring the study session. So we made a deliberateeffort to make as similar as possible 'the patientpopulations being given nursing care during theresearch time.)

Twelve study situations were set up in the radialand angular study units. Occupancy levels werealtered from high (90 percent filled or more) tolow (50 percent filled or less). Nurse staffing pat-terns were varied quantitatively from high throughaverage to low. They were varied qualitativelyby varying the ratios of registered nurses (RN),licensed practical nurses (LPN), and nurse aides(NA) to total staff. Our rationale for the originalstaffing mix was based on consultation with theNursing Department of the hospital.

Each study situation was conducted for 4 con-secutive weeks during a 5-day workweek (Mondaythrough Friday), excluding holidays. Work sam-

Fig

ure

1.R

adia

l and

ang

ular

hos

pita

luni

ts.

3217

2BED

ROOM

3216

RA

DIA

LA

tS

TU

DY

UN

IT

W^)

rw

,, ROOM c,

c%

cl.

imsTASES41)

q---

i')

;,,,...0.24050EnD

w...

..s

ul 2

pt.:

3300

use 0

^Iv)

liPri

04

..1%

.342

,,,.3

3 ROOM

BE

ST C

OPY

AV

AIL

AB

LE

SUN

DE

CK

3BED

SO

ON

3303

JAN EXAM

o.C

ON

Fcl C

'13

011

w

DA

Y

RO

OM

AN

GU

LAR

s

ifS

TU

DY

UN

IT

3200

11el

f"c

3206

-74 BEO

ROOm

3203

..m

46

2 BED

3001

3:::

N.1

)..ROOM

3222

4BED RM

12BED

UTILITY

3016

3522

1352

4,3

5 2

6

3207

LOBBY

15

3014

PRI

Ed

13=

1MSG

3520

STA

3501

0502

135

00co

la35

18P

RI

PR

IT

3503

-TuTIL

3504

3506

PRI

. .E.

PRI

T'

PRI

3516

3514

3508

P-35

12PRI

TH

IRD

FLO

OR

110.

.4SE

Ake

4BED

RM

4 BED RM

3528

3530

3540

353

8PRI

PRI

3534

PRI

In ELEV

Sain

t Luk

es H

ospi

tal

Stat

ion

3500

rn

62 BEST COPY AVAILABLE

piing methods were used to observe the total nurs-ing staff for 24 hours each of these days. At ran-domly assigned subareas within the units, they wereobserved at intervals of about 10 to 15 minutes eachhour by observors trained in work sampling techniques and in nursing tasks and functions. Theaverage number of observations was about 9,000for each study situation. The total observatetsfor the 12 situations came to nearly 205,000.

Certain variables were recorded by an additionalobservor on the unit during the day shift only.A 1 week resettling period was set up before eachnew study situation to help blur any experimental"halo" effects. An effort was made to use identicalnursing personnel in both types of units for allstudy situations, so that each nurse would be herown experimental control for individual differ-ences among the nursing staff. This was not alwaysachieved, because of staff turnover. No attempt was

RESEARCH ON NURSE STAFFING /N HOSPITAUS

made to determine what might be the best or mostappropriate staffing pattern for these hospital unitsor patients since such would comprise an entirelyseparate inquiry in. itself. Rather, a major purposewas to determine whether the basic physical designof the nursing unit had any effect on nursing care.The effects of different nurse staffing patterns, de-liberately varied both quantitatively and quali-tatively, were also measured, but other variablessuch as occupancy levels and divergent patientcharacteristics were held reasonably constant orrandomized.

This project was initiated because the hospitaladministration wanted an evaluation of their radialunits. They were quite new. Controlled experi-mental studies were, if not limited, not readilyavailable. So the study was set up to evaluate whatwas going on in these new units,

The nurse staffing patterns were established as follows:

a. Quantitative Shift RN LPN Nurse Aide Ward Clerk

High nurse staffing 7 3 4 (1 Head N.) 2 2

(above average) 3 11 2 2 2

11-7 1 1 1

Average nurse 7 3 3 (1 Head N.) 1 1 1

staffing 3 11 1 1 2

11-7 1 1

Low nurse staffing 7 3 2 (1 Head N.) 1 1 1

(below average) 3 11 1 2

11-7 1/2 1/21/2

b. Qualitative 7 3 5 (1 Head N.) 0 0 1

High RN staffing 3 11 3

11-7 2

High LPN staffing 7 3 1 Head Nurse 5 0 1

3 11 1 Head Nurse 3

11-7 1/2 2

High nurse aide 7 3 1 Head Nurse 0 5 1

staffing 3 11 1 Head Nurse 3

11-7 1/2 2

We felt that to do the job right we had tocontrol a number of factors that everyone in theprofession is aware could affect or contaminate theoutcome of the study unless they were controlled

in some manner experimentally during our study.So an effort was made in this direction. We did notagree to conduct the study in this particularfacility until everyone agreed to accept any nega-

NURSING STAFFING PATTERNS AND HOSPITAL. UNIT DESIGN: AN EXPERIMENTAL ANALYSIS 63

tive evidence that might be generated by the study,(So many inhouse studies are done to rationalizewhat the administration already wants to do or hasa tendency to be nervous about accepting evi-dence against.) One of the hospital trustees said,"Well, if we make a mistake in these units, thewhole field should know about it so they won'trepeat the same mistake."

The .same went for the nursing staff. This wasdifficult because we wanted to set up a numberof conditions that the research design required tobe sustained over a suffir!ent period of time toconduct a really valid project from the researchstandpoint. So a second factor essential to thistype of study is a willingness on the part of thestaff to conform reasonably well to the dictatesof the research design.This is the more difficultproblem, particularly when you're experimentingwith the number of variables we were trying todo in our study.

We began with what was an approximation ofthe existing average staffing pattern already goingon in the hospital and used that as our baselineto increase it quantitatively or decrease it in justsheer numbers of nursing personnel and then varyit qualitatively between the categories of nursingpersonnel. When you talk about understaffing andoverstaffing you should say whom youdare under-staffing and whom you are overstaffing. Is it theregistered nurses, the LPN's, the nurse aides, orwho? This turned out to be a very importantoutcome variable, the quantitative variation versusqualitative, the mix of nursing category personnel.

So we proceeded to set up our project in whichwe would have observers on the units routinelymaking observations of the activities of the nursingstaff for a 5-day workweek, Monday through Fri-day, and observe them on all three shifts, 24 hoursa day, in the unit. We did a good deal of trainingof the observers. We used LPN's primarily forobservers because one thing we found out quicklyis that RN's don't like to be observed by otherRN's, but they are more tolerant of the lessertrained staff looking at them. Other studies havehad the same problem.

After the training session, we checked out thereliability of coding their observations by a num-ber of coders and observers. We obtained veryhigh check/recheck reliability of about 95 per-cent because the recordings were quite detailed on

what the nurse was doing. (When they were answer-ing the telephone we wanted to know whetherthey were talking to the doctor or a relative of thepatient or whom they were talking to, and thingsof that sort helped make the eventual classifica-tion of their activities a little more, in fact ex-ceedingly, reliable.)

We also wanted to sustain a nursing staffingsituation for a period of time should the nursesput on a show for the observer, (which they cando, you know). If this occurred,' we would extendthe period of observation after their show wasover. This was somewhat of a problem when wegot into some of the more difficult staffing patternsituations.

Another thing we wanted to check out was thesignificance of the activities of the nursing staffon the night shiftif such were equal to, if notmore important than, what went on in the dayshift in terms of the patients' morale, reactions,and satisfactions for th it nursing care. So muchresearch has been oriented toward the day shiftwhen actually other times may be equally significant. Just because that's when the doctor is usuallythere doesn't mean that other things aren't hap-pening at other times of the day. So we wantedto check out these possibilities.

Then, too, we were quite concerned about thestatistical validity of our observations, not onlyin terms of reliability and validity but also to de-termine whether we really were getting a goodpicture of what was going on in these units. So wedivided the unit into 10 subareas: the nurses' sta-tion, medication area, patients' rooms, the corridor,utility room (they had a tub and shower room tosupplement the bathrooms in each of the patients'rooms) and all the different parts of the unit.We random! ted these areas in which the observa-tions would be conducted by the observers so thatthe nursing staff hardly knew when they werebeing observed. Although the observers may bein the unit, they may not really be recording atthat particular place in the unit, and this methodenabled us to get a random sample.

We found some feeling of resentment on the partof the aides and the LPN's toward the RN's andthe head or charge nurse in the radial unit whenthe latter supervised their activities too closely.If you were to see this unit you could see why.You can be in a patient's room and you can see

into all the other patients' rooms, too. so if thehead nurse would be attending a patient in oneroom she could still watch the other personnelon the same unit. She can't do that in theangular unit. Such visual contact or visibility ofactivity on the radial unit did not exist on theangular unit. and this IHIHPd out 10 be a keyvariable in the whole study that was related 10 theshape of the unit.

Also, because of differential preferences by thenursing staff for the different units in the hospital,we tried to equalize these preferences by choosingthe nursing personnel equally between those whopreferred the radial to the angular unit and thosewho preferred the angular to the radial unit.So we tried to get an equal mix so that any biasthat might affect how they attend patients andtheir activities would be equalized.

Also, we made a very strong effort to use thesame nursing personnel all throughout the 2 -yearproject. Instead of trying to get so-called matchedgroups of nurses, one in one type of unit, anotherin another type of unit, each nurse was her ownexperimental control. They wound run throughthe series of study situations in the radial unit,then he transferred into the control angular unitand repeat the staffing pattern and occupancysituations. We only had a certain amount of turn-over in the nurses over a 2-year period. About 90percent of the nursing personnel were the samethroughout the project in our studies, which wefelt was important.

Because we knew the study would last for a

long period of time, we were quite sensitive tothe cyclicity of the occupancy level of the hospital.Most hospitals do follow peak and trough periodsof occupancy. Summer time in this hospital wasits low patient occupancy period and was vacationtime for the staff. We were sensitive to the vacationneeds of the personnel of the staff so we set upour low staffing patterns and our low occupancylevels when the hospital would be ordinarily atlow occupancy and nursing personnel would likelybe on vacation. I think such sensitivity is veryoften important to the success or failure of suchprojects in hospitals of this kind.

When you study staffing patterns, you have totake into account the role and function of thefacility. the kinds of patients likely to he admittedby the medical taff, whether or not there is


emergency service, whether or not it has an out -patient service, whethe it takes a broader rangeof patients or is confined to a few. In effect, youmust consider the inflow of patients the nursingstaff must care for.

We were interested in some of the more tradi-tional outcome factors such as the potential re-lationship of nurse staffing patterns and oc-

cupancy levels to length of patient stay. We usedDr. Aydelotte's scales to see whether patient wel-fare would be in any' way affected. We checkedthe possible effects of basic demographic factors,medical diagnoses, amid a number of basic medicalbehavioral factors on the dependent variables.

As it turned out, the length of stay did notvary significantly among any of the so-called studysituations. The patient welfare measures and anumber of other variables did not affect our results.

Prior to discharge, each patient was interviewedprivately off the floor regarding the nursing andmedical care he had received, how he liked ordisliked his room, the layout of the unit and so on.We soon discovered a bias operating in favor ofthe radial unit. The radial unit was rather small,and we could quickly see a bond develop betweenthe nursing personnel and the patients that wedid not see at the same intensity on the angularunit.

Fortunately the interviewer, a social worker,was very skilled at interviewing and workedthrough the fact that many patients would simplynot be critical of the nurses, more on the radialunit than on the angular unit, at the start ofthe interview. finally, we did get some criticismsafter piercing through the facade of being a "niceguy who didn't want to embarrass the nurses," andso forth.

But this was a reflection of the close relation-ship that developed in the circular unit. This wasparticularly true in those study situations of highoccupancy and low staffing.

Another fact is that the patients on the radialunit could identify the different nurses by the capsand uniforms of personnel coming in and out ofthe unit, which they couldn't do on the angularunit. There were other indices of how patientsempathised with the nurses more on the radialthan on the angular unit. It was not the same forthe doctors, I might add.

One of our main purposes in the project was to

14,i`if 3

BEST COPY AVAILABLE

NURSING STAFFING PATTERNS AND HosPITAL UNIT nEsioN: AN EXPERIMENTAL. ANALYSIS

replicate many factors in a prior study conductedby Sturdavant and her associates (1) at Rochester,Minnesota. We wanted to see what factors we couldreplicate so we used at our facility a number ofcriteria similar to those in the Rochester study.

This is one thing needed in nursing care re-searchmore replication in different kinds of facili-ties with different personnel to see if we derivethe same or divergent results. So we used essentiallythe criteria for patient and nursing care that wereused at Rochester, with some variations.

For example, we had six basic categories of majortypes of nursing care essentially defined as directbedside care, which was basically composed oftreatments, supportive care, ambulation and soon and communication directly with patients. Thenwe had indirect care, general assistance, standby,

time out, and travel for patient purposes. Thecriteria used were similar to the Rochester meas-ures to permit replication of 'these factors.

.We took these criteria and categorized them intowhat we call medical, social, and physical care.

65

The categories of the direct care activity involv-ing treatments, where the activity was orientedtoward the medical problem of the patient, weretermed medical nursing care. Care that was gearedtoward physical comfort like adjusting the pillow,making the patient more comfortable, attendingto physical needs of patients not related to theirillness was called physical (bodily) care. The thirdwas social, which involved communicating, nursestalking with the patient or observing the patientdirectly in case there was a need in terms of thesocial category.

We had another category termed patient-involvedcare. That is, such care was defined essent:ally asa combination of direct care, indirect care, and

general assistanceall of which had directly to dowith involving the patient as opposed to suchfactors as standby and time out. So we had ratiosof time observed in these three activities comparedto other activities, and this varied enormously.These arc just some of the measures we developedin the course of the project.

Major Results

Type of Nursing Care

Since a major purpose of the overall originalstudy was to replicate many facets of Sturdavant'sRochester study (13), similar criteria and measure-ments of many variables were used in the presentinquiry. Six major categories of nursing care weredefined: direct (bedside) care, indirect care, gen-eral assistance, standby, time out, and travel. Forthe total of 12 previously mentioned study situa-tions it was found that major types of nursingcare differed significantly 1 tween the nursingunits, as shown in table 1. (Percentage differencesgreater than 1 percent were statistically significant,because of the large size of the samples involved.)

It was found also that the highest percentageof nursing care observed was in direct, bedsidecare (30 percent), followed in order by standby,indirect care, general assistance, time out, andlastly in travel, when nurse staffing pattern, oc-cupancy level and other characteristics were heldconstant statistically. On 204,793 observations onboth nursing units for somewhat over a year each,approximately identical nursing staffs provided

74

more indiiect care, general assistance, and tookmore time out, andeririiided less direct care,

standby, and traveron the radial units than on theangular unit.

An analysis of the six specific nurse staffing pat-terns designed for this study presented a somewhatdifferent picture when shape of the unit and oc-cupancy levels were pooled, as presented in table 2.

For the quantitative staffing patterns, the highstaffing pattern exhibited more time spent in

Table 1.Major type of nursing care in an acute,general hospital

Type of CareNumber of

observations Percent Rank

Direct care 61,338 30 1st

Indirect care 46,370 23 3rd

General assistance 19,567 9 4th

Standby 46,601 24 2nd

Time out 17,405 8 5th

Travel 11,512 6 6th

Total 204,793 100

66 BEST COPY AVAILABLE RESEARCH ON NURSE STAFFING IN HOSPITALS

Table 2.-Major type of nursing care, by six nurse staffing patterns

Nurse staffiing pattern

Type of care High Average low High RN High LPN High NA

Number Per.rent

Number Per-rent

Number ,"errent

Number Per-cent

Number Per.cent

Number Per.cent

Direct care 12,261 26 10,821 32 14.456 32 7,988 35 10,632 31 11,180 28Indirect care 9,849 21 7.832 23 t.161 23 6,170 27 9,240 27 7,118 18

Generalauistance 4,743 10 2,801 9 2,638 10 2,631 11 3,627 10 3.127 8Standby 13,088 27 7.075 21 5,832 22 3.686 16 6,424 19 12,496 32Time out 4,764 10 2,980 9 1,999 8 1,582 7 2,770 8 3,310 8Travel __ _ _ 2,824 6 2,114 6 1,335 5 961 4 1,879 5 2,399 6

Total _ . 47,529 100 33,623 100 26,421 100 23,018 100 34,572 100 39,630 100

Table 3.-Major type of nursing care, by nurse staffing patterns and by radial and angular units,percentages only

Major type of care (percent of observations only)

Staffing Direct Indirect Gen. Asst. Standby Time out Travelpattern

Radial Angu Radiallar

Angular

Radial Angular

Radial Angular

Radial Angular

Radial Angular

High 26 26 22 20 13 7 24 31 10 10 5 6Average 31 34 25 21 9 8 20 22 9 9 6 6Low 33 32 24 22 10 10 19 25 8 7 6 4High RN . 35 34 28 25 13 10 14 18 7 7 3 6High LPN ___ ____ _ 29 33 27 27 11 10 20 17 8 8 5 5

High NA 27 29 19 17 8 8 31 32 9 8 6 6

standby followed in order by direct care, indirectcare, equal amounts of general assistance and timeout, and last in travel. There was an increase instandby and time out and a decrease in directand indirect care compared to the percentages oftime observed giving these types of care for thetotal staff.

Average staffing showed an increase in directcare and a drop in standby compared to the totalstaff. Low staffing also exhibited an increase indirect care and less standby than for the averagestaffing pattern. For qualitative patterns, the highRN pattern showed an increase in direct care, in-direct care, and general assistance but a decreasein standby and travel. The high LPN patternexhibited an increase in indirect care and a de-crease in standby, and the high NA pattern indicated an increase in standby and a decrease indirect and indirect care compared to the totalstaff percentages for such care.

Thus, major types of care seemed to vary morefor the qualitative than for the quantitative nurse

staffing patterns, although type of care was affectedby all of the staffing patterns differently when oc-cupancy level and shape of the unit were pooledor held constant.

Were these major types of care for each of thepatterns affected by the shape of the unit? Table3 shows that type of care provided on the twounits did indeed differ for the six nurse staffingpatterns. For the quantitative patterns, highstaffing increased time spent on general assistanceand decreased standby on the radial unit, but theopposite occurred on the angular unit. Averagestaffing increased time on indirect care on the cir-cular unit and decreased such care while increas-ing direct care on the angular unit. Low staffingshowed opposite results primarily in standby, drop-ping on the radial unit and increasing on the an-gular unit.

For the qualitative patterns, increased time oc-curred for general assistance with a drop in standbyfor the high RN pattern on the circular unit,while indirect care decreased and standby and

NURSING STAFFING PATTERNS AND HOSPITAL. UNIT DESIGN: AN EXPERIMENTAL. ANALYSIS

travel increased on the angular unit. Direct caredropped on the radial unit, but increased whilestandby decreased on the angular unit for thehigh LPN pattern. Apparently, the high NAstaffing pattern was not significantly affected byshape of unit since type of care on both resembledclosely that of this staffing pattern totally, althoughsmall differences occurred for type of care betweenthe two units.

Thus we find that both nurse staffing patternsand the physical design of the hospital unit af-fected the major type of nursing care provided togeneral medical and surgical patients. Adding ad-ditional nurses to the staffing pattern when shapeof the unit was ignored led to more standby andtime out and less direct patient care.

This resembles similar findings by Aytlelotte andher associates (3) when compared to the averagestaffing pattern. The low quantitative staffing pat-tern devoted similar percentages of time to directand indirect care as for the average staffing level.But the qualitative staffing patterns indicated moreeffective differences in type of patient care, particu-larly for the high RN pattern where significantincreases in direct and indirect care and generalassistance occurred compared to the lesser qualitystaffing ratios.

The physical shape of the nursing unit seem-ingly affected the types of care given for most ofthese staffing patterns. Increasing the quantity ofnursing staff had differing results on the two units.For the radial unit an increase in indirect careand general assistance occurred more than for theangular shaped unit, with standby differences dif-ficult to assess. Varying unit shape led to opposingchanges in type of care when changes did takeplace for identical staffing patterns. Indirect care,general assistance, and standby were most affectedby the radial shape, but direct and indirect care,general assistance, and standby were most affectedby the angular unit although in oppofite ways.

Even so, the radial unit seemingly 1. ad more ef-fect on type of nursing care provided by qualita-tively different that by quantitatively divergentstaffing patterns. The high RN pattern gave moreindirect care and general assistance with less traveland standby on the circular than on the angularunit. The high LPN staffing pattern gave moredirect care on the angular than the radial unit.Shape of the unit, therefore, had more effect on

67

the high levels of quality of nursing care (RN) andthe least influence on the lowest level of care (NA).This was further confirmed when the major typesof care for specific nursing personnel were analyzed.

Ilead and charge nurses on the radial unit pro-vided more indirect care and general assistanceand less standby than on the angular unit. RN'sgave more general assistance and less direct careand travel on the radial than on the angular unit.LPN's on the radial unit gave more indirect careand general assistance and less standby than onthe angular unit. NA's traveled more and gaveless direct care on the radial than on the angularunit.

In studying the effect of nurse staffing patternsand design of the nursing units, the need to con-trol for differences in occupancy levels and stagesof illness is obvious. A separate analysis of theresults was made in terms of high and low levelsof occupancy. of the study units. This revealed sig-nificant differences in type of care provided forthe staffing patterns and shape of the units, as

differential demands for patient care were af-fected by these patient levels. Consequently, ourefforts to control for such divergences when analyz-ing the effects of staffing and shape of the unit wereempirically justified.

The type of care most affected was direct care.For example, by and large there was more directcare given in the angular unit than in the radialunit, although it was "easier" to givc it in theradial unit. Also to our surprise, there was lesssocial care in the radial than in the angular unit;that is, less verbal communication between pa-tients and nurses on this type of unit. .

So we had to come up with some answers as towhy this happened. Did the shape of the unithave anything to do with this or what was goingon between the staff and patients or what? Thecircle is architecturally the most rigid, inflexibleshape, and the only way to expand a circularunit is to add floors vertically like stacking pan-cakes. It cannot expand laterally as can an angularunit or other shapes. Yet, on most nurse staffingpatterns we found far more flexibility in whatcare is given by whom in the circular than inthe angular unit. I think that is why the reg-istered nurses preferred it, because they were ableto exercise more options in giving care to patientsby the vet), fact that they could see the patient

68

from the station. They could sit in the station (sitdown because the counter is low), and they couldlook into every room and see whether the patientneeded care or not. You can't do that in an angularunit because you don't have that visibility. In theangular unit, when a call light came on in thestation, the head nurse might delegate an aide, to"Go see what Mrs. Jones wants." But in theradial unit she could look up and see whethershe was needed, She could exercise options in thedelivery of nursing care that she could not do ina differently shaped unit.

So I think, because of this, there was a pref-erence for the radial unit by the RN's. Theycould exercise more professional nursing judg-ment in providing care. There was a lot of non-verbal communication going on between patientsand nursing staff waving and smiling at each

other hut they didn't go into patients' rooms andvisit as much. So there was a kind of vicariouscare going on between the staff and the patientsin the circular unit.

Now, we got into some very difficult study situa-tions where we really overloaded the staff. Wewere almost risking the well-being of the p;W:.nts,but we would terminate the study situation beforeactually jeopardizing the welfare of any patients.During these overloaded situations, many patients,when we interviewed them, said they never touchedthe call light in a 2- or 3-week stay, and we askedthem why. "Well," the patients would say, "thenurses were too busy. I didn't want to botherthem." They contr.l see the nurses in the unitrunning around helter skelter sometimes. and theydidn't want to bother them. They felt that, "If Ineed htlp, they'll come," so they had this confidencein the nursing staff in the radial unit.

We didn't find that in the angular unit. Theywould sometimes get lonely and just want to see ifanyone were still out there because they hadn'tseen them in a while. So they pushed the call light.

When we examined the factor of nurse initiatedvisits versus patient initiated visits, we found thatabout 82 percent of the trips to the patient's roomin the radial units was initiated by the nursescompared to about 79 percent on the angularunit. Well, you may think this doesn't make sensesince sometimes they may not need to make atrip to the rooms because they can see in. Buta related finding was that the average length of


time spent in the patient's room per nurse visitwas far less on the radial than on the angular unit.

Why? What happens? You can do a tempera-ture check and you pick up the tray and maybeyou do some comfort things in one trip in theangular unit, and those tasks can accumulate intoseveral functions per visit. A lot of them involvetalking and socializing with the patient, whiChyou don't see in the radial unit because thenurses don't feel there is as much need for it inthe radial unit as in the angular unit. This is partof the relationship between patient and nurse.

We could almost predict the day patients in thecircular unit would be discharged. One of thethings that happens when patients get out of in-tensive care in this type of unit is that as they aitconvalescing, feeling better after the crisis has

passed, they start closing the cubicle curtain be-cause they don't want ,to be bothered. They wantprivacy and want to read or watch TV. Theyleave their room, go to the lounge or the coffeeshop, and use other parts of the hospital morewhile convalescing on the radial unit than theydo the angular unit. The "curtain pulling syn-drome" was directly related to convalescence, andwe recorded the halfway pull, all the way pull,and finally when it started to be closed all thetime we knew the doctor was going to discharge thepatient because he was ready to leave. This oc-curred only on the circular shaped unit.

A number of other events occurred that youdon't ordinarily take for granted. What goeson in the nighttime in different units? This iswhere we had some problems with the radial unitbecause of the architecture. One patient, becausethe drape next to the outside window wouldn'tquite close, complained about being bothered bynoises. We found there was street light by herwindow, and the slightly opened drapes alloweda sliver of light to come into he;' room all nightlong,

We found these things out when we interviewedthe patients. It is amazing what you can learnabout hospitals from patients, about a lot of littlethings. The night light in the nurses' station both-ered one room but it didn't bother any of theother rooms on the radial unit because of theshape and room location. Also, the drug cabinetwas a problem in the radial unit because accessto the radial unit nurses' station was easier.

BEST COPY AVAILABLE

NURSING STAFFING PATTERNS AND HOSPITAL. UNIT DESIGN: AN EXPERIMENTAL. ANAI.YSIS

Now one other thing about our discovery wasthat a number of physicians, in fact every physicianwhom we interviewed, had patients to whom theradial unit was contraindicated, This was not asmuch for medical as for personality reasons, butnone of the physicians had any contraindicationsagainst putting patients into the angular units.This meant that if this hospital went strictly toradial units it would lose about 20 percent of itsoccupancy because the medical staff would admitthat many patients to a facility with other shapedunits-not for intensive care, that is, but onlyfor intermediate care.

I should mention too that we were further sur-prised when we did this same kind of study forself-care patients. The radial unit was designedfor intensive care so that nurses could constantlyobserve patients who need this type of monitoring,not electronically but by eyeball, with minimumeffort. When we dropped to the intermediate levelwe had problems. When we dropped to the min-imal self-care needs we were surprised. The self-care unit was far better from the standpointstaff and patients in the circular than in theexisting angular self-care facility.

Medical, Physical, and Social Types ofDirect Patient Care

69

Direct care was examined in terms of threesubcategories. These were medical, measured bythe frequency of tests, measurements, IV therapy,preoperative care, medication and prescription

administering (treatments as a major category);physical, measured by frequency of providing bath-ing and grooming, comfort, meals, and courtesyservices, elimination and ambulation services to pa-tients (general supportive care); and social, indi-cated by direct bedside communication and ob-

servation of patients. It was found that morephysical care was provided, followed by medicalcare and last social care, as presented in table 4.Between the two units, no differences occurred inmedical care, but more physical and less social careoccurred in the radial than in the angular unit.

For the nurse staffing patterns, the total patternof such care did not differ for the high staffingpattern. Less social care occurred for the averagepattern and less medical and more physical careoccurred for the low staffing pattern (table 5). Forthe qualitative patterns, more medical and less

Table 4.-Medical, physical, and social types of direct care, by number oE, observations and percentages,total study, by shape of unit

Type of care Total Radial Angular

N,.mber Percent Number Percent Number Percent

Medical _ 23,057 38 11,199 38 11,858 38

Physical . 29.631 48 15,135 51 14,496 46

Social 8,650 14 3,446 11 5,204 16

Total - 61,338 100 29,780 100 31,558 100

Table 5.-Type of direct patient care, by nurse staffing patterns

Nurse staffing prttern

Type of care High Average Low RN LPN NA

Num Per. Nun, Per Num Per Num- Per. Num Per. Num- Per-

ber cent ber cent ber cent ber cent ber cent ber cent

Medical 4.594 38 4,232 39 3,000 35 3,238 41 4,220 40 3,773 34

Physical _ 5.628 48 5,264 49 4,363 52 3,896 49 5,014 47 5,465 49

Social - ---. -- 1,739 15 1,324 12 1,093 13 854 10 1,398 13 1,942 17

Total 11,961 100 10,820 100 3,456 100 7,988 100 10,632 100 11,180 100

78

70 BEST COPY AVAILABLE RESEARCH ON NURSE. STAFFING IN HOSPITALS

Table 6. Type of direct patient care by nurse staffing patterns, by radial and angular units,percentages only

Staffing pattern

NIethcal

Type of direct care (percent only)

Physical Social

Radial Angular Radial Angular Radial AngularHigh 10 37 51 43 9 20Average 39 39 50 47 11 14Low :14 37 55 48 11 15High RN 39 42 51 46 '10 12High I.PN 39 40 51 44 10 16High NA 36 32 50 48 14 20

social care occurred for the high RN's, more medi-cal care for the high LPN's, and more social careand less direct care took place for the high NApattern when compared to the total staff results(table 5).

When such staffing patterns were analyzed bythe two shaped units, significant divergences againoccurred as shown in table 6. The quantitativestaffing patterns showed increases in medical andphysical care and reduced social care at highstaffing on the radial unit and an opposite resulton the angular unit. The angular unit caused moredivergences at average staffing level on the angularthan on the circular unit, increasing social andreducing physical care. Low quantitative staffinghad opposite results for these forms of care onthe two units. Qualitative differences in staffingalso had differing results on the two units. HighRN staffing exhibited more physical and less medi-cal care on t;:e radial unit and more social andless physical care on the angular unit.

Opposing results also took place for the highLPN pattern on the radial and angular units.There was more physical and less social care onthe former than on the latter unit. The high NApattern indicated an increase in medical and a de-crease in social care on the radial unit and anopposite result on the angular unit (table 6).

Again, the physical shape of the nursing unitaffected the types of direct patient care providedby the various nurse staffing patterns.

Patient. Involved Care

Patient involved care is measured by the ratioof direct nedside care combined with indirect care

and general assistance to nonpatient involved ac-tivity (standby and time out), all derived fromthe preceding data. As presented in table 7, theratio of such types of 'care for the total staff wasnearly 2 to 1 (1.93) for patient involved to non-patient involved care. For the specific nurse staff-ing patterns, the highest ratio was found for thehigh RN pattern (3.2), followed by the highLPN (2.6), average staffing (2.1), high and lowstaffing (1.5 each), and lastly the high NA pat-tern (1.3), (table 7).

Physical design of the nursing unit had a signif-icant effect on these ratios of patient involved tononpatient involved care for all six staffing pat-terns as shown in table 8. Such care increased forall three quantitative patterns and. the high NApattern and decreased for the high RN and highI,PN patterns on the radial shaped unit. Theseratios decreased on the angular unit for high andaverage staffing and high RN and increased forthe low staffing and high LPN nd high NA pat-terns on this latter unit.

Table 7.Ratios of patient involved to nonpatientinvolved care, by total staff and by nurse staffingpatterns

A. Total staff: 1)C, IC, GA / Standby, T.O. = 127,275/66,006 1.93

ft Staffing pattern / Nottpt.Pt. invol. invol. Ratio

High = 26,853 / 17,852 = 1.5Average = 21.454 / 10.055 = 2.1low = 17.255 / 11,831 = 1.5Iligh RN = 16.789 / 5,268 = 3.2High 1,PN = 23,499 / 9,194 = 2.6Fl igh NA 21,425 / 15,806 = 1.3

79

BEST COPY AVAILABLENURSING STAFFING PATTERNS AND HOSPITAL UNIT DESIGN: AN EXPERIMENTAL ANALYSIS 71

Table 8.Ratios of patient involved to nonpatientinvolved care, by nurse staffing patterns andradial and angular units

Staffing patternShape of nursing unit

AngularRadial

High 1.8 1.3

Average 2.2 2.0

I.ow 2.4 2.0

.....High RN 3.6 2.8

High LPN 2.4 2.8

High NA 1.4 1.4

Other differences in this ratio of care took placebetween the two units for these staffing patterns.The radial unit exhibited more patient involvedcare than the angular unit for all quantitativestaffing patterns. As quality of staffing dropped, pa-tient involved care also decreased on the ra,lialunit.

Amount of Nursing Care

The amount rit nursing care is measured quanti-tatively by the number of trips by the nursingstaff to the patients' rooms in the two study unitsand more specifically when such trips are initiatedby the nurse and by the patients' calls. This factorwas observed only during the day shift whensuch activity is most likely to occur. As illustratedin table 9, the nursing staff initiated by far themost trips (79 percent during the total project,while staff-to-staff trips were the least (1 percent),with patient initiated trips very low (3 percent).

When the average number of total trips perday and those that were nurse initiated were com-pared to those initiated by patients in the unitsby the six nurse staffing patterns, it was found

that the high NA staffing pattern had the highestaverage number of total trips per day. This wasfollowed in order by high staffing, RN, aver-age staffing, high LPN, and lastly low staffing, in-dicating a mixed pattern of trip activity (table10). However, the ranking shifted somewhat whenthe average per diem number of nurse initiatedtrips was examined. High staffing showed thehighest average number of such trips followed inorder by high NA, average, high LPN, and finallylow staffing.

The average number of patient initiated tripswas far lower than nurse initiated trips. Averagestaffing had the highest daily average followed inorder by high staffing and high RN (tied), highNA, high LPN, and low staffing patterns. The lowstaffing pattern was the lowest in such trips, andhigh NA, high staffing, and average staffing pat-terns were !iighest in total average trips, nurseinitiated trips, and patient initiated trips in thatorder.

The average number of total trips to patientrooms was more affecod than nurse initiatedtrips an patient initiated trips, as presented intable 11. Only the high NA pattern had a higherarc Age total of trips on the radial unit, and adecrease was noted for the remaining patterns,.except for average staffing when no change wasfound. On the angular unit the opposite resultoccurred with an increase in the average numberof total trips for high and low staffing and forhigh RN and high LPN. A reduction occurredfor high NA staffing and no difference was foundfor the average staffing pattern, which was truealso on the circular unit for this pattern.

For nurse initiated trips, fewer but sufficientdifferences occurred for the nurse staffing patterns.There was a decrease in such trips on the radial

Table 9.Types of trips to patients' rooms by nursing staff in radial and angular units, by number ofobservations and percentages

Types of trips Total Radial unit Angular unit

Number Percent Number Percent Number Percent

Nurse initiated 118,444 79 60.772 82 57,672 78

Patient initiated _ 4,235 3 2,100 3 2,135 3

Tripsp,. absent 11,843 8 5.619 7 8,224 8

Tripsno pts. involved 13,121 9 4,495 8 8,826 11

Staff to staff 2,409 1 1,251 2 1,158 2

Total __ .... 150,052 100 74,237 100 73,813 100

so

BEST COPY AVAILABLE72

Table 10.Average daily number of trips to pa-tints' rooms by nurse staffing patterns, by totaltrips and by types of trips

Staffingpattern

"fetaltrips Untiattil

Patientinitiated

High _ 732 688

Average 665 523 22

Low . 589 485

High RN 675 496 21

High LPN 626 497 16

High NA 797 638 Is

unit for high and average staffing and for highRN staffing, au increase for the high NA pattern,and no change for the low and high LPN staffing.On the angulai unit, decreases in such trips oc-curred for the high, average, and low staffingpatterns and also for the high N pattern. Nochange took place for the high I.PN pattern. Forpatient initiated trips, only the high staffing pat-tern showed a change. There was a decrease in


average ntuW)er of such nips on the radial unit,and an increase occurred for this high staffingpattern and a decrease for the high RN staffingpattern on the angular unit.

Thus both staffing pattern and shape of unitexhibited differential effects in the amount ofnursing care provided in this study.

Utilization of the Unit by Nursing Staff

Since differences in nursing care were foundfor all of the preceding variables in terms of nursestaffing patterns and physical design of the units,it is of special interest to analyze how the nursingpersonnel utilized these units. Table 12 shows thatthe muses were observed spending the greatestamount of their time in the nurses' station (31percent). Following in order were the patients'rooms when the patient was present, nurses' con-ference rooms located within the units, off the

Table 11.Average daily number of trips to patients' rooms by types of trips, by nurse staffing patterns,and by radial and angular units

Type of trip to rooms (daily average)Staffing

Total trip,. Nurse initiated Patient initiatedpat tan

Radial Angular Radial Angular Radial Angular

High . 705 7 i9 617 586 15 14

Average 661 665 5.11 502 20 22Low 557 621 488 462 15 14

High RN . _ 646 705 487 505 22 19

High LPN 593 659 500 495 17 16

High NA 853 742 692 585 18 19

Table 12.Utilization of nursing unit by total nursing staff in radial and angular units, by number ofobservations and percentages

Total1:nit subarea Radial Angular

Number Pert, nt Number Percent Number PercentNurses' station 73,399 34 39.985 37 33,414 31

Mulication arit 0,313 4 4,850 4 4,463 4

Corridor in unit 17,634 8 8,393 8 9.241 9Patients. rooms 46,608 22 22.002 21 24,606 23Utility room 2,750 I HIS 1,015 2Tub and shower room 1.064 657 1 407Nurses' conterence room 34,574 16 14.010 13 20.564 19

Off unit . 24,380 11 13,068 12 11,312 10

Other areas in unit 5.377 3 3,009 2,368 2

Total 215,099 100 106,789 100 108,310 100.....I,ess than I percent.

81

BEST COPY AVAILABLE

NURSING STAFFING PATTERNS AND nosPrrAt. UNIT DESIGN: AN EXPERIMENTAL. ANALYSIS

entire unit, corridors within the units, medica-tion arca, miscellaneous areas in the unit, pa-

tients' rooms when the patient was absent, utilityroom, a 10 lastly, in the tub and shower room inthe unit.

Under conditions of differing nurse staffing pat-terns, the nursing personnel were found to utilizethe two units somewhat difieremly than for thetotal study, as illustrated in table 13. For thequantitative staffing patterns, high staffing person-nel used the nurses' conference room and wereoff the unit more and spent less time in thenurses' station and in patients' rooms when pa-tients were present Average staffing indicated anincrease in time in patients' rooms when patientswere present and a reduction in time in the con-ference room. Low staffing increased time in pa-tients' rooms only when patient..., were present.

For the qualitative patterns, high RN staffingled to an increase in use of the muses' station,medication area, and patients' rooms with pa-tiznts present, and a decrease in use of the con-ference room and in being off the unit. HighLPN staffing led to more time in the nurses' sta-tion and to less use of the conference room. Thehigh NA staffing pattern led to increased use ofthe conference room and decreased time spent inthe nurses' station and medication area.

73

For each nurse staffing pattern observed on thetwo units some inter esting differences were foundin utilization of these subareas in both units, asshown in table 14. High quantitative staffing ledto an increase in use of the nurses' station andin being off the entire floor and a decrease itu

time spent in patients' rooms while the patientwas present and in the nurses' conference roomon the radial unit. Use of the conference roomincreased on the angular unit and use of thenurses' station decreased. Average staffing broughtabout an increase in use of the nurses' stationand a drop in time spent in patients' rooms, pa-tients present, and use of the conference roomon the cir:ular unit. The opposite occurred onthe angul. r unit. Low staffing resulted in an in-crease in time off the floor and a drop in use ofthe conference room.

In terms of the qualitative staffing patterns, highRN staffing led to an increase in time in thenurses' status and patients' rooms with patientspresent and a decline in use of the corridors andnurses' conference room on the radial unit. Therewas an increase in .use of the conference roomand a decrease in time spent in the nurses' stationand patients' rooms with patients present, on theAngular unit. High LPN staffing on the radialunit showed a drop in use of patients' rooms with

Table 13.-Utilization of cursing unit by nurse staffing patterns, by number of observationsand percentages

Unit subareaNurse staffing pattern

High Average Low RN LPN NA

Num-ber

Per.cent

Aft1111ber

Per-cet

A'um-ber

Per.cent

Num-ber

Percent

Num-ber

Percent

Num-ber

Per.cent

Nurses' station 16,471 32 12,218 35 9,679 35 8.699 36 14.265 40 12,742 30

Medication area 2,065 4 1.574 4 1,239 4 1.450 6 1,829 5 1,046 2

Corridors . _ 4,646 9 3,152 9 2.200 8 1.571 7 2,926 8 3,563 9

Patients' rooms,patients in 8,534 17 7378 23 5.951 22 5,558 23 7,273 20 8,378 20

Patients' rooms,patients out 1.035 2 346 420 I 482 2 366 1 632 2

Utility room 780 511 348 1 360 1 363 1 629 2

Tub and Showerroom _ -- 258 181 76 22 199 399 1

Nurses' confercute room 9.861 19 4.903 14 4,171 15 3,264 14 4,044 11 8,607 21

Off unit 6,955 13 3,325 10 2,741 10 2,296 9 4,387 12 4,814 11

Other areas 1,026 2 882 2 829 3 470 2 927 2 1,055 2

Total 51,631 100 34,008 100 27,654 100 24,172 100 36,579 100 41,165 100

Less than I percent.

74 BEST COPY AVAILABLERESEARCH ON NURSE STAFFING IN HOSPITAI.s

Table 14. Utilization of nursing unit by nurse staffing patterns and radial andpercentage of observations only

angular units, by

Nurse staffing pattern and unit (percent only)

Iligh A%erage

Radial Angular Radial

low RN

Radial Angular

I.PN NAsubarea

..%ngillar Angular Radial Angular Radial Angular

Nurses' station 37 27 39 :11 :15 :15 :19 33 '49 39 36 25

Medication ;tin 4 4 5 4 5 4 6 6 5 5 3 2

Corridors . 9 9 8 10 9 7 5 8 8 8 8 9Patients' rooms.

patients in 15 18 21 2122

25 21 18 92 19 21

Patients' rooms.patients out 2 2 I I 1 2 2 2 1 1 2

Utility room I 2 I0 2 1 2 1 I I 2

Tub and shover room 1

Conference room 24 12 16 13 17 11 16 12 10 16 25Off unit ;5 12 10 9 12 8 9 I0 12 12 12 11

Other ar-as 2 2 3 2 3 3 2 2 S 2 3 2

'Less than i percent.

patients present and the opposite result on theangular unit. The high NA staffing pattern ex-hibited an increase in use of the nurses' stationand a decrease in time spent in the nurses' con-ference room on the radial shaped unit, withthe opposite result on the angular unit.

Thus, it . staffing patterns, directly oppos-ing results . tilization of various subareas ofthe nursin ; unit took place between the twophysically shaped units. This again indicated some

effects of such design on nurse staffing patterns.

Conclusions and Discussion

Space permits only the mention of some majorconclusions from nut- findings. As indicated bythe measurements employed, nursing and patientcare differed both in terms of nurse staffing pat-tern and the physical design of two divergentlyshaped nursing units in an acute care generalhospital. The results were not always uniform orpredictable. Both quantitative and qualitativenurse staffing pattern differentials were affecteddifferently by the two diveigently shaped nursingunits tor different variables related to patient careand utilization of the unit by nursing personnel.

Undoubtedly, part of the reason for varyingresults was the rather gross measurements ern.plcyed in this study. Most of the data were ofthe enumerative type with simple percentages ofthe major measurement data. It seems likely thatmore precise measures of nursing and patientcare than were available for this study may wellhave produced more detailed and precise differ-

83

ences in care provided by the different staffingpatterns. Also, expanding the divergency in phys-ical designs beyond merely the circular andLshaped angular types included in this studymay have revealed even more differences in im-pact on patient:care. The same could well holdfor expanding the different types of nurse staffingpatterns along both qualitative and quantitativelines for additional divergences in results relatedto patent care.

Only the most basic types of patient care werereported here, and such variables would be easilyextended into many other phases of such care,even with the same level of gross measurement,We should perhaps also suggest that our overallfindings might have differed with other nursingpersonnel in other hospitals, to mention only afew necessary qualifications of the results of thisst tidy.

Our major conclusion nevertheless remains that

NURSING STAFFING PA1TFRNS AND HOSPITAL. UNIT DESIGN: AN EXPERIMENTAL. ANALYSIS 75

hwpital physical design affected differentially the,atient care provided by six different nurse staffingpatterns for general medical and surgical patientsat the intermediate level of nursing care. Froma strictly research standpoint, then, investigationsof nurse staffing patterns should take the physicalsetting into account as a potential factor affectingthe outcome of such research. Omitting these con-siderations and potential effects will likely bias ordistort the results of such research on nurse staffingpatterns.

0..ir study suggested a number of theoreticalproblems. First of all, to return to Dr. Aydelotte'ssuggestion that we need theory, I think we haveto start to do some pure Imsic research on justwhat wt' mean by currnot care in the nursingsense or medical sense but care in general. I re-

gard care as ba8ically a social phenomenon. The"caring" relationship is defined by society andculture. We can break the societal definition downinto "hospital care" and then into categories of"patioitt care," "nursing care," "medical care,"

and so forth. We need to do the type of basicresearch whereby we can delineate specific di-

mensions of the caring relationship that can becontrolled or enhanced when specified for patientcare.

Sonic of the medical staff of the hospital whereour study was done would like ideally for thehospital admitting office to pick out that roomand that 1,, : that would maximize the personal,physical, and medical needs of that patient, Wedreamed about this, for we found that there werea lot of personality Factors involved in both posi-tive and negative reactions to the unit, to certainnurses, to certain physicians, the way they weretaken care of, and so on. Such reactions had alot to do with the room and the whole settingin which all of this kind of care, the total medicalcare, was given to patients.

These are some of the things that we touchedon in our study and some conclusions we reachedas a result of the-study.

76

(I)

References

Jao, Gart ly. "Ecological Aspects of Pa-ticot Care and Hospital Organisation."Organization Research on Health Institu-tions, Basil S. Georgoponlos, (ed.) institutefor Social Research, University of Mich-igan, Ann Arbor, Michigan: 1972, chapter10, 418 pp.

(2) Price, Elmina NI. Staffing for Patient Care.New York: Springer Publishing Co., 1970.178 pp.

Aydelotte, Myrtle K., et al. An Investiga-tion of the Relation between Nursing Ac-tivity and Patient Welfare. Iowa City,Iowa: State University of Iowa, 1960. 412pp.

(4) Sturdavl.nt. Madelyne, et al. Comparisonsof Intensive Nursing Service in a Cirrularand a Rectangular Unit. Chicago: Amer-kan Hospital Association, Monograph Se-ries No. 8, 1960. 220 pp.Trites. David K., et al. "Radial NursingUnits Prove Best in Controlled Study."Modern Hospital, vol: 109: 94-99, April1969.

(6) "Social Interaction as the Definition of theGroup in Time and Space." Translatedfrom Georg Simmel, Soziologie, by AlbionW. Small, in Robert E. Park and Ercie-st W.Burgess, Introduction to the Science ofSociology Chicago: University of ChicagoPress, 1921.Hawley, Amos H .uman Ecology. NewYork: Ronald Press, 1950.Hall, Edward T. The Hidden Dimension,New York: Doubleday & Co., Inc., 1969.217 pp.

Sommer, Robert. Personal Sifre. Engle-wood Cliffs, NJ.: Prentice-HalL Inc., 1969.177 pp.

(10) Barker, R. G. Ecological Psychology.Stanford: Stanford University Press, 1968.

(11) Osmbnd, H. "Function as the Basis ofPsychiatric Ward Design." Mental Hospi-tals, April 1957 (architectural supplement),pp. 23-29.

(3)

(5)

(7)

(8)

(9)

RESEARCH ON NURSE STAFFING IN HOSPITAI S

(12) Hall, Edward T. "Proxemics." CurrentAnthropology, vol. 9: 83-108, AprilJune1968.

(13) Alihan, M. A. Social Ecology. New York:Columbia University Press, 1938.

;14) Lippert, S. "Travel as a Function of Nurs-ing Unit Layout," Tufts University,. HumanEngineering Information & Analysis Serv-ice Report No. 113, March 12, 1969.

(15) McLaughlin, H. P. "What Shape is Bestfor Nursing Units?" Modern Hospital, vol.103:84-89, December 1964.

(16) Pelletier, R. J., and Thompson, J. D."Yale Index Measures Design Efficiency."Modern Hospital, vol. 95, November 1960.

(17) Rosengren, William R., and De Vault, S.

"The Sociology of Time and Space in anObstetrical Hospital." in The Hospital inModern Society, Eliot Freidson, (ed.) NewYork: The Free Press, 1963, chapter 9.

(18) Sommer, Robert and Dewar, A. "ThePhysical Environment of the Ward." inThe Hospital in Modern Society, EliotFreidson, (ed.) New York: The Free Press,1963, .hapter 11.

;19) Sorokin, .Peterim A. Sociocultural Causality,Space, Time. Durham: Duke UniversityPress, 1943.

(20) Thompson, John D. "Efficiency and Designin the Hospital Inpatient Unit." Poceed-ings of the 13th International HospitalCongress, (Paris, 1963). London: Interna-tional Hospital Federation, 1963. pp.- 67-75.

(21) Trites, David K.: Galbraith, F. D. Jr.;Sturdavant, M.; and Leckwart, J. F. "FinalReprt of Research Project Entitled: In-fluence of Nursing Unit Design on the Ac-tivities and Subjective Feelings of NursingPersonnel." Rochester, Minnesota, Re-search Office, Rochester methodist Hospi-tal, 1969.

(22) Wheeler, E. Todd. Hospi:al Design andFunction. New York: McGraw-Hill BookCo., 1964. 296 pp.

DISCUSSION OF DR. JACO'S PAPER

Discussion Leader

Dr. Stanley JacobsAmerican Hospital Association

Chicago, Illkois

Dr. Jacobs

I was most impressed by the experimental de-sign of Dr. jaco's study. The number of variableshe was able to manipulate in a controlled situ-ation is not often encountered in hospital re'.

search. The study was more like an investigationcarried mu in a laboratory than one performedin an operating organization. He is to be com-mended for "selling" the hospital managementon cooperating in the research.

Dr. Jaco, as most of you know, is a sociologist.

I, by training, am an industrial psychologist. Veryfew nursing or personnel research studies in hos-pitals are being designed and directed by -in-dividuals with degrees in the social sciences. Most

such studies reported in the literature have beenconducted by industrial engineers. They have mademany significant contributions. I do riot believe,however, that most of them would conceive of

experimental designs as sophisticated as that de-scribed by Dr. Taro.

Although I have been out of industrial psychol-ogy for some time, I still scan the journals ofthis discipline. Very seldom do 1 see articles de-scribing research in a hospital setting by psychol-ogists, other than clinical psychologists. In fact

the only one that comes to mind is a study atIowa State in which size of staff and personalrelations training for nurses were varied in orderto determine impact upon patient reaction to

care received.The rapid growth in the number of industrial

engineers working in the hospital field was greatly

77

stimulated by the Kellogg Fouadation. Perhaps

the entry of a substantial number of social scientistscould be similarly stimulated by the U.S. Depart-

ment of Health, Education, and Welfare'(DHEW). The American Psychological Assoeik-

tion (APA) is located in the Washington area:As a starter; DHEW might make arrangementsfor a symposium at the APA convention to pointout some of the hospital and nursing problemsand to inform psychologists of the research op-poi tuni t ies.

In some cases there is already a sufficient body

of psychological knowledge to predict the out-

comes of staffing decisions. Earlier todayDr. jelinek pointed out that nurses are reluctantto float when they are identified primarily withunits in which morale is high, but they are wilting

to float from units in which morale is low. Mostpsychologists would have predicted this finding.

My basic appeal in these comments is that,through research, social scientists should he ableto make a significant contribution to the solutionof many of the problems we have been discussingin this conference, In addition, they could cer-tainly assist hospital management in making per-sonnel decisions.

I had hoped when I first saw the title of Dr.jaco's paper that he was going to tell its whichwas "best," radial or conventional units. He didnot do that. However, he told us of a wealthof other findings discovered during the course of

the study.

78

In regard to hospital architecture, there are tworeports recently prepared for the Department ofDefense concerning a "new generation of mili-tary hospitals." I procured them because I wasinterested in what they had to say about com-puterization. The two reports consist of nine antiseven volumes, respectively. One was rrepared by1Vestinghouse and the other by Arthur D. Little.They have a good deal to say concerning archi-tecttire for hospitals and nursing units therein.


Finally, I strongly concur in Dr. Jaco's commentabout replication of researches. In psychology,replication of researches (or cross validation ofpersonnel tests) is most important before one ac-cepts findings with no reservations. We seldomdo this in hospital research. No matte.: howsophisticated our statistical tests, significi:nt studiesshould be replicated before hndings are ac-

cepted, as doctrine and decisions are based uponthem.

Impact of Administrative and Cost Factors

on Nurse Staffing

Mr. John R. GriffithPTOYMOT and Director

Program and Bureau of Hospital AdministrationThe University of Michigan

Ann Arbor, Michigan

To explore the impact of administrative andcost factors on nurse staffing. one might use themodel developed by ',con, which implies that thepolicies and operating procedures of a hospitalat any given time represent the sum of two in-fluences. These are the influences generated tyy

forces from the outside environment and thosegenerated from within the organization by pro-fessional and employee attitudes and knowl-

edge (1). In the case of nursing, the parties tothe negotiation bring quite divergent backgroundsand therefore have, as a result, quite different setsof cognitive information, attitudes, and values.The bargaining process would incorporate the

usual searches for mutually acceptable solutions,

compromises, and innovations that meet or at

least partially satisfy multiple goals simultane-ously. The principal parties from the outside en-vironment would be third party financing agen-cies, trustees. and hospital administrators. Fromwithin the organization, they would be the pro-fessional and employee groups of nurses, doctors,and nonprofessional nursing employees.

79

While this model of the establishment of nursestaffing policies is simplistic, it suggests that theterm "administrative and cost factors" would beall those demands and desiderata of the outsideparties. Such a list is constantly changing and ispotentially quite lengthy. If the major goals ofthe outside bargainers are postulated, it is possibleto predict something of the nature of the outcomeof the bargaining process. This then can be aguide to understanding the current evidence onthe actual state of nurse staffing and to drawingsome tentative conclusions on the accuracy of themodel and postulate. Finally, the model canframe %peculation upon the impact of the majorchanges in the demands of the outside partiesthat currently appear possible.

Obviously, there are some weaknesses in this

approach. Any assumption about the goals of in-dividuals or groups is subject to question. Thelists of both outside and inside parties have been

arbitrarily curtailed, and because of the pressureof time, the limited power of our analytic tools,and the shortage of empirical evidence, we will

be limited to some general conclusions.

80

Problem Definition

The goal of nurse staffing is delivery of anappropriate quantity and skill level of nursingservice to each individual patient. The issues thatmust be resolved are the definitions of appropri-ate quantity and appropriate skill level and themechanisms of delivery, including the tolerancesto which the final nurse staffing system is to beheld. These issues are related to, but can beseparated from, a variety of organizationalparameters that influence the outcome. The or-ganizational parameters include the training andrecognition of capability of each of the possibleskill levels, the physical facilities of the wards,the nature of supervisory and management struc-ture under which patient care is given, and theuganivatiunal structure of ate hospice iNN v.;titthe nursing staff performs.

For example, one might assume that the up-grading of registered nurse skills to the bacca-laureate level would influence the relative quan-tities of registered nurses and practical nurses.Similarly, decisions that registered nurses maydraw blood specimens or that they are responsi-ble for admitting histories would influence theappropriate quantities of RN skill levels. Thearchitecture of the patient care units might aflectthe staffing (2,3), and so might decisions thatchange the management style of the head nursesand higher levels of supervision (4), and so wouldorganizational decisions that create patient unitmanagement departments.

The goal definition suggests that once theseorganizational parameters have been fixed thereexists a unique quantity of nursing resource thatwould be judged appropriate to the needs of anygiven patient for a given time period, such as aday or a shift.. This seems to he consistent withmost thought on the problem. Such comment asthere is on the appropriate level of care is givenin hours per patient days (5). Many of the existingempirical imestigations tend to treat this as hav-ing face validity as the dependent variable (6).Further, there appears to be a consensus that theappropriate quantity and skill level for an in-dividual patient is a function of the severity ofhis illness (7,8,9).

The bargaining model suggests that the e en-tual definition of "appropriate" will be negoti-


ated. That does not preclude the use of substantialfactual material. In fact, one might infer that ifthe bargaining were proceeding normally, a widevariety of both evidence and argument would beintroduced. The lines of argument will followrecognizable patterns, however, of which two areperceptible. These can be labeled the engineeringargument and the economic argument.

Engineering Approach

The more commonly encountered definition ofthe appropriate quantity of nursing service couldbe labeled as An engineering approach, both be-cause enginefrs frequently erl t and becauseit can be traced to Taylor's (11) analysis of jobshop time standards. This approach assumes thereis a fixed, describable quantity of activities thatmust be performed for a given patient in a giventime period. Further, given a set of organizationalparameters, the incremental times for each ofthese activities c.r the total time for the full setcan be measured by standard work measurementtechniques. This approach implies that at leasta major portion of nursing care can be describedin terms of discrete, identifiable, measurable com-ponents, and that for any given patient there isreasonable consensus as to which of these com-ponents is appropriate.

The approach has a reasonably successful his-tory in industrial job shops. The underlying as-sumptions, however, have not been tested in a

nursing environment. While existing work showsthat time estimates can be made with satisfactoryvariances For both large and small componentsof activity, the author is not aware of any workthat shows consensus on the components desirablefor specific patients. Neither has it been demon-strated that the components are trvtly additive.

Another difficulty has been oveooked. That isthat in the industrial job shop it is the usualpractice for managemtlit to specify exactly whatshould be done, but in the ho ')ital it is theusual practice for management to delegate to theregistered nurse it, charge of the patient the iden.tification of the patient's needs and the work tobe performed. Even if nursing supervision were

IMPACT OF ADMINISTRATIVE AND COST FACTORS ON NURSE STAFFING Sl

willing arid able to approve musing care plansor some other specification of work for individualpatients, it :s quite difficult to determine whetherthese specifications have been met.

At the present time there appear to he nopractical systems for assessing the quality of theresult. Thus, while the engineer may measure tohis heart's content what he thinks might be donefor the patient, the nurse on the floor will con-tinue to do what she thinks is desirable for thepatient. The two decision systems can coexistnicely so long as the engineer does not force thenurse into a situation where she is routinely dis-satisfied with her professional performance or herwork pace.

Economic Approach

The other approach to the question of the ap-propriate quantity and skill level of nursing carecan be called an economic approach. Althoughit is conceptually at least equally appealing, ithas received very little attention in the literature.This approach would assume that nursing care isa resource sought by a consumer or a physicianon his behalf because he perceives that it hassome utility for him. 'Thus, in some more or less

imperfect fashion, the amount of nursing carethat will he judged appropriate can he predictedby economic theories rather than by strictly em-pirical considerations as the engineering approachwould suggest. Under these theories, nursing carehas a certain utility, and the utility is generallydependent upon the quantity rf care received.

Although it is difficult to define this relation-ship, the marginal utility, or the value of an ad-ditional increment of a given good or service, is

generally envisioned as a downward sloping curve,as 3 howr in figure 1. If we can assume for themoment that a typical hospital patient buys hisown ntrsing care, this relationship is not un-reasonable. The first care that he receives mayhe lifesaving, :end therefore lie would valtic it

infinitely highly. As the quantity goes up, theutility of additional units goes down and even-malt reaches a value very close to zero.

At least theoretically, it is now possible to de-cide the appropriate amount of care. It will bethat amount at which his perception of th.: mar-ginal utility of a given number of dollars af ex.

pend t tire on nursing care is equal to hisperception of the utility of the same number ofdollars if spent for other goods and services. Sincehe will also buy every other good and serviceaccording to the same standard, he will makealternative purchases as, for example, betweennursing care and laboratory services, until theirmarginal utilities are also equal. This means thathe will purchase nursing care over any alternativeservice until he has purchased the quantity ofnursing care where its marginal utility is exactlyequal to the marginal utility of the last quantityof another service that he purchased.

Put another and more concrete way, the pa-tient will spend $10 for nursing care so long ashe feels he can afford the $10 and so long as hefeels that $10 worth of nursing care is at least

,as valuable as $10 worth of laboratory service orany other good or service available to him. Theequilibrium that is reached will be a function ofthe following:

the patient's income, because this determineshow willing he is to spend the money;the patient's preference or tastes, because thesedetermine on what he wishes to spend hismoney;the nurse's income, because this determinesthe cost of the marginal unit;the state of health care technology, becausethis otp! Ilipes the utility of a given phys-ic41 qnant4y of tmrsing resource and alsothe 9114Y of r titer goods and services;the general economy, because this influences

not only the relative incomes of the patientand the nurse but also tie relative incomes ofother resources, including both labor and cap-ital resources.

To summarize the economic approach, onemight assume that the appropriate quantity ofnurse staffing, than, is determined by people's per-ceptions of relative incomes of patients andmuses, by u.chnology, and by the desirability ofspending money generally for health care.

Patients, of course, do not purchase their ownnursing care. The decisions that must be madeare collectivized in several different ways. Thenursing resource is provided to a nursing unit,the quantity of resource available is sometimeslimited, the impact of the patient's personal in-come is to a large extent alleviated by third party

90

82

payment mechanisms, and the problem of evalu-ating the marginal utilities of various hospitalgoods and services is assigned to the bargainingprocess between the outside parties and thenurses and other professionals within the hospital.In this regard, however, it must be noted thatwhile the physician acts as the patient's agent inselecting the qua titles of most of the other goodsand services in the hospital, it remains the nursewho makes must of the decisions in regard to'nursing care.

In some ways, this collectivization is an im-portant asset to our problem. If the severity ofillness is a random phenomenon, it is reduced' bytreating patients in groups. The elimination ofindividual patient income from consideration is

clearly desirable. There are a number of obviouseconomies that result from grouping patients.Finally, the data (rem th engineering approachare not discarded. They enter the arguments asa way to evaluate the translation of inputsnurs-ing hours, to benefitsnursing service.

The normative roles usually assigned to thevarious decision makers in the hospital suggesthow the bargainers might behave. The clinical


professionsprimarily nurses in this casewillbargain for the purchase of additional nursingservices. They will attempt to argue that the util-ity of another nursing service exceeds its cost andexceeds the value of any other good or serviceavailable. Their arguments will center on qualityand on the desirability of identifiable elements ofservice that will be foregone if their demands arenot met. The administrator and the trustees wil:bargain to protect the patient's purse, arguingthat he needs the money or that there are othergoods or services that have a higher utility. Ifthese stereotypes apply and there is "hard bar-.gaining" or a considerable impact of adminis-trative and cost factors, there will be certainpredictable results.

First, within the same labor market, under thesair, set of third party arrangemems, and forsimilar groups 01 patients, similar groups of bar-gainers would tend to have the same numberof hours per patient day. Second, there would beconsiderable pressure to identify and install theleast-cost organizational parameters. All those thatcost less than the value of the nursing iime theyrelease would be installed because this 6 to the

Figure 1.Marginal utility of nursing care.

0Quantity of nursing care

IMPACT OF ADMINISTRATIVE AND COST FACTORS ON NURSE. STAFFING 83

advantage of both parties. Again, one would ex-pect consensus to develop am ng hospitals in shn-ilar situations. Third, given tat nursing care is

a high priced and scarce. resource, considerable,,ttention would be given to controlling the actualdelivery of nursing care so that it would not devi-ate from the level agreed on as appropriate.

There are some special cases of this analysisthat must be discussed briefly. One is the casewhere the supply of nurses is consistently less

than the demand at the prevailing wages. Ashas been indicated by Yett (12), this situationwould be characterized by steadily rising hospitalnurse wages at a rate faster than the rise ofwages in other nurse employment or other similaremployment opportunities. It also would be char-cterized by an even greater pressure to seek nurse

saving changes in the organizational parameters.he more interesting case is where there is no

utility consensus as a result of the bargainingprocess. One possibility is that there is no lardbargaining" or that the bargainers perceive themarginal utility of additional nurses as being moredifficult to estimate than the impact it will haveon the outcome. Under this situation, the deci-sion as to the amount of nurse staffing may betreated as unimportant and the amount of nursestaffing would tend to be reduced slowly to thepoint where a strong consensus could be reachedon the undesirable impact of reducing it further.

The second case of lack of consensus would hewhere the utility is perceived very differently bydifferent parties to the....bargaining. Under thiscase the solution would reflect the most powerfulbargaining group. One might expect actualsolutions to be subject to sharp variations de-pending on which group is in control. The situ-ation would probably be accompanied by con-siderable tension and dissatisfaction. These twocases of lack of consensus can be compared to aship. In the first case, no one is interested inwhere the ship is going and it is allowed to drift.In the second case, the ship sails smartly in a

variety of directions depending on who is at thewheel.

It is of some interest to consider what happensto the engineering approach under these eco-

nomic cases. Where there is consensus there willbe considerable interest in the engineering ap-proach to the extent that it rationalizes the agreedupon decision and is useful in finding least-costorganizational parameters. Otherwise it will beignored. It also will be ignored in both cases

where there is no strong consensus. Where theship is drifting aimlessly, no one will wish toinvest the money necessary to evaluate variousnursing activities. Where there is a continualstruggle for the helm, all energies will go to-wards the battle.

The Current Performance

The implications of the engineering approachtaken alone and of the economic approach in theanalysis of the evidence with regard to nursestaffing are quite different. The engineering ap-proach would imply that hospitals that have agiven set of organizational parameters and thattreat patients of similar average severity of illnessshould have the same quantities of nursing care:The economic approach would suggta that fac-tors quite outside the organization itself, such asthe extent and kind of third party coverage andthe wages of nurses relative to other health careresources, would also affect the amount of nursestaffing.

Also, the economic approach provides additionalexplanatory power. It suggests that if the varia-

tion in nurse staffing between hospitals that aresimilar in these characteristics is large, then eitherthe bargainers do not care about the results orthey cannot reach a stable agreement. Neither oneof these finch J tad be so. 'ally useful. It is

incongruous to say that no one cares about themost expensive single resource that the hospitalpurchases. Neither is it a hopeful sign to contem-plate continued tension and abrupt changes inthe quantity of nursing resource.

Actual Nurse Staffing

Ilnfortunately, the definitive research suggestedhere does not appear to have been clone. There arc

84 BEST COPY AVAILABLE

a number of empirical problems that make onehesitate to attack such research. Such evidence aswe do have is far from encouraging. One finds, forexample, that the number of nurses employed inhospitals by the State bears little relation to thebed supply of those States (refer to table 1). Notonly do the States range from 171 nurses per1,000 beds to 436 nurses per 1,000 beds, a differenceof over 150 percent, bat adjacent States vary by10 percent or more.

For example, see New jersey at 300 nt,rses per1,000 beds and New Yo '4, Pennsylvania, and Delaware, or Massachusetts, Rhode Island, and Con-necticut. The widest of these variations is betweenArizona, the second highest State, and New Mexico-424 and 264 RN's per 100 beds. There also is


no pattern between economically similar States.Michigan is 20 percent lower than Illinois, forexample. The causes arc likely to lie in licensureexamination levels, acceptance of reciprocity, andthe presence of schools of nursing. Rut these arcitems that over long periods of time are withinthe control of the bargainers.

Data compiled by Hospital Administrative Serv-ices (HAS) on the number of nurse hours perpatient day similarly show an extraordinary varia-tion. HAS segregated hospitals by State, district,function, and size, thus making at least crude ap-proximations for many of the explanatory varia-bles. These data, shown in table 2, are quite recentbut include only 200 self-selected hospitals. The

Table 1.-Loca*ion of RW8 employed in hospitnts ift gelation to hospital beds: 1968

Location RN'sRN's per

1,000 hospital beds Location RN'sRN's per

1,000 hospital beds

United States .

New EnglandConnecticutMaineMassachusettsNew HampshireRhode IslandVermont

Middle AtlanticNew Jersey -

New York _. _

Pennsylvania _

South AtlanticDelawareDistrict of ColumbiaFloridaGeorgiaMarylandNorth CarolinaSouth CarolinaVirginiaWest Virginia

East South CentralAlabamaKentuckyMississippiTennessee

West South CentralArkansasLouisianaOklahomaTexas

445.243

8,179

2,498

20,007

2,142

2,527

1,361

15.715

49,290

33,392

1,242

3,610

13,079

6,877

8,172

8,642

3,876

8,707

4,037

4,986

5,104

2,837

6.051

2,340

,978

3,741

16,338

268

336

257

313

319

289

289

300239

287

248

236

294

202

245

243

206226

232

175

222

171

183

211

186

225

218

East North CentralIllinoisIndiana ..Michigan

Wisconsin __ .

West North CentralIowa -

Kansas

MinnesotaMissouriNebraskaNorth DakotaSouth Dakota

MountainArizonaColoradoIdahoMontana _

NevadaNew Mexicolitah .

Wyoming

PacificAlaskaCaliforniaHawaiiOregonWashington

27,653

9.43216,338

22,894

10,357

7,284

5,129

11.668

8,594

4.204

1,832

1,711

3,936

6,084

1,391

1,960

896

1.580

1,933

796

713

44,137

1,775

5,032

8,186

269

242

229

281

275

353

258

349

221

323

294

267

424

367

352

425

322

204

384

197

357

330

284

333

436

SOURCE: Nursing Personnel in Hospitals, 1964. P.S. Department of Health, Education. and Welfare. Public Health Service,National Institutes of Health, Washington: U.S. Government Printing Office, 1970.

IMPACT OF ADMINISTRATIVE AND COST FACTORS ON NURSE. STAFFING 85

Table 2.-Median nursing hours per adult medical surgical patient day, HAS subscriber hospitals,6 months ending 12/31/71

Bed site

Region Under 50- 75- 100- 150- 200- 300- Over50 74 99 149 199 299 399 400

New England 14.04 7.07 6.90 6.71 6.56 6 23 6.40 6.46

Middle Athol tic 8.22 6.24 5.91 5.53 5.84 5.60 5.72 5.54

South Atlantic 7.27 6.41 6.80 6.26 6.09 6.12 6.47 6.04

East South Central 6.39 6.60 6.31 6.46 5.98 6.03 5.93 5.8?

West South Central 6.62 6.65 7.49 6.80 6.25 6.08 5.42 5.36

East North Central 6.65 6.72 6.73 6.49 6.74 6.48 6.33 5.98

West North Central 7.05 6.58 6.87 6.71 6.93 6.32 6.28 6.51

Mountain 6.78 6.73 6.47 6.84 7.81 6.95 6.53 7.01

Pacific 7.51 7.83 6.92 7.07 6.87 7.14 6.50 7.10

SOL'RCE SisMonth National Comparison. Hospital Administrative Services, (HAS) The American Hospital Association, Chi.cago.

patient care on ,nredical-surgical units, not simplyfor registered nurses.

Table 2 shows the sii adjusted regional medians.

Size High Region

The follow:ngin size gr mtp:

The data

chart summarizes the variation

show some pattern. The Pacific

Low Region

with-

region

Under 50 beds Mid Atlantic 8.22 E.N. eentral 6.39

50-74 beds Pacific 7.83 Mid Atlantic 6.24

75-99 beds E.S. Central 7.49 Mid Atlantic 5.91

100-149 beds Pacific 7.07 Mid Atlantic 5.53

150-199 beds Mountain 7.81 Mid Atlantic 5.84

200-299 beds Pacific 7.14 _Mid Atlantic 5.60

300-399 beds Mountain 6.53 E.S. Central 5.42

400 and over Pacific 7.10 E.S. Central 5.36

tends to be high and Mid-Atlantic low. But bothregions are reasonably well supplied with registerednurses, according to the older data on table 1, andthe Mid-Atlantic States have the peculiarity ofbeing high for small hospitals aid at or near thelow for all other sites. Further. one would inferfrom these data a demapd for. more nursing per-sonnel for very small hospitals, fewer for hospitalsup to 150 beds, and fewest for large hospitalseither 300-400 beds or over 400 (see figure 2).There may he explanations for this. The analysisof returns to scale at ,.'aiying levels of service byCarr and Feldstin is suggested by the shape ofthe cost curves (13). Problems of reporting andnonrepresentative samples cannot he ignored.

The ranges of lmhavibr for all hospitals in thesize groups is shown in table 3. One is forcedto conclude from these data that size matters very

94

little, at least at the low end of the scale. All thehtth percentile values are within 10 percent of thelowest hospital, which is an over-400 bed, non-teaching hospital with 4.4 hours per adult medicalsurgical patient day. The variation increases sub-stantially for the 95th percentile, where the lowvalue is 7.15 and the high is one-third more, 10.02.The ranges within size groups, however, cover in-creases of 80 to 150 percent.

The conclusion the author draws from thesedata is that there is no consensus on the questionof nursing t-tility. The reurring wide variationsand the apparent failure of known theory to

explain the cause suggest that nurse staffing is

simply rn)t subject to administrative and cost fac-tors. Uodoubtedly, more careful research wouldreveal considerable impact tr:)in organizational



Figure 2.High and low region medium AM&S mrsing hours per patient day.

9.0

8.0

7.0

13 6.0

5.0a.

k 4.0

3.0

2.0

1.0

< 50 50-74 75-99 .i00-149 150-19S 200-299 300-399 4004

BedsSOURCE : TABLE 2

High MK: Low

IMPACT OF ADMINISTRATIVE. AND COST FACTORS ON NURSE STAFFING

Table 3.-Percentiles of nursing hours per medical surgical patient days in HAS subscriber hospitals,3 months ending 10/31/71

87

lied site5th

percentileIst

quartile Median3rd

quartile95th

percentile

Und'r 50 4.68 5.92 6.88 8.02 10.02

50-74 . 4.87 6.04 8.79 7.51 9.52

75-99 4.82 5.97 6,59 7.29 9.13

100-149 4.84 5.79 6.47 7.22 8.63

150-199 4.58 5.54 6.31 6.94 8.09

200-299 _ _ 4.76 5.51 6.04 6.82 8.18

300-399 4.61 5.55 6.11 6.70 7.93

Over 400 . 4,40 5.22 5.74 6.38 7.45

Teaching _ _ 4.53 5.67 6.36 7.23 8.98

SOURCZ: Hospital Administrative Services (HAS). prepared on request of author.

parameters and environmental factors, but still,substantial random or unexplained variance is

likely to remain.One study that relates to this question yields

essentially this finding. Ingbar, Whitney, and Taylorperformed multivaria:D analyses on nursing hoursper patient day in 72 Massachusetts communityhospitali using 1958 and 1959 data (14). At thattime, five factors measuring the type and scope ofservices explained 34 percent of the variance. Sur-gical activity was the largest, explaining 22 percent.A measure of ambulatory activity (X-rays) ex-plained 6 percent. Measures of ward patients andeducational activities explained the balance. Theuse of nursing students explained an additional20 percent of the overall variance. Forty-six per-cent of the variance remained unexplained. Thisrather large residual is not suggestive of clearconsensus.

Contrcl 7'cchnolagy

Although available data on actual nurse staffingand distribution of nurses reveal substantial varia-tions in act. .1 performance, it appears that atleast crude technologies are available that couldpermit more careful control of this resource. Pro-gressive patient care concepts of grouping patientsaccording to the severity and kind of need havebeen around for many years, and at least the in-tensive type care units have proved popular. Thesehelp in controlling severity-of-illness variance andestablishing appropriate care, but they do noteliminate tt.,: staffing problem (15). Manual meth-ods for maintenance of control of nurse staffing

96

have also been available and have been demon-strated to be feasible. Computer aided schedulingsystems have also been developed and at least oneof these has been marketed commercially (16).

Both manual and computer scheduling systemsrely upon the use of "float pools" or reassignmentof nurses to maintain desired levels of staffing.They have an additional advantage in that theyincorporate measures of their own success. Thus,where these models have been put into effect, itis possible not only to strike a bargain regardingthe appropriate amount of nurse care but also tomonitor how closely it has been achieved. Evenmanual systems are susceptible to a variety of kindsof monitoring, either of total hours delivered or ofhours per patient day delivered. The HAS systemcited above provides routine data on this questiontogether with standards of comparison for individ-ual hospitals.1 So do several regional systems ofmanpower planning and control such as CASH(10,17).

It seems reasonable to conclude that the tech-nology for monitoring and controlling nurse staffingis feasible. There is no evidence, however, that itis widely applied-say in as many as 20 percent ofU.S. community hospitals. Without extensive sur-veys it is difficult to make this conclusion defini-tively. Yet the existence of continued variation inthe very measures that would be used for controlis strongly suggestive of a lack of what one wouldcall a closely bargained outcome.

The same argument may be extended to thevarious organizational parameters. Other than in-tensive care units and coronary care units, these

' HAS, The American Hospital Association, Chicago. Illinois.

88

have not even received enough attention to heproperly validated.' Often, ill lact, the goals ofinstalling these systems, such as Service Unit Nlan-agement System, appear to he different from a goalof freeing nursing resources (18). Thus changesin organizational parameters have only infrequentlyreached the level of consensus in the bargainingprocess, and when they have, the reasons havesometimes been tangential to the direct problemof nurse staffing.

The evidence, unfortunately, is circumstantialand unsatisfactory, but on a number of fronts it

RESEARCH (.)N ,NURSE STAFFING IN II)SPITALS

tends to the conclusion that the trustees, Adminis-trators, directors, and nurses who are central tothe bargaining process do not appear to have actedin a manner that either an economic or an engi-neering approach would dictate. To return to thesailing analogy, the ship is adrift. The adminis-trative and cost factors of nurse stalling have nothad any perceptible impact. It might be desirableto explore some of the assumptions and implica-tions of the ',odel in order to identify potentialreasons that this is so.

Difficulties in the Control of Nurse Staffing

If we postulate economically rational bargainersworking within the framework of feasible tech-nology, there seems to be au identifiable set ofpotential and actual difficulties. Empirical speci-fication of these is again difficult, partially becausethe documentation of why things do not work inthe health field is quite limited (19). Such dif-ficulties as are known, however, can be classifiedas conceptual, technical, and professional.

Conceptual Difficulties

Cone eptual difficulties obscure the bargainingprocess and impede the solution because confusionexists in the bargainers's mind as to what thebargaining issues are. The most serious conceptualdifficulty in nursing appears to be the lack of cleardefinitions of nursing roles and the lack of meas-ures of quality of nursing activity. Quality, ofcourse, is closely related to utility. Although" thereare some problems in assuming that they are

synonymous, the measurement of quality clearlysharpens the understanding of utility. Conceptsand methods of measurement of quality need tobe much more clearly developed.

In medical (that is, physician) care, it is pos-

sible_ to identify structural, procedural, and out-comes measures of care. Although structural meas-ures turn out to be of limited use pod outcomesmeasures are incomplete procedural measures

exist in considerable detail, at least for the in-patient hospital portion (20). Whole systems ofmanagement and analysis, such as the Professional

Activities Study (PAS), exist and are routinelyused (21). Research in medicine tends to validatethe procedures against outcomes. A body of vali-dated procedures exists as a result of the decadesof intense effort.

Some efforts are underway at Michigan to developprocedo, ' and outcomes measures of nursing per-formance in a manner analogous to PAS.= In thisapproach, the activities and the outcomes thatclinical nursing judgment suggests are desirablecan be specified for patients classified accordingto their diagnoses. Some desirable activities wouldbe common to every patient. The list of qualityimliators compiled by CASH and developed andused by others is an example (22). Other activi-ties would be quite specific to the diagnoses, asthe attached prototype for diabetic patients slug-g6ts (see figure 3).

Preliminary work on this approach is alreadyunderway.2 The specification of disease relateditems of care that can be reliably assessed by anindependent observer appears feasible. Work thissummer will investigate whether economicallypractical surveys will yield sensitive and reliableassessments. The effort is still some years fromthe point of universal day-to-day use. It will bea costly instrument until it can be. reduced to alimited utunber of representative indicators. Moreimportant, it appears that nursing staffs will re-(wire considerable reeducation to orient themselvesto this form of purposive professional behavior,

The effort is consistent with the philosophy of

smith. II L. and Horn, B. I.., personal communication.

IMPACT OF ADMINISTRATIVE AND COST FACTORS ON NURSE STAFFING 89

Figure 3.Prototype outcomes quality assessment instrument, patient observation section, diabetes mellitus.

Patient Observations

I. Have you been giving your own insulin? YES_ NO(a) (If YES) When did you begin?(b) (If YES) How competent do you feel? VERY__ SOMEWHAT_ NOT AT ALL____(c) (It FONIEWHAT or NOT AT ALL) Why? _ _

(d) What material and aids do you have?

2. Have you been selecting your own diet?(a) (If YES) When did you begin? .

(b) (If YES) How competent do you feel?(c) (If SOMEWHAT or NOT AT ALL) Why?

YES_ NO_VERY_ SOMEWHAT_ NOT AT ALL___


3. Have your been doing your own urine testing? YES_ NO(a) (If YES) When did you begin?(b) (If YES) How competent do you feel? VERY SOMEWHAT_ NOT AT ALL(c) (If SOMEWHAT or NOT AT ALL) Why?


4. Injection observation(a) What insulin did the doctor tell you to take?

Type Strength Units Freq(b) What unit syringe do you have or plan to buy? 40 80 40/80____(c) Observation of patient's selfinjection:

Touch needle YES_ NO_Clean vial top YES__ NOInsert air in vial YES__ NOCorrect dosage YES NO If NO, specifyCleanse site YES_ NO If NO, specifyInjection: into subcutaneous tissue YES NO

If NC), adipose YES_intramuscular YES_

Post injection: Gently wipe site with sponge YES NO__Comments

(d) Observed condition of injection site:Hardness YES_ NOInflamed YES_ NOBruised YES_ NOComments _ _

5. Urine testing(a) What tests has your doctor asked you to use? Clinitest Acetest

Freq Dyestick Ketostix

Timing of testing__ _ __ TesTape(b) Describe procedure you will be doing:

Empty bladder YES_ NO--__

Figure 3. (con.)

Drink waterVoid and test

(c) Test time with respect to meal time:Prior (minutes)__ Skipped (

(d) Observation of patient's selfurinalysis:Sugar test CORRECT READING _ INCORRECT READING.__..Acetone test CORRECT READING_ _ _ INCORRE.:T READINGComments

RESEARCH ON NURSE S'IAFFING IN HOSPITALS

YES__ _YES__

NO_._NO

6. Diet knowledge

(a) What are your doctor's orders concerning your own diet?Calories__ _ CHO_ _ Protein__..__ Fats__ .

Comments

(b) Observation of patient's menu selection (see attached copy of menu)Correct calories: CFI()

ProteinFats

Comments

YES_YES___YES___

NO__ ..

NO____

(c) What materials do you have pertaining to your diet to assist you with diet selection?Exchange ___ Diet selection sheet__ ___ General food discussions_Diet record form______ Exchange instructions_ ____ Other_ ___

7. Foot care observation

(a) Cleanliness(b) Toenails (cut)(c) Dryness

(If, SOME or SEVERE) Isdays ago?DesquamationDaily use of lubricants

(d) Calluses(If SOME or SEVERE) Are they the SAME___ WORSE____ BETTER_

ago?

(If SOME or SEVERE) What arc you doing about it?

ADEQUATE_ _ NOT ADEQUATE___STRAIGHT CURVED__ LONG _

NONE SOME SEVERE _ _

this condition the SAME__ WORSE____ BETTER__ than a few

NONE ._. SOME._YES__

NONE__ SOME__

SEVERE._NO_____

SEVERE__. than a few days

8. Foot care information

For each of the followinr,, is it or is it not recommended for your feet?Wearing wool socksWearing loafersWearing elastic gartersHot water bottle

Metal arch supportsWalking barefootWearing sneakersFoot powder

SOURCE: R. L. Smith and B. j. Horn, The University of Michigan, Bureau of Hospital Administration.(This figure is a draft and should not be reproduced.)

99


clinical nurse specialitation, Imwever, and it shouldbe useful soon in at least research or nonrepetitiveapplications. In the meamime, most of the existingquality measures, such a.; large part of the CASHinstrument, have been tumid to be lacking in re-liability or sensitivity or validity..3 In the absenceof workable quality measurement systems, debateon nurse staffing takes the character of specula-tion, and a consensus from the bargaining processbecomes much less likely.

Also, there is a semitechnical, semiconceptualdifficulty in nurse staffing as to the meaning ofthe average number of hours per patient daydeemed appropriate for a group of patients ofgiven illness severity in a given set of organiza-tional parameters. In an uncontrolled situationthere is considerable daily variance in the amountof nursing care required on a given ward. This istraceable to both illness severity and census varia-tion and introduces the possibility of a significantambiguity. The appropriate level may be perceivedas that level which on the average is within theworking performance of a given set of personnel;that is, the level that can be sustained- by a workgroup under varying conditions over a reasonablylong period of time.

Alternatively, the appropriate level may be per..ceived as that amount necessary for the individualpatients on a given day. The average of the latteris likely to be smaller because it does not includeallowance for occupancy variation. If it were thechosen level, the variation would have to he con-trolled or there would be substantial quality de-cline and an increase in complaints from peoplewho are responsible for giving care. The concep-tual difficulty is that in many people's minds thetwo problems have never been separated. If oneasks a director of nurses what staffing levels shethinks are required by her patients one cannotaccept the answer at face value. It may containallowances for occupancy fluctuation.

Technical Difficulties

While feasible technology for nurse staffing exl3ts,it is far from perfect. There are two or jx;ssiblythree problems that the technology attacks, butthere are residual difficulties; in each area. Tile

'Smith, R. 1... unpublished paper.

first task is for the hospital to arrange a predictablesupply of nurses and other nursing personnel foreach day and shift and to set that level somewherenear the point where an appropriate quantity ofservice can be delivered. This is the schet4dingproblem that arranges days on, days off, and shiftassignntents for t' ;eh individual employee, subjectto wage and hour legislation, union contracts, ab-senteeism, and (Ls'ttl number of personnel report-ing. The most corttnonly reported solutions havebeen cyclic schedules that claim to permit the nurs-ing department to make individual work plansin a relatively economical manner (23).

The major difficulty with these appears to be inthe variability of human needs. Sometimes theseare expressed in the form of a request for excep-.dons from agreed upon schedules. Other times theyappear simpl) as absenteeism. The result is thatthe schedule seems always to be in tra6sition andit never reaches a steady state. The ManagementMonitoring Systems Demonstration (24) going onat Michigan indicates that for one demonstrationhospital nursing department this problem is cen-tral and disabling. It is simply not possible to pre-dict the quantity of nurse manpower that willappear on any given day, even though considerableeffort is expended by the nursing supervisory per-sonnel.

Warner of the Michigan group is working on anew approach to the scheduling problem that willmake an optimal assignment by mathematical pro-graming methods based upon nursing personnel'sown preferences as to work times (25,26). Thismay reduce some of the inherent causes of dissatis-faction with long range schedules, as well as drasti-cally reduce the amount mf middle managementtime involved

The second technical problem can be describedas an assignment problem. Given that a known orclosely predicted number of nursing personnel areavailable for distribution to a described inpatientcensus, one must assign them to nursing units ina way that maximizes the overall utility subject tocertain constraints. This problem has receivedconsiderable prior attention. As was noted above,feasible models for its solution exist, both in amanual and in a computer oriented mode.

The manual modes appear to be uneconomical.They make large requirements on personnel andrequire considerable dexterity of the operator. In

100

92

the McPherson Community Health Center, forexample, the director of nurses handled the as-

signment problems quite well. Her assistant, whowas forced to do it 2 days a week, had extremedifficulty, as evidenced by less desirable solutionsand much greater expenditures of time and effort.

A semiautomated solution, which uses the com-puter to suggest an apparently optimal solutionfor review by the responsible nursing supervisor. ap-pears to he the method of choice. One of the mostimportant technical difficulties in this approachfrom the administrative and cost viewpoint is

the specification of the objective function. This isthe element that will determine the relative utilityof various possible assignments and select the onethat makes the greatest 'contribution.

There are a number of possibilities. One thatthe Michigan group is attempting to explore wouldattempt to use the floor supervisor's own prefer-ences for additional staff. While this evaluation istoo subjective to provide much improvement overpresent practice directly, if certain more objectivepredictors of that evaluation can be determined,such as a Connor type severity of illness scale (/7),the objective measures C111 be used as a proxyutility function for day-to-day assignments.'

Even if the Michigan approach provides an im-provement in the assignment technology, anotherportion of the objective function retnains for systeniatic exploration. The assignment model assumesthat at least sonic fraction of the nursing personnelare available for assignment on any two or morenursing floors or units. It is well known that thereare some costs involved in transfers or reassign-ments. These costs appear to include attitudinalcosts on the part of the employees, training costnecessary to orient personnel for two or more workassignments, and quality costs involved in the lackof knowledge of patients for a recently transferredperson.

These costs can he minimized by reducing thevariation in number of patients and severity ofillness, which will have the 'result of reducingthe number of transfers required. Thus, stabilizingthe occupancy level of the floor and utilizing dif-ferent floors for different purposes in the progres-sive patient care manner seem to he desirable. Ifa situation existed where occupancy were reason-ably stable, patients were assigned to patient units

Trivedi. V.. dnctoral dissertation, in progress,


according to their severity, a scheduling systemresulted in predictable numbers of nursing person-nel reporting each day and quality evaluationschemes were available, it would become possibleto evaluate the tradeoffs between transfers and ap-propriate staffing. The result of this evaluationmight be that fixed staffing for floors of stabilizedcensus and illness severity is the cheapest possi-bility. In this case, the assignment model wouldbe unnecessary.

It is difficult to evaluate the severity of thetechnological difficulties. To he sure there areopportunities for improving the technical models,but the nature of the technology is such that thisis likely always to be the case, In the meantimethere is little evidence to support a conclusionthat the presently available modelsAould not con-tribute to the control of nurse staffitig.

Professional Difficulties

Day to day work on the Management MonitoringSystems Demonstration with two community hospi-tals in Michigan leads the investigation team toidentify a problem concerning the ability and un-derstanding of nursing supervision that is tooserious to leave unnoticed. The conceptual andtechnical issues involved in the problem of nursestaffing ;kre obviously not simple. It has been rarefor the project staff to encounter any significantinsight into these problems by nursing supervisionin our demonstration hospitals.

Of course one would not expect expertise ineconomics, work measurement, mathematical pro-graining, or sociological analyses of decision makingin hospitals. On the other hand one might expectat least intuitive comprehension that there probablyis an appropriate level of care for each individualpatient, that the consumer or his agents have areasonable interest in the definition of what is ap-propriate, and that individual vaiiations tendto average out when patients are taken as groups.

In each of these situations we have encounteredwhat appears to he an absence of understanding.In both of our demonstrations there has beenserious (inestion raised by both research staff andadministrative staff as to the capability of nursingsupervision to deal effectively with improved con-trol of staffing. At the very least it appears necessary

101


either to begin rather extensive and fundamentalretraining or to supplement the nursing super-visory structure with managers with a different

form of training. It is impossible to tell how wide-spread this problem is, but there is nothing tojustify optimism.

Validity of the Economic Bargaining Yodel

The implementation of the engineering modelalone is that carefully developed measurement tech-niques would yield consistent values for a givenseverity of illness, and these would be adaptedsmoothly and uniformly by the hospitals acrossthe United States. The only cause of residual varia-tion after a reasonable period of time would bevariations in severity of illness. Since individualhospitals each admit several thousand' patients eachyear and make their admissions tinder at leastroughly comparable systems, differences in theaverage severity of illness would tend to be quitesmall as would the differences in staffing patterns.This obviously is not consistent with the facts.

The economic bargaining model brings severaladditional possible explanations into play, butone would expect it to lead to uniformity, at leastamong economically similar situations. Althoughevidence for this is not fully explored, the outlookis distinctly unpromising. Short of abandoning theeconomic bargaining model (a step not to be takenlightly), the most reasonable conclusion is thathospitals are operating in the special case where noutility consensus results from the bargaining proc-ess. More specifically, the bargaining parties per-ceive the marginal utility of nurses as very indeter-minant or diffic dt to estimate.

Thus they have allowed nurse staffing to wanderin an essentially uncontrolled manner. The rea-sons for this appear to lie in the conceptual dif-ficulties, particularly the lack of any Dbjectivemeasurement of utility and in the technical andprofessional problems. If, however, either party inthe bargaining process had strong convictions, aconsensus would develop around those convictions.No such movement is perceptible. Neither nursesnor doctors, for example, have come forward withdemands as to the tequired quantity of nursestaffing,s and hospital administrators and trustees

'"The American Nurses' Association does not recommendany specific formula, ratio, or numerical concept which cotildbe applied generally. . ." American Nurses' Association,Committee on Nursing Services, "Statement on Nurse Staff

apparently have not resisted any of the numeroussolutions that seem to be reflected in the variety ofcurrent practices.

The problem presented by this conclusion is thatthere seems to be conflict with the normative valuesabout the appropriate roles of the bargainingparties. One assumes that trustees and adminis-trators should be concerned with costs, that doctorsand nurses should be concerned with quality, andthat the four of them will bargain in a more or lesseffective manner. This may be where the problemlies in the control of nurse staffing. In fact, trusteesevidence little interest in the control of hospitalcosts, often do not have the proper informationnecessary to show interest in costs (27), and cer-tainly cannot be described as having been effectivein the control of costs.

The nature of the prevailing cost reimbursementcontracts and the sharing of costs through thirdparty coverage provides little direct incentive forcost control. The record of continuing inflationspeaks for itself. So, unfortunately, do the planningdecisions of hospital trustees. There has been notendency toward improved occupancy of Americanhospitals, little activity toward mergers, and fre-quently documented cases of almost deliberatecompetition and over expansion of hospital facil-ities. These decisions are very largely within therealm of trustee authority. They imply that, atleast up until the very recent months, control ofcosts has not been a primary activity of hospitaltrustees.

A less important factor in the attitude of thebargaining parties may have been the prevailingbelief about the existence of a "nursing shortage."If all the parties approached the bargaining tablewith the attitude that no matter what solution isreached for nurse staffing it is inadequate, noserious bargaining will result. Not much has beenheard lately about the nursing shortage. Both ofthe demonstration hospitals have established

Requirements for Inpatient Health Car' Services." TheAssociation: New York, March I, 1967.


guidelines as to Cie appropriate average nursinghours per patient day and both appear to be expect-ing their nursing department to stay at or below theguideline. Other hospital administrators in south-

east Michigan have also commented that they nolonger have chronic vacancies in nursing. There isno way of knowing whether this is also true inother States.

Some Speculations on the Future

If the model and the facts are not misleading.one can infer that the question of nurse staffinghas been a nonissue in most community hospitals.The author is not willing to predict that this willchange. Although there has been a great deal ofpublic "viewing with alarm" of hospital costs, therehas been relatively little serious activity aimed attheir control. One must look to the political arenafor firm signs of such action. The evidence is mixedat the Federal level and, on the balance, negativeat the State level.

Federal action, for example, could have effec-tively forced national franchising: it could havemoved more strongly against the economic posi-tion of pathologists and radiologists: it could havepasitd the so-called Bennett Amendment.'for pro-fessional peer review or even something strongerthan the Bennett Amendment; and it could havemoved much more strongly towards prospectiverating for Medicare payments and uniform na-tional payment schemes for Medicaid. As of thiswriting none of these things have come about. Theclearest sign has been the 6 percent limit on priceincreases imposed by the Price Commission. Eventhere, there is a provision for exceptions. It will beseveral months before it is dear how firm- that_guideline is to be. The permanence of the Com-mission is also a question.

At the State level, it would appear that a smallminority of States feel the cost of hospital careto be a central political issue. After some yearsexperience in New York with the Metcalf-Mc-Closky Act. franchising or certification of need ofhospital e-pansion has been successfully introducedin only a handful of other States. Insurance com-missioners who might insist on more control ofcosts through third party payment mechanisms havegenerally not done so. Dennenberg of Pennsylvaniaand a few others have attracted publicity partiallybecause they are the exception rather than therule. Steps toward public utility type regulation ofhospitals have been minimal. Only California, for

103

example, requires that hospitals publish informa-tion about their rates and costs.

One occasionally encounters activity at the Statelevel that appears to be in the exactly oppositedirection. A bill was introduced in the Michiganlegislature in 1971 that would have effectively re-moved all control of either costs, quality, or utili-zation from Blue Cross in Michigan!' In oneversion it specified that Michigan Blue Cross wasobligated to pay audited costs of any licensedhospital. The bill was eventually passed-in a lessdrastic form at the same time as a bill to providefranchising or certification of need legislation.Fortunately, the latter act restores many consumerprotections the former destroys.

The trend overall seems in the direction of in-creased control of costs. It is easy to predict theresult from the economic bargaining model if thistrend solidifies. Increases in nurse staffing will bestrongly resisted and hospitals will begin to noticethe range of behavior within their own regions.This will 'lead administrators and trustees to raisequestions about the staffing levels in all hospitalsexcept those at the bottom of the range.

To the extent this occurs, it looks as thoughnurses will be unprepared. In the absence of aquality measure they will be forced to rely uponarguments based upon the subjective assessments ofutility and the existence of the "nursing shortage."The evidence of the range of current performancewill he used against them in the subjective qualityarguments. They may do better with the argumentsabout the nursing shortage, but it is likely thatpeople will begin to say that if the shortage hasexisted for all these many years, perhaps it isn'tas bad as we initially thought it was.

In any case, a move toward tighter cost controlwill result in considerable pressure to install andimprove the technology. Thus interest in themodest for nurse sc eduling and nurse assignment

'H. B. 4030, Michigan Legislature. 1971.


will continue and probably increase. So will in-terest in the evaluation of the organizational para-meters, including the appropriate roles for variousskill levels, supervisory structures and service unitmanagement, and related proposals. The evaluationof all these, both organizational parameters andtechnological schemes, will depend on betterquality instruments as will the defense of presentor increased levels of nurse staffing.

The conclusion seems to be that although theimpaCt of administrative and cost factors on nursestaffing is at the moment relatively insignificant,there is quitc a powerful chain reaction that mightoccur if there is a basic change in the attitudeof the American people toward the cost of hospitalcare. Given attention to this it is inevitable thatthere will be attention to nurse staffing, if forno other reason than that it is the largest single

component of hospital costs. The resulting attention will lead to calls for improved technology,better defense of present practices, and evaluationof proposals for improved organizational para-meters.

In each of these areas the central item be-

comes the development of measures of quality ofactual nursing performance. Advances on this frontwill clarify nursing roles, aid in the evaluation oforganizational parameters, permit the improvementof scheduling technology, and improve the knowl-edge of all parties to the bargaining processes.In the author's opinion it is the most critical ofthe various opportunities and is worth substantialFederal support. A 5-year multidisciplinary andmultiuniversity effort measured in millions of dol-lars each year would be entirely appropriate tonational health care needs.

104

96

References

(1) Scott, W. R. "Professionals in Hospitals:Technology and the Organization of Work."Organization Research on Health Institu.lions, Georgopoulos, B. R., (ed). Instituteof Social Research, University of Michigan,Ann Arbor, Michigan, 1972.

(2) Trites, D. K.,, et al. "Influence of NursingUnit Design on the Activity and Subjec-tive Feeling of Nursing Personnel." FinalResearch Report, Rochester, Minnesota, un-dated.Freeman, J. R. "Quantitative Criteria forHospital Inpatient Nursing Unit Design."Doctoral Dissertation, School of IndustrialEngineeting, Georgia Institute of Technol-ogy, Atlanta, Georgia, 1967.

(4) Revans, R. W. "Morale ;Ind Effectivenessof General Hospitals." Mclachlan, (ed.)Problem; and Progress in Medical Care,Oxford University Press, for the NuffieldProvincial Hospitals Trust, 1964.

Pfefferkorn, B., and Rovetta, C. A. "Ad-ministrative Cost Analysis for Nursing Serv-ice and Nursing Education." American Hos-pital Association and National League forNursing, Chicago, Illinois, 1940.

(6) --"Departmental Indicators." Hospital Ad-ministrative Services, American HospitalAssociation, Chicago, Illinois.Connor, R. J. "A Hospital Inpatient Classi-fication System." Unpublished Doctoral Dis-sertation, Johns Hopkins University, Bald-more, Maryland, 1968.

Jelinek, R. C. "Nursing, the Developmentof an Activity Model." Unpublished Doc-toral Dissertation, University of Michigan,Ann Arbor, Michigan, 1964.

Warner, D. M. "A Two Phase Model forScheduling Nursing Personnel in a Hos-pital." Unpublished Doctoral Dissertation,Tu lane University, 1971.

(10) Commission for Administrative Services inHospitals. Nursing Procedure Manual. TheCommission. Los Angeles, California, 1968.

(11) Taylor, F. W. The 'Principles of ScientificManagement. Harper and Sons, New York1911.

(3)

(5)

(7)

(8)

(9)


(12) Yett, D. E. "The Chronic 'Shortage' of Nurses.A Public Policy Dilemma." In Klarman,H., Empirical Studies in Health Econom-ics. Johns Hopkins Press, Baltimore, Mary-land, 1970. p. 357 ff.

(13) Carr, W. J., and Feldstein, P. J. "TheRelationship of Cost to Hospital Size."Inquiry, vol. IV, no. 3, September 1968, p.48 ff.

(14) Ingbar, M. L., et al. "Differences in theCost of Nursing Service: A Statistical Studyof Community Hospitals in Massachusetts."American Journal of Public Health, 56, 10(October 1966), pp. 1708-1709.

(15) Griffith, J. R., et al. The McPhersonExperiment. The Bureau of Hospital Art.ministration, University of Michigan, AnnArbor, Michigan, p. 179 ff.

(16) Medinet Corporation. "Nuriing Staff Allo-cation." 100 Galen Street, Watertown,Massachusetts. 1969.

(17) Edgecumbe, R. H. "How CASH HelpsNurses Improve Patient Care." ModernHospital, 106, May 1966. pp. 97-99.

(18) Jelink, R. C.: Munson, F.; Smith R. L.Service Unit Management: An Organiza-tional Approach to Improved Patient Care.The W. K. Kellogg Foundation, BattleCreek, Michigan, 1971.

(19) Perrow, C. "Interorganizational Research inthe Fieid of HealthA Radical Criticism."Paper prepared for the Johns Hopkins Con-ference on Interorganization Research inHealth, sponsored by DHEW, October 1972.(Mimeographed.)

(20) Donabedian, A. "Evaluation of the Qualityof Medical Care." Milbank Memorial FundQuarterly, vol. XLIV, no. 3 (July 1966).Part 2. pp. 166-206.

(21) Professional Activity Study. The elm.mission on Professional and Hospital 'Activi-ties, Ann Arbor, Michigan.

(22) c.f. Committee for Administrative Servicesin Hospitals. A Quality Control Plan forNursing, Los Angeles, California, 1965. Anumber of modifications of this approachare in existence.

105


(23) Morrish, R. A., and O'Connor, A. R."Cyclic Scheduling." Hospitals, MHA, vol.44, February 16, 1970, p. 67 ff. There are anumber of similar publications.

(24) HS 00228. "Demonstration of a HospitalManagement Monitoring System." Granr bythe National Center for Health ServicesResearch to the University of Michigan,Program and Bureau of Hospital Adminis-tration, Ann Arbor, Michigan, J. R. Griffith,Project Director, 1969-73.

(25) Warner, D. M., Bureau of Hospital Admin-istration, The University of Michigan, Ann

Arbor, Michigan, contributed paper: TheOperations Research Society of America,41st National Meeting, April 28, 1972.

(26) Sanders, J. L. "The Scheduling of NursingUnits for Small Hospitals." Contributedpaper: The Operations Research Society ofAmerica, 41st National Meeting, April 28,1972.

(271 Neuhauser; D. "The Relationship BetweenAdministrative Activities and Hospital Per-formance." Center for Health Administra-tion Studies, Research Series 28, Universityof Chicago, Chicago, Illinois, 1971.

106

DISCUSSION OF MR. GRIFFITH'S PAPER

Discussion Leaders

Miss Ruby M. MartinDirector of Nursing Services

Ohio State University HospitalColumbus, Ohio

Dr. Frank A. SloanDepartment of Economics

University of FloridaGainesville, Florida

Miss Martin

I think Mr. Griffith has left us with the im-pression that nurses spend the bulk of their timeworrying about how much they are going to bepaid in relation to all the people working underthem, and that we really care little about overallcosts. I would have to disagree with him. We arevery concerned about costs. As a group we do notspend any more time on these kinds of concerns,about what others are going to be paid in tLenursing hierarchy, than administrators spend dis-cussing the administrator hierarchy salaries. Thisis not something about which nursing alone ex-presses concern.

Dr. Aydelotte mentioned a void in the literatureshe has reviewed that deals with the administrativeand cost factors in nurse staffing. Traditionallythere has been a void regarding this in the educa-tion of professionals, both nurses and doctors. Theemphasis, of course, has been on educating theprofessional person in the clinical practice of nurs-ing or medicine, and more attention needs to begiven to the financial aspects of health care inprograms of formal preparation. Perhaps MurielPoulin would like to talk about this as far as herprogram of nursing administration at BostonUniversity is concerned.

In the HAS program, the American HospitalAssociation designed the format for the figuresthat are reported there. These figures representthe manhours paid for all those assigned to a

98

nursing unit without any thought as ..t) what theyare doing. As I understand from our personneldepartment, these are actual hours worked or paidfor without any thought of who the people areor what they do as long as they are assigned toa nursing unit. This can be a very misleadingfigure, and we ought to take that into considera-tion. All hospitals do not necessarily assign person-nel to cost centers using the same rationale.

I agree with Mr. Griffith when he says we cedto beginI would say continuean extcnsive andfundamental retraining of our nursing supeisorypeople. There are many instances where this isbeing done and being clone very well. If you look athow people have been assigned to supervisory.positioas in hospitals -traditionally, you will findthey have been chosen because they have beenthere the longest and not because they have hadany particular preparatior to assume this respon-sibility. If you look at the majority of hospitalsin our country, those around the 100-bed size, thisis still the way these people are being chosen. Weneed to provide opportunity for these people tolearn to function as managers of nursing ratherthan to criticize them for not functioning in thisrole.

In his paper Nfr. Griffith does not suggest thatwe do this or that we provide managers with adifferent form of background and training. Weare only muddying the water, and we have had

107

DISCUSSION OF MR. GRIFFITH'S PAPER

quite a lot of this with all the people added tothe structures within the hospitals, I see it evenwith the Unit \tanager System in some instances.Instead of relieving nurses of duties, it adds timeto their workload in some other way. Communi-

cation becomes more difficult when there are morepeople with whom one must communicate to gettliv job done. I ant sure there are many nursesin the room who will want to speak on this later.

Dr. Sloan

I was glad to sec that we are discussing paperson implementation at this conference. Yesterdaywe discussed the very important problems of whatshould be done or how yon could start from scratchto staff a hospital. The problem is that one candevise many schemes on paper, but things maynot work that way, which brings us to the questionof implementation.

We face parallels elsewhere in the health sector.For instance, there has been much talk aboutHMO's and how desirable they are, but they willnot be successful if producers and consumers don'tlike them. In other words, one may design some-thing very nice and find out it is not desirablefor less "objective" or "quantifiable" reasons.

I was also glad to see that you made the follow-ing considerations explicit, even though many ifnot most of us are generally aware of these matters.The first pertains to substitutions. There are sub-stitutions of various types of nursing personnelthat we might want to consider. We have RN's butwe also have other categories. We have capitalthat may be in some limited sense substitutedfor labor. We actually discussed that very brieflyyesterday.

Second, with respect to staffing ratio, hii;herratios of nurses to patients yield (holding otherfactors constant) a product that is of greater valueto patients. If one increases the ratio one gets more,but there is a question of how much more onecan get. There is no particular ratio of nursingper hour or nursing hours per patient that is

appropriate. Rather, there are a number of thesenumbers, and the question is. What are they worth?This was adequately discussed in the formal presen-tation.

This, of course, gets into the question ofassessment of what a nurse is producing. Thesame issue is found in the area of educationplanning where evaluators attempt to gauge effec-tiveness. In education thereare different success

108

criteria; there are numerous different instruments.Results 'differ in large part because instrumentsdiffer. I certainly would not like to see this kindof research conducted at the cost of understandinghow hospitals actually. behave, especially with re-gard to how they select and use inputs for theproduction of care.

The third point concerns the discussion in thepaper of different types of staffing models. Thereis an assessment of whether hospitals really behaveas if they were implementing this engineeringmodel, and this notion is rejected. Likewise thereis a consideration of an economic model wherein some sense some administrator or trustee evalu-ates the utility of the service and weighs thisagainst the cost of providing that service.

I would argue with the author about both ofthese models. I think he rejected them too quickly.He takes an unduly pessimistic st.-nce, and I shalldiscuss that in a moment.

There is the point that the third party payerhas to be considered, the point that most hospitalreimbursement (or at least a considerable percent-age) is cost-plus, and as long as you have cost-plus reimbursement there is no incentive on thepart of trustee or administrator to look at costconsiderations. There is a very widespread use ofcost-plus in the hospital industry, and, of course,it is not the only industry that uses it. Aerospaceis a good example; biomedical research is another.Hut hospitals are one example in which somebehavior that we observe reflects very much thefact that we do have cost-plus reimbursement ar-rangements. And when we look at how staffing takesplace we should remember that we do have to takethat into account. Reimbursement is not somethingto leave to the economists and staffing to thesociologists and radial units to the architects. Wemust work together.

This paper has stated that we do have to worryabout implementation, that there are sortie patterns

100

in staffing that are not explained. I think we haveto develop formal model., a hit more, ones thatare testable with data. The problem in any kindof empirical work is that if you build an inaccur-ate model or you do not build a complete model,you always end up accepting the null hypothe-sis. Here the null hypothesis is that hospitals donot behave in a systematic way with respect tostaffing. If you do not build a complete model,you are always going to have to accept tt at hy-pothesis. It is not an easy hypothesis to accept.

The author does cite some work by Ingbar andTaylor that seems to indicate that staffing patternsreflect the type of service, the case mix, and soforth. In fact, 30 to 40 percent of the variance instaffing is explained by hospital type. The greaterchallenge lies in explaining the remainder of thevariance.

For instance, we should look at hospital owner-ship. An economist colleagueI do not neces-sarily subscribe to his researchhas hypothesizedthat proprietory hospitals would he much moreconsistent in their staffing patterns than wouldvoluntary hospitals. His reasoning is that after all,proprietory hospitals are subject to some sort ofprofit motive and there exist prices of these variousinputs and they somehow maximize something andcome up with a unique set of inputs that theywant to impletnen. The voluntary hospital facesso much constraint that therefore you find allthis variation.

There are problems with that, The differencesin case mix and different kinds of patients are

109

RESEARCH ON NUltsE STAFFING IN HOSPITALS

treated differently, and I do not know that I be-lieve it, but this is the hypothesis that is stig-

gestetlthat we should be wondering about theownership in addition to the reimbursement ar-rangement.

In addition, wehave to consider that there aredifferences in consumer incomes. This is. broughtout in the paper. If you go to a county hospital inCalifornia, Los Angeles County or Cook County,there is going to be a variation in staffing andthere is going to he a variation in the consumer'sability to pay. There is nothing irrational aboutit. It may he objectionable in terms of what oursocial objectives are, but it is not irrational. Therewill also be different staffing patterns, presumably,if there are great variations in the wages differentkinds of nurses are paid. The question of differentkinds of licensure arrangements has been broughtup. There are various types of bargaining situations, in fact there are a host of factors I thinkshould he investigated. They should be investi-gated primarily so that we may have an idea ofwhat kind of staffing is desirable. We have tofind out why hospitals are not staffing the way wewant them to and determine how we might getthem to go in the right direction.

In closing, I applaud Mr. Griffith for gettinginto some very interesting questions. His modelwas not specific enough for the conclusion thatwe do not see staffing as a random or nonsystematicprocess. We should have fur'.her research in thisarea with a view toward implementing the staffingrequirements that may be desirable,

Nursing Directors as Participants

in Hospital Peer Review Boards

Dr. Gerrit WolfAssistant Professor of Administration

Sciences and PsychologyYale University

New Haven, Connecticut

Introduction

Nursing personnel comprises one of the largestdepartments in a hospital, and next to medicaldepartments, the hospital's most central. In thishighly important department, the quality of patientcare is of paramount importance. This may be dueto the training and interest on the part of nursingpersonnel and supervisors (1) as well as to de-

mands by consumers.However, the consumers and payers of health

services strain under present hospital costs anddesire some cost control. Any program in costcontainment in a hospital must involve nursingbecause of its size as a cost center. However, becausecost control is not of the highest importance to nurs-ing, any cost control program must provide ameans of satisfying nursing's needs of offeringquality patient care and service.

One wq). of trying to contain costs and maintainquality patient care is to use a combination ofthe carrot and the stick, The stick we report hereis that of peer review in connection with the useof engineering standards. The carrot, detailed

101

elsewhere,1 is that of monetary incentive for betterthan budgetary performance. These are only twoof the many possibilities, but these are of interestto this researcher because of their use of inter-personal communication (2) and decision mak-ing (3).

Let me illustrate with some problems. Place adirector of nursing on a budgetary review boardof peer administrators and what happens? Whatkind of impact does the director have on a nursingbudget and what is the effect relative to the influ-ence of "peer" administrators? How does the peerreview process operate in relation to cost contain-ment and patient care?

The peer review notion assumes that a nursingdirector is a peer of the head administrator andcontroller. The director, like all managers, must'schedule, supervise, control, plan, and encourage.What kinds of interests and skills does the nursingdirector have to do this, and does peer review

Elnicki, Richard, personal communication, 1971.

102

improve them? If the administrative knowledv in-creases, is this seen in relationships with subordi-nates and in relation to cost and performance?

Answers to these questions may vary with thesize of hospital, the social and technical opera-tion of the review board, the social and technicalenvironment of the board. and the background ofthe nurse, such as the nurse's training and expe-rience. That is to say that organizational, group,environmental, and individual variables may affectthe nursing director's performance and learningin peer review.

RESEARC:11 ON NURSE STAFFING IN HOSPITALS

We report 'here aspects of a 3.year experimentin cost containment using peer review, engineeringstandards. and monetary incentives. We explorethe above questions about the nurse in ;, peerreview setting. The important organizing distinc-tions are technical features outside and insidehoard meetings and organizationsl problems outside and inside the meetings. The next sectiondescribes the setting and design of the peer reviewexperiment. Sequel sections report results relevantto nursing directors for the 3 years of the experi-molt.

The Experimental Design

Half of the 3ti hospitals in a small State "vol-unteered" to participate in an experiment in costcontainment by peer review boards. The 18 hos-pitals not in the experiment were used as a controlgroup for economic analyses and comparisons,reported elsewhere. Some wished to partici-pate, but others were persuaded to participateas necessary for the sake of the health industry.

The 18 experimental hospitals were divided intothree setssmall. medium, and largeof six each.Small hospitals had fewer than 100 beds, mediumhospitals, 100-250 beds and large hospitals, 250-1,000 beds. Each of these three sets was divided inhalf. From each of the hospitals in this final clas-sification, peer review boards were constructed.During the first year, each hoard reviewed fourdepartmental budgets (nrsing, medical records,housekeeping, laundry-linen) of each of the otherthree hospitals of similar size. The planned lifeof the experiment was 3 years with expansion inthe last 2 years for a greater number of depart-.meets to come under review.

Setting

A meeting of a hoard took place at the hospitalunder review and lasted a hill clay. All IS reviewstook place during the last 3 months of the calendaryear, which were the first mouths of the nextfiscal year. Meeting rooms varied from a spartanhall with assembled table and chairs to plashboard rooms. Lunch and coffee break snacks werepart of the routine at all meetings,

The target hospital lined itself up along one

side of the table and each of three review hospi-tals queued up along the other three sides witheach hospital staying together as a group. Thisprocedure tended to structure the situation as acourt of inquiry. If department heads participated,they, waiting like witnesses, sat around the pe-riphery of .he room. There was no discussion ofalternative designs, such as not sitting by groupsor constructing circular tables.

Subjects

The board consisted of nine members: threeadministrators, three controllers, two nurses, and atrustee. This means that each hospital contributedan administrator and a controller and that two ofthe three hospitals contributed a director of nurs-ing and the third contributed a trustee. Most ofthe members knew each other before the experi-ment through their professions. The larger thehospital, the longer the tenure in office of eachperson and, consequently, the more known andrespected he was. Because of the division by hos-pital size, board members were still relatively peers.

Controllers usually sat next to administratorsfrom the same hospital, facilitating consultation be-tween them. This structure changed slightly inlater meetings with nurses, the least experiencedin budgeting, tending to sit together. The staffcoordinator for the group (described in a moment)acting as secretary- consultant, was located near thetarget hospital ill small and medium size hospitalmeetings, The large hospital boards found the co-

NURSING DIRECTORS AS PARTICIPANTS IN HOSPITAL. PEER REVIEW BOARDS

ordinator, acting as titanium', sitting opposite thetarget hospital.

The chairman was to he selected at each meet.ing. In the small hospital hoards, the coordinatorpreferred not to he chairman. (*OBI la IN to the

board's wishes, and an administrator was elected.The full responsibility of the chairmanship becameapparent when the coordinator felt the hoard

should accept the responsibility for informing thetarget hospital of the board's decision on the firstdepartment. This procedure apparently went a longway to create connnitment and responsibility bythis group for its decisions.

In the boards of medium size hospitals, the

coordinator took a strictly advisory role. Theboards tried functioning without a chairman atthe first meeting. However, it was apparent tothem that a chairmanless group floundered. Acontroller was elected chairman for subsequentmeetings, and this responsibility generated moreextensive preparation on his part.

The coordinator of a large hospital took therole of chairman, which satisfied his own inclina-tion to get things done and relieved the group's

anxiety about what to do. This violation of theintended functioning had its positive and nega-tive effects. The coordinator positively led withpriming questions about the budget, trying to getthe board to take an active part, while the groupnegatively became dependent on the coordina-

tor le.. leadership.

Task Environment

Each board was supplied a staff member, thecoordinator, who helped construct budgets andgather together correlative data. The budget itself,the first year, was constructed for each of fourdepartments: nursing, medical records, laundry.linen, and housekeeping. In the second year ad-ministration, pharmacy, and operation of plantwere added. For each department, salary, nonsal-ary, and depreciation figures were generated, line-by-line, according to a standardized recording sys-tem. Data sheets reported the expense budget forthe previous year, the actual results of the pre-ceding year, and the expense budget for the cam-ing year.

Correlative and supplementary data were ofthree kinds. One was quartile expense comparisons

103

with all other hospitals in the State, for each de-imminent. This report is a standard report gener-ated by a national association. The second reportwas a biography of the hospital's operation. Thisincluded kinds of nursing care and bed comple-ment, or how housekeeping handles the cleaningand janitorial jobs. The last report was a systemsengineering report on each department specifyingthe number of personnel needed to do the job ineach department. These data were generated byhaving an industrial engineer visit each depart-ment, interview various personnel, and then havethe engineering firin convert the data based onstandard formulae into personnel requirements.

Instructions

Boards were instructed to use any of the infor-mation to come up with a decision concerningthe budget. The board had the authority tochange any part of the budget they saw fit thatdid not stay within reasonable historical practicesof the hospital. The board's judgments were to bemade with the hospital under review removedfrom the meeting. Also, the target hospital didnot have the right to bargain, debate, or changethe board's decision, However, it had the rightand responsibility of informing the board of jus-tification for the submitted budget.

The board was not told how to organize otherthan that a chairman was to be elected fromthe board at each of the meetings. The boardcould use the staff coordinator as it wanted.

Incentives

A monetary reason for participating was that ahospital would be reimbursed for performing bet-ter than its budget. This reimbursement would bedistributed at the end of each year and supplied

by the Social Security Administration and Blue

Cross. However, once the experiment started, im-mediate problems of planning and learning aboutmanagement took on an autonomy of their own.The opportunity to learn was an inducement also,

although the need to learn how to control costsmay have been driven by a fear that "outsiders"might force something to be done to control costs.

These observations stemmed from interviewswith board members. Attendance at the meetings

11Z

104

yielded some further ideas on motivations afterthe experiment started. These were desires to learnand desires to look good in front of peers.

Relevant to the desire to learn...small hospitalshistorically had done little specific breakdown budg-eting. Participating forced them to budget andlearn how to budget effectively. Although thelarge hospital hail budgeted for years, its budgethad grown so that the experiment provided ameans for sorting out budgetary complexities.

Also, the turnover in administrative personnelin the experiment's 18 hospitals had produced asituation where about one-third of the memberswere relatively new to their jobs. In these cases,the experiment provided a means for learningabout their own hospital's operation as well as

getting some professional feedback to their initialdecision.

Taking a look at self-presentation motivation,we found that for those who had been in theirpositions for a considerable time the meetings pro-vided a means to show off their facilities andpersonal skills. The carpeting it the meeting rooms


of the experienced hospitals was thicker, food bet-ter, and visitation of new wings occurred moreoften.

This looking-good motive carried through inconstructing a budget. There appeared to be a de-sire to construct a tighter budget than that of asimilar hospital. This desire could conflict withthe monetary incentive, because the tighter thebudget, the less likely the possibility of receivinga monetary reward.

Data Collection

Three kinds of data were collected. 'These werequestionnaire, content analysis of audiotapes of allboard meetings, and personal observations andinterview. Questionnaires were sent to the 54 par-ticipants before and after the set of meetings eachyear. Fewer than half the participants responded.All meetings were tape recorded, and observationoccurred in more than two-thirds. Details of ques-tionnaires, coiling systems and analyses can befound elsewhere (4,5).

Results

One way of classifying the social-psychologi-cal results is to describe the social and technicalproblems inside and outside the board meetings.Where appropriate, we note differences betweendifferent size hospitals and nurses with differentbackgl.ounds. We do not report the details of theeconomic results other than to note that as of thistime, the .first year resulted in no differences be-tween the experimental and the control group hos-pitals in their actual financial performance.2

Technical Problems Outside the BoardMeetings

Before the boards could meet, budgets that werecomparable and consistent between hospitals hadto be constructed. Engineering standards needed tobe set, and rules for reimbursement were called forin sonic detail.

Because of the shortage of time and experience,these three problems were handled relatively poorly

' F.Inicki, Richard, personal communication, 1971.

the first year. Except for the small site hospitals,all hospitals had budgeted using State hospitalassociation guidelines. However, there were stilldifferences between hospitals on allocation of coststo departments. Even within a hospital, budgetaryprocedures changed annually by small degrees.

The engineers had to hustle to get data on the.frequency of microtasks performed in each depart-ment. Often the hospital personnel had to supplydata that were hard to obtain or were inaccuratelyreported. The result was that engineering stand-ards lacked credibility.

Finally, the development of rules for reimburse-ment was delayed until the end of the first yearwhen they were to be actually used. This procedureraised decision making problems for the boards,which is taken up in a subsequent section.

By the second year, the first two problems ofuniform budgets and complete engineering datahad been, for the most part, corrected. However,rules for reimbursement had not been developedand were not considered until after the secondyear's review board meetings. When they were

113

NURSING MRECTORS AS PARTICIPANTS IN tlosPITAI, PEER REVIEW BOARDS 105

considered, they were treated differently by differ-ent hospitals.The small hospitals decided to use a step vari-able procedure that called for adjustment of budg-ets to volume, not on a linear basis but on a stepbasis. For example. for every x-tousand patientdays' deviation from the projected patient days.staffings wo,.1d be adjusted by one FTE. Rewardswould then be based on a justified adjustedbudget.

The medium size group talked at length aboutusing a form of step variable but was unable tocome to agreement. The large hospitals decidedthere was no way to adjust budgets and, therefore,gave no rewards or penalties. Some thought thereshould be no adjustment for volume. The argu-ment used was that a household does not changeits income and budget for olume but insteadtightens its kit. Others thought there should beadjustments but people could not agree and saidthe problem was too complex.

The nursing department exemplifies these com-plexities. Staffing depends on the available part-time and full-time employees in different gradelevels, the number of nursing units, and the op-portunity to switch personnel from unit to uniton a daily or hourly' basis. It also depends on thevariability of patient demand. With ample part-time help and relatively continuous units, staffingcan adjust to volume if the director has the skillsto do the juggling act. On the other hand, if ahospital had a few part timers, separate units andstable demand, clear step function adjustmentscould be made. These two extremes made it diffi-cult for participants, including nursing directors,to agree on reimbursement and budget adjust-ment.

Technical Problems Within the BoardMeetings

This area includes the kind of data supplied tothe hoard, the form that they were in, and therules for how the data were to be used. The formsfor data reporting were literally hard to read andhad too much raw data without useful summarystatistics. Also, the use of data, particularly if con-tradictory, was left to the boards to work out. Theresult was that complex data were ignored, andsimple decision rule of normal historical increases

was used. This course of action was of least effort

for them.This mem that between hospitals, comparisons

were difficult to make, and within hospitals, com-parison to engineering Aandards was rejected. Thereason for the latter can be understood by lookingagain at the nursing department as an example.

The engineers constructed staffing standardsbased on national norms for microlevel tasks. Thenursing department supplied information on howoften each microlevel task had to be performed.The engineers then cranked the frequency of tasksand time per task through a formula to comeup with a standard. This was a classical indus-trial engineering approach of time and motion.

The problems with time and motion studies arewell known and these problems occurred in thissetting. The problems were as follows:

The shortness of the study gave the nursingdirector little help in determining how tochange and implement the staffing change ifthe engineering standards, called for fewernurses.

114

The time pressure to get data forced the nurs-ing employees who filled out the frequency oftask questionnaire to fill in fictitious data.

There was no one engineer regularly availableto help the nurses collect the data and answerquestions.

The final staffing standard was not explainedto nursing's satisfaction; the engineers said theformulas were too complicated and their ex-pertise would have to be trusted instead.

The standard was a single number with noconfidence interval. Neither the nurse nor thehoard could determine how far from thestandard was really acceptable or unacceptable.

In summary, had data, hidden procedures. andunaided involvement produced a relative rejec-tion of the engineering standards by the nursesand the boards. The hoards' suspicion of thestandards led them to using the standards onlywhen it was to their benefit. One nurse said, "Idon't believe the standards, but if it makes mydepartment look good, I'll push the board to usethem so that my budget is accepted."

106

Organizational Problems Within the BoardMeetings

The concepts of intergroup relationship anddecision process became clear during the first yearas useful for describing peer review. We had ex-pected that concepts such as conflict and influencebetween peers might be more important (6). Inthe first year the strong difficulties of technicalproblems inhibited these phenomena from clearlyemerging. Instead, all participants were in thesame boat of grappling for understanding. Withno one knowing the rules of the ,game, conflictcould not appear.

The decision process for the review of a partic-ular budget followed a three phase sequence ofsearch, analysis, and decision. These parts havepreviously been noted as essential components ofa group decision process (7,8,9,10). Search includeslearning, concept formation, and information gen-eration. Analysis encompasses problem solving,comparison and contrast, and evaluation. Decisioninvolves the acceptance, rejection, or change of thebudget.

The intergroup relationship was seen as onebetween a review board and a hospital. This re-lationship was viewed as consisting of the co-occurrence of strategies by board and hospital.The strategies and relationships were thought ofas varying with the search, analysis, or decisionstage of the decision process.

The strategies, as measured on the question-naire the first year, were not evident to the partic-ipants before the meetings started, Because ofthis uncertainty, the nurses met in a preliminarysession to acquaint themselves with the data andpossible selection of strategies. We observed thefollowing strategies at the meetings. During thesearch phase, the board was passive and the hos-pital made a presentation of the budget. Theanalysis stage usually found the board proceedingthrough a budget line-by-line. Also, much timewas spent on managerial discussion. The decisionphase usually saw some discussion and consensuson a decision.

The nursing budget was presented with percent-age . breakdown for subunits and changes fromprevious years. Often, background informationwas supplied by the director on the wage strut=ture and staffing patterns. Nurses on the hoardusually questioned the figures by comparing them


with their own hospitals. The general tone wasone of trying to understand the nursing depart-ment. This type of strategy was not used as much,if at all, in the other departments. This may havebeen because directors of other departments wereusually not present.

The reasons given for increased costs in nursingwere salary increases or increased staffing. The lat-ter created confusion when the hospital budgetedfor unfilled positions that it really did not expectto fill. It seemed that nurses would set budgetsaccording to an idea that was not ever possibleto attain. The board feared cutting the vacanciesin case the nurse might actually be able to hire.

There was time spent discussing the number ofnursing hours per patient day and the staffing mix.These areas seemed to be ones of hospital philos-ophy and generated much managerial discussion.The total time spent on analysis of a nursingbudget was 47 minutes on the average.

The decision phase took about 15 minutes andno budget was cut. Some budgets were tabbed fora further look at reward time for what to do withunfilled positions (vacancies). The nurses werenot as active in the decisicro phase as in the pre-vious two phases. In ratings of who contributedto the group's decision, controllers rated highest,administrators next, and nurses and trustees least.

The second year, a fixed set of strategies foreach phase was selected by the researcher formeasurement. Participants, acting in the role ofboard members and then in the role of hospitalmembers, rated their preferences before the meet-ing for the kinds of strategies preferred. The ob-servers rated the actual use of strategies. Table 1reports strategies for each phase and the inter-coder reliabilities. From these data, we look at theintergroup problem and the influence of nurses.

Across all budgets we found no differences be-tween the hoard and the hospital in their prefer-ences for joint strategies. This held up for allthree decision phases. The key finding, instead,was that during the search and analysis stagesthe participants' strategy preferences were nega-tively correlated with their actual strategies per-formed. The negative correlation was highest forcontrollers. During the decision phase the corre-lation was positive, with the nurses' preferencescorrelating highest with performance. However,the nurses were rated by their peers as havingthe lowest influence over all participants the sec-

115

NURSING ontrcroas AS PARTICIPANTS IN tiosPITAI. PEER REV1EW BOARDS 107

Table 1.-Interater reliabilities for allstrategy items

Hospital st tegyDoan! siva! egy

Learning stage1:. rept ion

Preset' ta t ion velum Nothing

Croup studyIndia i(111.11 studyNo st tidy

0.90 0.90 ($60.98 0.23 1.00

0.99 1.00 0.98

Compare and contrastLine bv.line reviewNlanagentent discussion

Accepts _

Delays

Changes

Problem solving stagetuts tta 1 Niu 1Ia I tagemen t

increases increases (list ussion_________

0.1r) 0.71 0.33

0.74 0.72 0.82

1.00 1.00 0.71

Sat isliet1

0.94

0.99

0.69

Decision stageQuest ions

1.00

0.82

0.84

Protests

1.00

0.84

1.00

Table 2.-Mean contribution and influence ratingaccording to role (N 16)

Role ContriI utiuil Influence Average

Controller 3.4 3.5 3.5

Hospital 3.3 3.3 3.3

Chairman 2.9 2.5 2.7

p 0.01

Coordinator 2.4 2.4 2.4

Administrator 2.0 2.4 2.2

Nurse 1.5 1.2 1.4

and year, averaged over all the budgets (refer totable 2).

These second year results support the resultsof the first year that, generally, the nursing direc-tors were relatively powerless during most of the

review but influential during the nursing budget.However, this could he due to the diminishedattendance of nursing directors from the small andlarge hospitals. The nurses from medium sizehospitals attended meetings regularly and werevery involved in their own budgets.

The third year saw the large hospitals droppingout of the experiment and nursing directors insmall hospitals not attending meetings, often be-cause of a turnover ill this position. Of the fourdirectors for the medium size hospitals, all at-

tended the meetings, and three of the four found

the peer review experience valuable for increas-ing awareness in how to manage a budget andthe department.

Organizational Problems Outside the BoardMeetings

In the section on experiment design, we notednonsystematically some of the organizations in-

volved in this study.: listing of the organizationsand their relationship at the beginning and dur-ing the experiment supplies a further basis forunderstanding the experimental outcomes.

The experiment was designed by the State Hos-pital Association with the advice of an engineer-ing firm and was supported by the Federal Gov-ernment and Blue Cross. Hospitals of various s'feswere involved in the design to the extent ti:ey

wished, which often meant that administrators en-couraged the idea. The professional associationsfor administrators, controllers, or nurses were notinvolved.

Because there was a short lead time in startingthe experiment, the actual participants had littlehand in the design, little time to tome to gripswith the experiment. In addition, it became ap-parent that in the large hospitals the administra-tors who pledged the hospital's participationwere too busy to participate. Also, the controll2rand nurse were nut sold on the value of the ex-periment. The large hospitals dropped out afterthe second year without administering rewards orpenalties.

During the experiment the large hospitals be-came interested in prospective reimbursment. Al-though they did not openly endorse this alterna-tive, privately they spoke of its advantages, whichwere being able to control costs and place theblame on outsiders.

The engineering firm had trouble resolving the.problems with engineering standards. This waspartly because they had vested interests in thedesign of the experiment and because they as-sumed that having been apparently successful

with the sante hospitals outside the experiment,the failures within the experiment were not theirfault. It appeared that hospitals were willing touse engineering standards in behalf of a prioriplans they wanted Co institute. Hospitals were notinterested in using engineering standards if the

116

108

engineers went counter to hosp tal plan andpolicy.

As for the nurses, they did not coalesce as agroup. Some half of the eight who had some mini-mal administrative experience and training saidthey increased their awareness of cost problems,but they could not identify any changes in thenursing department. The other half thought itwas a waste of time. The four with little ad-


ministrative experience had left the nursing direc-tor job during tl..e experiment. Their replacementswere hired with better managerial skills.

The problems the first year were blamed bythe hospital, on the experiment staff and en-gineers, by the staff on the hospitals and engineers,and by the engineers on the other two. The rela-tionship among these three groups varied, butgeneral:), it was distant to hostile.

Discussion

Looking at all four areas of technical and or-ganizational problems inside and outside theboard meeting, we might speculate on the orderof importance of the areas. This speculation re-quires an assumption. We assume that the threadlinking the areas is that of utilization and devel-opment of human resources for the goal of costcontainment. Or more simply, conscious partici-pation is the key. We have in casual order then,the external organization, the external technical,the internal technical, and the internal organiza-tion.

We suggest that the lack of participation of allrelevant parties in designing the experiment ledto low understanding in the execution of technicalproblems of budget construction and engineeringstandard estimates. This low understanding wascritical in producing problems in these areas.Nurses did not participate in design and conse-

quently fed fictitious data to the engineers becauseof low understanding that came from low partici-pation. This analysis assumes that participation isuseful both for utilizing resources, such as thecontrollers, and for developing resources, such asthe nurses.

These technical problems were carried into theboard meeting with bad data, few ways to processthe data, and rejection of engineering standards.At this point, only awareness of the problems anddevelopment of ways to overcome them wouldhave worked. But neither the board's chairmannor the nurses, among others, showed skills atorganizing, such as listing goals and search, andanalysis and decision strategies to accomplish thegoals. The idea of participation becomes furtherrefined because of the boards. Participation with-out awareness of alternatives can be somethingeven less than muddling through (11).

NURSING DIRECTORS AS PARTICIPANTS IN IIOSPITAI. PEER REVIEW BOARDS

References

(1) Argyris, Chris. Human Relations in theNursing Division. Labor and ManagementCenter Report, Yale University, 1955.

(2) Wolf, Gerrit. "A Model of Conversation."Comparative Group Studies, 1970, 1, 275-305.

(3) Wolf, Gerrit. ''Evaluation Within Inter-personal Systems," General Systems, (inpress).

(4) Wolf, Gerrit. Behavior in Decision MakingGroups: Budget Approval Boards: 1969.

Technical Report No. 33, Yale University,Department of Administrative Sciences,New Haven, Connecticut, 1970.

(5) Wolf, Gerrit. Behavioral Aspects of theConnecticut Hospital Experiment 1970-71.Technical Report No. 49, Yale University,Department of Administrative Sciences,New Haven, Connecticut, 1971.

109

(6) Bather, James. Power in Committees: AnExperiment in the Governmental Process.Chicago, Rand McNally, 1966,Bales, Robert, and Strodbeck, Fred. "Phasesin Group Problem Solving." Journal of Ab-

normal and Social Psychology, 1951, 46, 485 -495.

Davis, James. Group Performance. Reading,Massachusetts, Addison-Wesley, 1969.

Simon, Herbert. Models of Man. New York,Wiley, 1957.

(7)

(8)

(9)

(10) Weick, Karl. The Social Psychology of Or-ganizing. Reading, Nlassachusetts, Addison-Wesley, 1969. .

(11) Lindblom, Charles E. "The Science ofMuddling Through." Public Administra-tion Review. 1959, 19,79-88.

118

Social Psychological Factors Relating

to the Employment of Nurses

Dr. Virginia S. ClelandProfessor of Nursing

Wayne State UniversityDetroit, Michigan

Mary Lee SulkowskiGraduate Student, Psychiatric Nursing


Introduction

The general purpose of this paper is to examinefactors that are antecedent or contiguous to the

decision of a nurse to be employed. The decision

to participate in the occupation is of great signifi-cance to nursing and to the public. Not only hasthe occupation of nursing been unsuccessful inproducing adequate numbers of professional nursesbut, also, large numbers of the best prepared mem-bers are utilized in educating nurses who neverwork after their first child is born.

With greater knowledge of the nature of thedifferences between nurses who work only duringperiods of financial exigencies versus those whoactively seek long term professional careers, nurs-

ing school faculties might attempt to select students

who are more likely to work for extended periods.

Also, with better u9derstanding of the culturaldetermination of certain values held by the stud-

ents, it would be possible to modify these through

the process of socialization into the profession.

The general theme of this paper is related to aview of the traditional role of women and the

119

deleterious effect this has had upon the develop-ment of women and thereby the development ofthe profession itself. Consideration will be givento alternative ways to view "women's place" andthe manner in which nursing can align itself withsome of the current social and psychological forcesthat are redefining the roles of men and women.There are economic and political forces that cansuppott an expansion of nursing's contribution tothe health services of this Nation as the professiondocuments that it can assume these new responsi-

bilities.In 1975, one hundred years will have passed

since the first schools of nursing were establishedin the United States. However, the next 10 years

will afford nursing its most crucial tests. Can theprofession adapt and grow in the face of socialchanges that will require that it modify its verynature? The new, expanding clinical roles in

nursing demand career continuity, independenceof judgment, and initiation of action that are notcongruent with the feminine image of employ.meat inactivity, dependency, and passivity.

112RESEARCH ON NURSE STAFFING IN HOSPITALS

Psychological Development

A brief review of selected aspects of the psycho-social development of identity in women and numfollows. Bardwick's Psychology of Women (1971)provides a useful frame of reference, as doesBranden's The Psychology of Self-Esteem (1971).The Adolescent Experience (Douvan and Adel-son, 1966) and extensive study of adolescence pro-vides supportive data.

Bardwick (1) has reviewed the most influentialformulations of personality development. Al-though she describes Freud's genius as "unassail-able," she rejects his ideas about the psychologyof women as reflecting political sentiments ratherthan scientific investigation. Freud saw a maledominated culture and believed that the penis wasthe symbolic source of this power. Karen Horneyrejected the idea that women suffered from "penisenvy." Horney believed that the female's capabilityof creating another human being left men withprofound envy that could only be resolved byseizing cultural and political power. Bardwickviews this as a reaction-formation idea and rejectsmost of it. The third point of view, currentlypopular, is that there are no real differences be-tween the sexes that extend beyond the biological,and thus all differences between the sexes havemerely been imposed by a masculine dominatedsociety. Bardwick rejects this notion also. Bardwicktakes the position that there are fundamental bio-logical differences between the sexes but that tra-ditional role divisions are too restrictive and thatcultural changes in sex roles are necessary.

It is unlikely that the nature of biological dif-ferences upon cognition, for example, can be ac-curately assessed until cultural restraints have beenremoved. The general psychosocial development ofboys and girls is outlined below.

ChildhoodA Sense of Self

Bardwick states "the basic sense of self, as aseparate entity, probably begins to develop aroundthe age of 2, when a child begins referring tohimself as 'I.' " Previously dependent on his par-ents for emotional and physical well-being, hissense of self develops to a great extent through hisinteractions with his parents and is dependent forits well-being on parental approval, Equally prom-

inent at this stage are the developmental tasks ofthe childthe cognitive, physical, verbal, and inter-personal skills that the child seeks to master. Suc-cess in this area receives parental approval, and thechild recognizes a new source of enhancing hiswell-being.

140

'Sexual IdentityThe child seeks a stable self definition and one

of the "givens" he can readily grasp is his sex,which he begins to recognize at the age of 5. Atapproximately the same age the boy begins re-ceiving less parental approval for dependent be-havior; likewise, emphasis is increasingly placed onhis mastery of skills. As parental approval becomesa less reliable sovrce of well-being, as independ-ence and skill mastery are stressed, he begins torely less on others and more on self for his feelingof well-being. He comes to understand that mas-culine identity (masculinity) correlates with in-dependence and achievement and adjusts his be-havior accordingly. He develops internal criteriaby which to judge and reward himself and, cor-respondingly, internal means of control. Bardwickbelieves masculinity becomes a crucial issuc for theboy at age 5, that he uses the latency years toformulate his concept of masculinity and comes toa healthy adolescence with a fairly stable concep-tion.

In line with the cultural definition of femininitythe girl is allowed, in fact encouraged, to remaindependent. While achievement is important to herself concept, she continues to depend primarily onothers for rewards and controls, in effect for herself definition. Thus her emphasis is on the inter-personal; others rather than self remain her mosteffective source of well-being. Her skill in dealingwith others (interpersonal' skill) becomes impor-tant. Our culture is more tolerant and indulgentof girls. Accordingly, there is less emphasis duringthe latency years on femininity; "tomboys" are al-lowed. Hardwick believes that feminine identitydoes not become a crucial issue until adolescence,

Self-Esteem

The child's efforts to gain a sense of self and afeeling of well-being refer to the dual processes of

SOCIAL. PSYCHOLOGICAL. FACTORS RELATING TO EMPLOYMENT OF NURSES

achieving. identity and esteem. Branden definesself-esteem as the "integrated stun of self.confidenceand self-respect. It is the conviction that one iscorn petent to live and worth.) of living (2).

As the boy develops, his self esteem increasinglycomes to depend less on others, more on self. Thestress on the independent securing of satisfactionand control foreshadows the development of au-tonomy in adolescence. But the dependence fos-tered in the girl makes "the other" crucial to herself-esteem. Bardwick states: "While the interper-sonal rewards remain salient for girls, internalizedcriteria of achievement and personal satisfactionbecome more important for boys yr

AdolescenceIdentity

Branden defines personal identity as "the senseof being a clearly defined psychological entity (2)."Douvan and Adelson note that although identitydoes not begin in adolescence, it becomes criticalat this time:

During this period, the youngster mist synthe-size earlier identifications with personal quali-ties and relate them to social opportunities andsocial ideals. Who the child is to be will beinfluenced (and in some cases determined) bywhat the environment permits and encourages(3).The key terms in male adolescent development

are "the erotic, autonomy (assertiveness, independ-ence, achievement), and identity," the key termsfor the female"the erotic, the interpersonal, andidentity (3)." The one discrepancy in termsautonomy vs. the interpersonalshades the entirecourse of adolescent development. The identityproblem for the girl revolves around "the inter-personal," for the boy "autonomy" is the centralissue:

Identity is for the boy a matter u: individuatinginternal bases for action and defending theseagainst domination by others . . . (A. the girl)it is a process of finding and defining the internaland individual through attachment to others(3).This feminine dependence on "the other" leads

Douvan and Adelson to question the validity andrelevance of the concept "identity crisis" to theadolescent girl. In her early years, the girl learnsto depend on others for her self-definition and her

113

self-esteem. In adolescence, with cultural, parentaland peer pressure focusing on the importance of

femininity as preparation for the wife-mother role,heterosexual affiliation becomes crucial: Douvanand Adelson believe the adolescent girl does nothave the time, the energy or the motivation tofocus on identity resolution. Popularity, especiallywith the opposite sex, absorbs most of the energy.Further, "too sharp a self-definition and too fullan investment in a unique personal integrationare not considered highly feminine (3)," and mighttherefore reduce eligibility.

Marital Status and Female Identity

The girl's identity is bound up not so much inwhat she is but in what her husband will be.Someone has spoken of marriage as a "mutualmobility bed." We may add that for the girl itis equally an identity bed (3).Bardwick arrives at the same conclusion:The only way to achieve a feminine sense ofidentity, if one has internalized the generalnorms, is to succeed in the roles of wife, help-mate, and mother. . . . As a man's self-esteemis linked to his appraisal of his masculinity, ap-parently a woman's self-esteem is similarlylinked to her feelings of feminity. The differencebetween them is that the goal of esteem andidentity is achieved much later in life by womenand is primarily achieved in the traditional in-tense and important relationships rather than inoccupational achievement (I).The little girl's potential wife-mother role deter-

mines the pattern of her socialization. Although itis allowed that she may at some time work outsidethe home, the focus is her traditional sex role andthe "feminine" qualities of dependence, deference,nurturance, and passivity. She has been success-fully socialized when she assumes the socially ap-proved status of "Mrs." Mead has written, "Successfor a woman means success in finding and keepinga husband (1)."

Although it is assumed that the little boy willbe a father, his socialization focuses on hisfuture occupational role. He is taught to be"masculine"independent, aggressive, competitive,achievement oriented.

Marriage provides a socially approved arrange-ment for bearing, rearing, and caring for children.

121.

114

But marriage also "offers social recognitions, so-cial place, and designated roles and thus gives itsmembers a, vital form of social security (5),"While this statement has implications for bothsexes. it has special and essential implications forwomen, whose primary role is wife-mother, whose"place- is marriage and motherhood.

A recent article in Cosmopolitan opens with thefollowing statements:

As an unmarried woman, I am suffering a pro-found identification crisis. I am a spinster.Spinster is, quite simply, the legal description ofthe unmarried adult female. Why should theunmarried adult male, or bachelor, be a figureof glamor, and the unmarried woman, or spin-ster, a pitiful lump (6)?Although the psychological implications of

being single are not well explored in the scientificliterature, Sakol focuses on two concepts pertintMto an understanding of single women. The first,"identification crisis," is self-evident. The second,lack of self-esteem., is implied in the phrase "piti-ful lump."

The unmarried woman does not assume thetraditional feminine role and thereby exists on thefringes of the social structure. She may work andcontribute to society through her work, but it is

not quite acceptable for her to define herself interms of an occupational rolethat is a man's roleand calls for "masculine" behavior. A woman maybe a sister, a daughter, a community participator,but these are not primary roles that readily lead toself-definition.

There is no acceptable status setting name forthe unmarried woman . . . except for the des-ignation "unmarried," which is a designation bysocial default, as it were, not of a social status.The woman who remains single is ambiguouslyplaced in the social structure (5).By defining woman in terms of marriage and

motherhood, society essentially limits woman'ssource of status, recognition, and security to thewife-mother role. By socializing little girls to de-pendency and passivity and a focus on the inter-personal, society effectively circumscribes woman'ssource of identity and esteem around husband andchildren.

The female's dependence on others for self-definition leaves all women vulnerablethe mar-ried and the unmarried. The married woman is


too dependent upon husband and children for herself definition. The belief that ir.otherhood is a

drive, anchored deeply in the biological sphere,leaves the woman who remains single to be viewedas abnormal and deviant. This may cause the sin-gle woman to question her normalcy. Having in-ternalized the cultural definition of femininity, shequestions her feminine idertity. Unlike the man.who has developed internal standards and criteriafor judging and rewarding himself, the singlewoman has learned to define and value herself asshe is defined and valued by others. Society neitheradequately defines nor values singleness.

Through processes of definition and social rein-forcement the essentially broad patterns of nor-malcy in sex roles have become rigid sex stereo-types. Once the characteristics of the problem areidentified, modification of the process of develop-ment generally becomes possible by altering thesocial influences.

Toward Resolution

Women's ambitions are not fixedeither bio-logically or culturally. They are subject tochange in response to circumstances. . . . Thereare crises, or critical transitional experienceswhich stimulate a reorganization of values andpatterns of aspiration and gratification (7).One can question the health of a society that so

narrowly constricts a woman's source of identityand esteem to one primary role. Mannes credits themass media for women's obsession "with an idealof femininity as the guarantee of happiness (8)."Bettelheim feels our view of women is "psycho-logically immature (9)."

Bardwick reflects on the awareness of parentswho encourage their daughters "tr ew the love ofa man and the raising of child, en as the self-defining, self-esteeming achievement, limitingachievement to the traditional asks and the needto affiliate (1)." She feels it i, 'dangerous for awoman's sense of worth to be enormously depend-ent upon her husband's reaction to her and to hercontribution to his welfare (1)."

In the process of resolution, women may con-sider Branden's examination of the psychology ofself-esteem, contrasting his statements with the cur-rent cultural definition of femininity. Brandenstates:

SOCIAL. PSYCHOLOGICAL. FACTORS RELATING TO EMPLOYMENT OF NURSES

A strong sense of personal identity is the productof two things: a policy of independent thinkingand the possession of an integrated set of valueS(2).

Further in his discussion:A psychologically health man does not dependon others for his self-esteem: he expects othersto perceive his value, not to create it (2).

Branden briefly mentions:

. . . the incalculable damage that has beenwrought by the conventional view that the pur-suit of a productive career is an exclusivelymasculine prerogative, and that women shouldnot aspire to any role or function other thanthat of wife and mother. A woman's psycholog-ical well-being requires that she be engaged in along range career: she is not some sort of secondclass citizen, for whom mental passivity and de-pendence are a natural condition (2).

Before she car find identity and esteem in herwork the woman must have new values that ensuefrom a fresh, healthier self-conception. Brandenbelievls "the cause of authentic self-esteem is . . .

the rational, reality directed character of mind'sthinking processes (2). He designates "five inter-connected areas that produce enjoyment of life:productive work, human relationships, recreation,art and sex (2)." While one may disagree with hiscategories, it is worthwhile to note he does notrefer specifically to marriage and motherhood.

To take another perspective, Baker notes thatmarriage and parenthood arc not a man's "primaryavenue to personal identity" . . . nor is his fulfill-

115

men t as it human being dependent on "the bio-logical process of fathering a child (10)."

Nfead states the problem:Every known society creates and maintains arti-ficial occupational divisions and personality ex-pectations for each sex that limit the humanityof the other sex (4).What needs to be questioned is not a woman's

desire to marry and have children but rather thefact that her self-definition primarily Focuses onher biological role. It is time to question the rele-vancy and desirability of archaic conceptions of"masculine" and "feminine" that arbitrarily limitone's personal potential to biology.

What is needed is a more human definition ofmale and fethale. In a recent article. Ulmer dis-cusses three sex role systems. In the "patriarchalsystem," the male is the provider and thereforesuperior. This system may be efficient within thefamily but Limier feels it has negative COH5C-

Times for society. The "Complementary sex rolesystem," although claiming men and women areequal, retains an emphasis on basic differences be-tween male and female. The "human role system"acknowledges the differences between the sexes butsees them as secondary (II). This system proposesthe possibility of balancing male and female sexroles and working toward a new "human role." Ashuman beings we, ti reality, play many roles. 'Cofocus so exclusively on one limits our humanness.A society threatened with extinction may need toimpose such limits. Our society not only needs therich complexity of. individual human potential,but it can not really afford to do without it.

Implications for Nursing in Resolving Rigid Sex Role Conformity

Career Choice

In the childhood years a girl is encouraged toachieve in school. But adolescence brings a cul-tural emphasis on Femininity and feminine be-havior. Passivity, dependency, and deference arequalities that conflict with those that spell aca-demic success; for example. competition, aggres-siveness, achievement. Commenting on the girl'seducational dilemma, Meade noted that girls areexpected to display enough ability to be "successful, but not too successful; enough ability to get

and keep a job but without the sort of commit-ment that will make her either too successful orunwilling to give up the job entirely For marriageand motherhood (4)." A recent study by Constan-tinople reveals that college reduces conflict for mensince it equips them for occupational success butincreases conflict for women. The emphasis oncompetition and occupational competence andachievement does not correspond to the feminineideal (12).

The natural congruence between the traditionalrole of women and the occupation of nursing has

1a

116

made it easy to recruit girls for nursing. Whennursing programs were relatively inexpensive andstudents 11-1 paid, in tuit:on and service, for theireducation ny the time they graduated, the factthat many chose not to be employed was not aserious problem. Nursing has continued to con-done this practice when it can no longer be de-fended.

Vaillot has written that the committed nurse isone who views nursing as a means of achievingbeing, as a vital way of expressing and defining theself. Although the subjects of her study overwhelm-ingly ranked marriage and motherhood as pri-mary, she postulated that this does not precludecommitment because students see marriage andmotherhood as a natural extension of their com-mitment to nursing (13). Douvan and Adelsondemonstrate that girls who choose feminine occu-pations such as nursing view these occupations as

extentions of the traditional feminine role and asa means of expressing feminine qualities (3). Ifnursing is an extension of the traditional femininerole, if the primary commitment is to marriage andmotherhood, commitment to nursing is secondaryand temporary.

Extrapolating from tables in Facts about Nurs-ing, 1969, and liA"s 1966 Inventory (the mostrecent national survey of registered nurses), it canbe estimated that out of every 100 ?flurried gradu-ate nurses 31.5 are inactive and not registered, 31.5are inactive but registered, and 37 are employed.However, of these 37 nurses who are employed,10 work only part time. Thus, out of every 100graduates from schools of nursing who marry, only27 nurses are employed full time at any one time(14,15).

Nurses, individually and collectively, must ridthemselves of rigid sex stereotyped thinking.Women who happen to be nurses must learn to seethemselves as Iminan beings who have a contribu-tion to make and who have an identity greaterthan thci feminiue selves. Nursing as a humanservice is so tied to the nurturant role of womenthat nursing's professional role can expand only asthe societal role for women expands. Nursing as anoccupation and nurses as women must seek newdefinitions and new roles. Tile two processes musthe concurrent and mutually supportive.

To suggest that nursing be made acceptable tomales is to suggest that female nurses make them-


selves professionally attrative to males. This is thetraditional approach that women have always used,11 that were to happen, nursing would merely re-semble more closely the professions of teaching,social work, and library science wherein womenare dominant numerically but men hold all of thepower positions. The image of nursing will notbe altered simply by recruiting more men into thefeminine occupation. Rather women who arenurses must free themselves from the narrow normsof lemininit-

To conforn. to the mean of the norm leads todull, unimaginative, steeotyped behavior. Theimportant ial changes are brought about bypersons exploring the outermost edges of therange of behavior, which in turn is defined byone's personal system of values. Women who arealive and growing will develop new insights andvalues. This involves granting the individualnurse the right to be ditTerent without beinglabeled as deviant. Female nurses can be helped tolive their full human potential so that nursing as

an occupation will change and become a field at-.tractive to both men and women.

Astin, Stmiewick, and Divech, in their over-view of studies on career determinants, have re-ported that there is general agreement that womenwith career orientations tend to be high achieversand in aptitude testing show high need for achieve-ment and higher levels of endurance, introspection,and independence. These same tests tend to de-scribe career orients women as having a mascu-line orientation. The authors conclude,

11. unfortunately the present clinical inter-pretation of a person who has interests and/ortalents in areas that are descriptive of the op-posite sex is that such interests and aptitudesreflect one's lack of acceptance of his/her sex-uality and that the person is in some kind ofpsychological conflict. This interpretation war-rants reconsideration in light of evidence thatgreat numbers of women desire careers that havetraditionally been viewed as careers for men(16)."

The entire practice of classifying occupations asfeminine or masculine needs to be restudied. It is

!gised upon cultural biases and produces poorscience. The Strong Vocational Interest Blankshave been cited for their sex discrimination. It isnot only that certain score configurations for

124

SOCIAL PSYCHOLOGICAL. FACTORS RELATING TO EMPLOYMENT OF NURSES 117

women will suggest psychiatric social work and formen, psychiatry, but that the portrayal of thewomen's vocations are biased and restrictive (17).There is need to develop new vocational aptitudeinstruments that reflect potential for nursing ca-reers of the 1970's and 1980's rather than the1930's.

Career Without Marriage

A generation ago women often chose betweenmarriage and a career, or at least, these wereseen as the principal alternatives. . . . Thechoice for career in place of marriage is prob-ably no longer a consideration of young women.The new decision is for marriage alone, or mar-riage with career (18).The single nurse internalizes the cultural defi-

nition of femininity and desires the socially ap-proved stunts of marriage and motherhood. Copingwith her "failure" to achieve woman's traditionalrole, the single nurse under 30 faces a crisis ofidentity and esteem. The dilemma delays, if nottotally precludes, a serious commitment to nursing.

In 1966, there were over 150.000 unmarriednurses in the United States. Of those under 30,95 percent were actively employed in the pro-fession compared to 64 percent of married nursesin a comparable age range (11). Unencumberedby the responsibilities of marriage and mother-hood, is the single nurse more likely to concen-trate full energy on her nursing career? Is sEelikely to possess a higher level of career commit-ment? Or does failure to achieve the traditionalrole pose problems that preclude authentic com-mitment?

There are few studies of single women. Free-man, Levine, and Reeder note that historicallymiddle aged unmarried nurses have been the mostcommitted to the profession (19). In a 1968 study,Baker found single women achieve "adjustmentand fulfillment through creative contributions tosociety" but the mean age of his subjects is 51(10). Although it is assumed that women whopass age 30 without marrying review their lifegoals and expectations and readjust their self-image, there have been no studies of how this isachieved. Similarly, there has been little explora-tion of the dynamics of "being unmarried" for the21 to 29 age group. Davis and Olson found nurs-

ing students primarily oriented to the traditionalrole, and they view this orientation as precludingserious career commitment (20). Mayes, Schultz,and Pierce considered student satisfaction withnursing as it influences commitment. They foundstudents consider the career vs. marriage conflict asource of dissatisfaction and interpret this as anobstacle to commitment (21). Freeman, et al.,document studies that identify marriage as theprimary goal of nursing students, a goal they be-lieve prevents many from a serious commitment tonursing.

The prospect of marriage and children per-meates every aspect of nursing; no aspect of theprofession can be completely understood apartfrom the influence of marriage plans and theirfrustration (19).Nursing can be a viable means of achieving

"being" and self-definition. However, as long asthe traditional role model predominates in thesocial and psychological development of youngwomen, women will be incapable of achieving"being" through nursing. Their commitmer,ts willbe half hearted. Nursing has never explored thenature of the problem whereby its services areprovided by women who essentially remain alooffrom a deep occupational commitment.

125

Marriage and Career

The research interest of the senior author hasfocused upon the decision of married nurses toengage in employment. The Barnard-Simon the-ory of organizational equilbrium provides a rela-tional explanation of the high level of job turn-over and underemployment within nursing. Oneof the central postulates of the Barnard-Simontheory states, "Each participant will continue inan organization only as long as the inducementsoffered him are as great as or greater (measuredin terms of his values and in terms of the alterna-tives open to him) than the contributions he isasked to make." Another postulate of the sametheory states, "An organization is 'solvent' and willcontinue in existence only so long as the contribu-tions are sufficient to provide inducements in largeenough measure to draw forth these contribu-tions (22)."

To a very great extent nursing is dependent up-on the employment of married nurses with child-

118

ren in order to keep the vocation "solvent." Yetthese are the very nurses for whom there are thefewest employment inducements and the greatestpersonal contributions associated with employ-met, Simultaneously, the ready availability of thehousewife role has made it easy to leave the fieldwhenever the expected contributions appear to hetoo great and the inducements insufficient.

From a research point of view, the problem hasbeen to make the theory operational; that is, tobe able to measure the contributions and induce-ments as perceived by the individual nurse in orderto predict when and under what conditions shewill move toward labor force participation. Thewriter has been the investigator for two projectsin this field of research (NU 002 Hi-Decision toReactivate Nursing Career and NIH 70 4128-

Recruitment, ti li/ation, and Retention of Part-time Nurses). In the former study, use was madeof a Consequence Questionnaire developed by'Rosen and modeled after Vroom's cognitive modelof motivation (23).

This instrument was designed to measure theexpectancy ami valence of the consequence of achange in work status. The nurse listed the ad-"antages and disadvantages of changing her workstatus (inactive to active or active to inactive)and attached to each an estimate of the probabil.by that the consequence would occur. The totalconsequence score was determined by the alge-braic sum of the positive and negative consu-quences alter earl: was multiplied by its probabil-ity. 'Ellis instrument did not measure the theory'sconcepts adequately and thus was not a good pre-dictor of changes in employment status, However,the listing of advantages and disadvantages whengrouped by demographic variables such as age,education, and employment status resulted in anexcellent reservoir of descriptive data.

From that study the writer concluded that "em-ployment of married nurses can most often becharacteri/ed by the principle of immediate grati-fication, namely:

rapid job turnover and change of work sta-tus when family problems arise,acceptance of dead end jobs that are compati-ble with family needs and demands,work goals that focus upon maintenance of

I

126


existing skills rather than upon advancementor growth (24)."

The goal of this research has been to explainor predict the employment behavior of marriednurses. This has seemed appropriate since theseare the women who have an option toexercise (25).

The writer's previous research (26) supportsthat of the other investigators, that financial needand absence of young children are the best predic-tors of which women will choose to be actively em-ployed (27,28,29). However, there are large numbersof nurses who choose to work in the absence offinancial need and before children are grown.Knowledge of the differences between nurses whoseek employment under these conditions versusthose who do not could provide valuable insightfor the maintenance of a nursing work force.

Two of the variables hypothesized to have aneffect would be the strength of the self-concept ofthe nurse and the family ideology subscribed toby the couple. The influence of the self-concepthas been discussed. Whether the family ideology(husband-wife power relationships) is "pa-triarchal," "complementary" or "equalitarian"could be predicted to influence the extensivenessof the work role involvement of the wife.

Under the patriarchal system, employment ofthe wife would not be expected except in in-stances of financial distress. if the couple subscribeto a complementary ideology there would be em-ployment of the wife if she wished to he employedand if the work role could be made compatibleto her family role. Vacillating conditions in thehome or in employment would be expected to pro-duce irregular employment participation. Finally,the equalitarian sex role system could be expectedto correlate most highly with serious long-termcareer involvement. The equalitarian marriagew .yid afford the housewife-careerist the greatestpersonal support for the full development of herprofessional potential plus the greatest sharing ofhome and family responsibilities. Not all adjust-ments for home or family complications should heimposed upon the wife's employer. Increasinglysuch occurrences will he classified as parentalrather than maternal responsibilities within equal-itarian families.

SOCIAL. PSYC,i0LOGICAL FACTORS RELATING TO EMPLOYMENT OF NURSES 119

Conclusion

The common delay of a serious commitment tonursing until the issue of marriage has been re-solved causes many female nurses to waste, pro-fessionally, those age years between 21 and 30.These are the same years in which society dictatesthat young men will become vocationally estab-lished. Nurses under 30 years of age should be.come an important concern of the profession forit is the manner in which these women resolvetheir own identity (within or outside of marriage)and establish some sense of self-esteem and self.worth that will determine their long term involve.ment and contribution to nursing. The womanwho views herself as a personal failure can neverbe a professional success. There is some evidenceto suggest that women who marry after the age of25 are more likely to combine career and marriage.It is probable that the woman's identity as a per-son, independent of any spouse or child, is suffi-ciently developed by this age for career involve-ment to be viewed as personally satisfying andrewarding.

Certainly there is evidence from the feministmovement that identity can be significantly in.fluenced long. after adolescence. "Rap" sessions inschools of nursing might well be used to helpstudents (graduate or undergraduate) perceivetheir options for achieving full human potential,including or excluding the roles of wife and

1..47

mother. There is need for controlled studies ofthe effectiveness of such socialization.

The passivity, deference, and dependency inherent in femininity and inherent in nursing limitprofessional potential. Woman's need for affilia-tion and approval makes her susceptible to others,whether she is seduced by sex or authority.Energy directed towards securing esteem from"others," be they medical or administrative per-sonnel, diminishes energy to be directed towardchange. Acceptance of the status quo results. Toobtain a clearly defined professional identity,nurses must develop an caupational orientation.Traditionally, women have not defined themse1vesin a work role. Nursing independence requires anachievement orientation along with the qualitiesof aggressiveness, assertiveness, and autonomyqualities that threaten feminine identity.

Nursing needs to question and challenge a tra-ditional role concept that limits a woman's poten-tial, creativity, and humanity. New deF---' ions,new alternatives need to be explored. As lung aswoman's primary definition is wife and mother,nursing will be primarily defined as a feminine,nurturing, mothering profession. An emphasis onthe "human role" would allow women and men,within nursing, to seek broader self-definitions,untapped avenues of participation in society, andimproved ways of delivering health care.

120

References

(1) Bardwick, J. M. Psychology of Women. (NewYork: Harper and Row, 1971), pp. 5 -16,p. 175, p. 189, p. 188, p. 151.

Braude!), N. The Psychology of Self-Esteem.(New York: Bantani Books, 1971), p. 110,p. 173, p. 174, p. 204, p. 131, p. 133.

Douvan, E., and Adelson, J. The AdolescentExperience. (New York: John Wiley andSons, Inc., 1966), p. 157, p. 347, p. 348,p. 349, p. 18.

(4) Mead, Margaret. Male and Female. (NewYork: Wm. Morrow and Co., 1949), p. 323,p. 372, p. 320.

Perlman, H. H. Persona. (Chicago: Universityof Chicago Press, 1968), p. 102, p. 100.Sakol, J. "The Case for Spinsterhood: Bet-ter Dead than Wed?" Cosmopolitan, 1971,171 (6), pp. 92-100.

Fogarty, M. P., Rapoport, R., and Rapo-port, R. N. Sex, Career, and Family. Lon-don: Allen and Unwin Ltd., 1971, p. 197.Mannes, M. "The Problems of CreativeWomen." In S. M. Farber and R. H. Wilson(eds.), The Potential of Women, (New York:McGraw-Hill, 1963), p. 123.

Bettelheim, B. "Growing Up Female." Har-per's Magazine, 1962, 225; pp. 120-128.

(10) Baker, Jr. L. "The Personal and Social Ad-justment of the Never-Married Women."ournal of Marriage and the Family, 1968,

30, p. 478, p. 437.(11) Linner, S. "What Does Equality Between

Sexes Imply?" American Journal of Ortho-psychiatry, 1971, 41, pp. 747-756.

(12) Constantinople, A. "An Eriksonian Measureof Personality Development in CollegeStudents." Developmental Psychology, 1969,1, pp. 357-372.

(13) Vaillot, M. C. Commitment to Nursing.(Philadelphia: J. B. Lippincott Co., 1962).

(14) American Nurses' Association, Facts AboutNursing. (New York: American Nurses' As-sociation, 1971), p. 9.

(15) RN's 1966 Inventory. (New York: Ameri-can Nurses' Association. 1969).

(16) Astin, H. S., Suniewick, N and Diveck, S.Women; A Bibliography of their Education

(2)

(3)

(5)

(6)

(7)

(8)

(9)

123


and Careers, (New York: Washington, D.C.,Human Services Press, 1971), p. 4.

(17) Schlossberg, N., and Goodman, J. Resolu-tion to the American Personality and Guid-ance Association: Revision of Strong Voca-tional Interest Blanks (Detroit: College ofEducation, 1971).

(18) Epstein, C. F. Woman's Place. (Los An-geles: University of California Press, 1971),p. 94.

(19) Freeman, H. E.; Levine, S.; and Reeder,L. G. (eds.). Handbook of Medical Soci-ology. (Englewood Cliffs, N.J.: Prentice-Hall, 1963), pp. 200-201.

(20) Davis, F., and Olesen, V. L. "The CareerOutlook of Professionally EducatedWomen7 Psychiatry, 1965, 28, pp. 334-345.

(21) Mates, N.; Schultz, M. N.; and Pierce, C. M."Commitment to Nursing. -How is itAchieved?" Nursing Outlook, 1968, 16(7),pp. 29-31.

(22) March, J. G., and Simon, H. A. Organiza-tions. New York: John Wiley and Sons, Inc.,1958, p. 84.

(23) Rosen, H., and Komorita, S. S. "A DecisionParidigm for Action Research." Applied Be-havioral Science, 1970, 19, pr 509-518.

(24) Cleland, V., and Bellinger, A., et al. "Decisionto Reactivate Nursing Career." Nursing Re-search, 1970, 19, p. 451.

(25) Cleland, V. "Role Bargaining for WorkingWives." American Journal of Nursing, 1970,70, pp. 1242-1246.

(26) Cleland, V.; Bellinger, A.; and Meek L."Husband's Income and Children's Ages asMotivational Factors in Nurse Em-ployment." in V. Cleland (ed.), Part-timeNurses, (Detroit: Center for Health Re-search, Wayne State University, 1971).

(27) Nye, F., and Hoffman, L. The EmployedMother in America. (Chicago: RandMcNally, 1963).

(28) Shea, J. R. Dual Careers: A LongitudinalStudy of the Labor Market Experience ofWomen. (Washington, D.C.: Department ofLabor, Manpower Research Monograph No.21, 1970).

SOCIAL PSYCHOLOGICAL. FACTORS RELATING TO EMPLOYMENT OF NURSES 121

(29) Women's Bureau, Handbook on WomenWorkers. Washington, D.C.: Department ofLabor, Women's Bureau Bulletin 294, 1969).

(30) Personal Communication with Dr. RobertNfendohlsolm, Associate Professor, LafayettClinic, Wayne State University, Detroit,

Michigan. (Date from a questionnaire, thefindings of which have not yet been pub-lished).

(31) Cleland, V. "Sex Discrimination." Ameri-can Journal of Nursing, 1971, 71, pp, 1542-1547.

DISCUSSION OF DR. WOLF'SAND DR. CLELAND'S PAPERS

Discussion Leaders

Miss Marjorie SimpsonNursing Officer in Research

Department of Health and Social SecurityLondon, England

Miss Simpson

I would like to discuss the two papers h. re-verse order, starting with Dr. Cleland's. She hastouched on the universal problem of recruitmentselection and wastage of staff for nursing posi-tions. Her arguments are very good; however oneor two points did occur to me.

The first is whether we need to follow slavishlythe career pattern that men set. I think we haveto consider the resources we have and the needswe have to meet. We should I am sure be happyto have men with us as nurses and I believe thatin my own country we have gone farther thanyou have here in incorporating them into theprofession, not as men, but as nurses.

Single women living away from home in medi-eval times were known as "anilepi" women. Theywere thought to have no viable life of their ownnut throughout their lives they were expected tobe dependent on a male member of the family,usually the senior male member. They held posi-dons as daughters, sisters, or aunts. They were notindependent. This has lingered with us, but notI think to the extent that we haven't brokenaway from it for women to hold happy, successfulpositions in their own fields of work.

This varies greatly from culture to culture. I was,for instance, astonished recently when I happenedto be in a hospital. I had as a ward sister a veryexcellent nurse from Singapore who refused ab-

Dr. Mario BognannoAssistant Professor

Industrial Relations CenterUniversity of MinnesotaMinneapolis, Minnesota

solutely to address me as "Miss" because clearlythis was insulting. To this day I am on the recordsof the hospital as "Mrs." I do not believe thiswould have happened had the ward sister beenEnglish.

It should be possible today for single women,for men, and for married women to mingle to-gether in the nursing profession, finding their ownposition. But for the married woman with chil-dren, I personally have swallowed hook, line, andsinker the things researchers have identified in re-lation to the deprivation that children suffer whenthey are deprived of their mother's care in theearly stages of their lives. In this way, of course,my public health biases come across very strongly,and I cannot reconcile the idea of the nurse whoshould be the health teacher setting the exampleof leaving the children. This I think you maystrongly wish to dispute with me. This is my bias,which I am now declaring.

But I do wonder if we couldn't cash in on thissituation, whether it is biologically determined orculturally determinedI don't know which it is.

In nursing there are very many positions at thefoot of the ladder. There are far fewer higher up,and I am not sure that wastage for child rearingis entirely dysfunctional. If we are preparing ournurses for health care as well as for sickness care,we may be able to rear a much healthier future

1.22Id()

DISCUSSION OF DR. WOLF'S AND DR. CLELAND'S PAPERS

generation if the mother understands the import-ance of child health. This is not only within herown family. As a member of the community inher peer group she can have a gt eat influence onother mothers, and we should make it possible forher to contribute in voluntary service or in otherways to the health service during the time she isrearing her own family.

One of the problems is that we prepare nursesfor a caring role that they have to change atsome stage in their career either for a much moretechnical role or for an administrative or academicrole. And I do wonder whether we could use thisgap to train or retrain the women probably ofabout age 35.

Now, there is experimental work going on inthe training of older women. We have done farless in thinking about the possibility of retrainingour nurses for a rather different role when theycome back. There is quite a lot of talk at presentabout the need for people in many occupationsto retrain at 40. We might examine this possibilityin relation to nurses who have a natural breakin the late 1920's.

This leads us rather closely to the %object of. the second paper. Here we saw all too clearlythe problem of the nurse unaccustomed to thebudgeting, management approach. This will needto be incorporated somewhere in her trainingperiod. In England, as you probably know, we aremoving into a unified health service, and 1974for many of us has a ring both of challenge and ofanxiety. We have a critical path analysis program

123

going on with the idea that in 1974 the transitionwill be made smoothly from a tripartite organiza-tion of the health service to a unified organization.One small segment of this scheme is looking atthe management of areas, and it has been quiteclearly stated that the nurse director, whatevershe will be called in this new setup, will have tobe able to develop and control the nursing budget,not only for the hospital but for the area as well.This is another thing for which we are not atpresent well prepared, and certainly for us thisis one of the things we shall need to look at.

The other feature that came out so clearly inDr. Wolf's paper was the problem of action re-search. It is not too difficult for the outsider to goin and describe a situation. We have many, manyof these studies. It is when you try to work withpeople who are affected by the study that the prob-lems start to arise. We should congratulate Dr.Wolf most warmly on his courage in coming hereand telling us of the problems we shall meet.Those of us who have tried already know the diffi-culties to some extent.

He gave a most masterly exposition of thethings that are going to happen and that we needto look at to arrive at a situation in which theresearch worker can genuinely help the people onthe field and not merely find ways in which to de-scribe them and criticize them. It is not difficultto go in and describe how badly a thing is working.It is a very different proposition to go in and helppeople to adjust.

Dr. Bognanno

I would like to make a few very general com-ments about each paper and then bring to yourattention another body of research that is cur-rently being undertaken by a handful of econo-mists in the area of nurse staffing.

In my mind there is little doubt that the nursestaffing research dollar will yield a considerablereturn if it results in the successful developmentof operational schemes to motivate hospital econ-omy. Two-thirds or more of most hospital expendi-tures are allocated to salaries, and we all knowwhat has been happening to hospital costs in theUnited States over the past 15 years relative tocosts in other sectors of the economy. Thus the

study of the effect of peer review on the efficiency(lower cost, quality constant) with which healthmanpower, particularly nurses, is utilized by thehospital is a very worthwhile problem area.

Dr. Wolf's study tried to get at answers to ques-tions like the following: (1) how does the peerreview p mess operate in relation to cost contain-ment and in relation to patient care, and (2) if thenursing director is on the peer review board, doesher administrative knowledge increase; if it does,is this knowledge manifest in relationships withsubordinates and in relation to cost and perform-.ante?

Well, these were the questions to which Dr. Wolf

131

124

and his colleagues tried to find answers. In con-currence with the previous discussant, I too thinkthese questions still go unanswered. However, inour continual search for answers to them we canlearn from Connecticut's experimental design andfrom the many *research problems encountered dur-ing the study as discussed in the concluding sec-tion of Dr. 1Volf's paper.

The ideas that 1)1'. Cleland and Mary I PC Sul-kowski develop are rooted in social psychology.From their literature review it would appear thatconsiderable agreement exists among social psy-chologists regarding the point that a woman inthe United States today does not engage in (atleast during early years) an occupationalcommitmentjob achievement fot her identityand self-esteem. Rather, the latter is found inthe woman's role as a wife and mother. Speakingas an economist, I consider the formal develop-ment of these theories to be excellent, and I ampersuaded to agree with their specific explanationof the work behavior of professional nurses.

In effect, what Dr. Cleland and her colleagueare saying is that the supply of nursing resourcesiu society will increase, holding everything elseconstant, if we can negotiate or bring about achange in the social status of women as they re-late to labor market work. In this instance thechange being recommended is to broaden thewoman's source of identity and self-esteem fromthe mother-wife role only to that which also in-cludes her role as a professional in the labormarket.

If, on a graph, we label age on the abscissaand the labor force participation rate on the or-dinate, then for nurses we would see a bimodallabor force pattern develop. The probability oflabor force participation would be relatively highduring a woman's younger years after college; thisprobability would decline during her childbearingand childrearing years, and then following theseyears it would increase until the woman entersher retirement years at which time the probability


of labor force participation begins to decline. Dr.Cleland and Mary Lee Sulkowski see the entiredistribution of participation rates shifting upwardif it were possible to, say, create the same labormarket role for women as that which currentlyexists for men. That is, such a change in thinking(roles) would result in more human (nursing)resour "es being allocated to market work.

Thi,- brings me to the last point I wanted tomake regarding research currently being conductedby me and a few colleagues like Drs. Stuart Alt-man, Lee Benham, Jesse Hixon, and Donald Yett.We have been studying the relationship betweenvarious measures of nursing labor supply and anumber of socioeconomic variables that impacton the former. In general, we have not consideredthe effects of change in sociopsychological variables(as discussed by Dr. Cleland) on the nurses' laborsupply and labor force participation behavior.Rather, we implicitly assume that at a point intimein cross sectionthere is little or no labormarket role variation among nurses that couldaccount for variations in the nurses' labor marketbehavior.

This assumption is realistic enough when itcomes to cross sectional studies; however, throughtime, cultural changes do occur and their effect onmarket work behavior must be taken into account.As an aside, I should note that this is preciselythe point where Dr. Cleland's study complementsour work in this area. Indeed, one of the positiveaspects of this conference is that it has given usthe opportunity to gain interdisciplinary insightsinto a common research problem.

The variables we have used in analyzing thenurses' market work behavior include husband'searnings, the nurse's wage, family size, age of chil-dren, and the age of the nurse, to mention a few.I am confident that most of you will find thepartial relationship between these variables andthe time a nurse spends in the labor market mostinteresting.

Evaluation of Quality of Nursing Care

Dr. Mabel A. WandeltProfessor of Nursing


Introduction

In any project to evaluate the quality of nurs-ing care, whether for research, educational, oradministrative purposes, those directly responsiblefor nursing care of patients must be involved.Though it means beginning on a rather negativenote, it seems to me that it may serve us to identifysome of the difficulties these persons encounter asthey are involved in evaluation endeavors. It hasbeen my experience that nurses, of whatever levelof administrative hierarchy and in relation to

whatever purpose for evaluation, express dissatis-faction with evaluation. They may be very satis-fied with all other elements of a program, or vari-ous nurses may identify particular areas aboutwhich one or another hap mild unease, but whenevaluation is mentioned there is unanimous ex-pression of dissatisfaction. There is dissatisfaction

either with all aspects of the program or, frequently,with the complete absence of a program.

There are four sources of dissatisfaction withevaluation programs in which nursing staffs areinvolved. Perhaps the fundamental source of dis-satisfaction is failure to recognize the magnitudeand complexity of the task. The second most basicsource is the imprecision with which many inherentterms are used. The third is the failure to dis-tinguish between and to make explicit the reasonsand purposes for undertaking evaluation. Thefourth is the lack of neatness possible in any eval-uation endeavor where human action, interaction,and reaction are involved.

There of course is interrelationship among these,and the sequence of listing is not meant to indi-cate greater or less importance, except for theprimacy of the first one.

Understanding and Precision in Use of Terms

For reasons that will later be clear, 1 shall begindiscussions with the second listed source of dis-satisfaction. Among the difficulties in relation touse of terms is failure to define the terms on a level

of sufficient concreteness and specificity, in con-trast to abstractness and generalization, so that acommon understanding is achieved among thoseusing the term. Perhaps the source of greatest

125

133

126

difficulty is the interchange of words with similaror closely allied meanings. These arc often usedas though they had identical meanings in onesentence but had distinctly different meanings inthe next sentence.

My protest is not that every word or term hadbut one meaning or that there are not wordshaving two or more meanings. The point I wish tomake is that interchange of words or use of oneword for two or more connotations in a singlegeneral context frequently presents problems ofcommunication. In most such instances it is pos-sible to designate distinguishing definitions of theconfusing terms. Further, it is reasonable to sug-gest that, in any one context under consideration,each word be used only in its defined meaning,and that two terms, which in other context mighthave interchangeable meanings, not be so used inthe particular context.

For example. of immediate concern to us arethe words "evaluation" and "measurement." Youmay propose that, since their meanings are quitedistinct, it is unlikely that these two words wouldbe used interchangeably. Indeed, it is true thattheir meanings are distinct, but it is also true thatpersons frequently fail to make the distinctionas they use them. A most common phrase, "evalua-tion and measurement," would seem to indicatethat many do not use the terms with precision.The structure of the phrase seems to indicatethat individuals do not recognize that measurementis a prerequisite to evaluation.

Delineation of varied usages of terms pertinentin considerations about evaluation and possiblereasons for varied usages, along with the accom-panying confusions or limitations in communica-tion, would involve detail beyond the scope ofthis paper. Clarification and suggestions for allevi-ation of some of the difficulties may be served bydefining some of the involved terms, followed byidentification of some difficulties in usage. To pre-clude confusion in the meanings of the variousterms employed in this paper, I have included atthe back a list of definitions.

It seemed a good idea to start with a definitionof definition. Some of the definitions are excerptsstraight from Webster. Many, however, are con-structed on the basis of logic and fit for the pur-poses of the context in which they will herein beused. Brief discussion of some of the words seems

134


warranted. For example, recognition of the dis-tinct meanings of measurement and evaluationmight well expedite the work in all projects forevaluation. Failure to recognize. differences in themeanings of the .two terms leads to attempting tobegin in the middle or near the end of the processof planning evaluation. Measurements arc identi-fied without establishment of a base on which tojudge their relevance to the purpose for whichthe evaluation is to be undertaken.

Standard, and Standard of Measurement

Another area of difficulty derives from the inter-change and failure to distinguish between mean-ings of standard and standard of measurement.Actually, except for the undesirable connotationof average inherent in- the concept of norm, itmight be well to adopt the use of the term "norm"where ordinal fly the term "standard" is meant.The connotation of average is here consideredunfortunate because frequently, when standardsare set, the intent is that the goal or what is de-sirable is something above average. The latter doesseem to be implied by the concept of standardwhen it is defined as "a level of excellence."The failure to distinguish among the meaningsof standard, standard of measurement, and unitof measurement is frequently observed in nursingliterature, particularly where nurses are strugglingwith evaluations of clinical competence. In al-most any article on the subject, there is lengthydescription of the various standards of measure-ment to be used for evaluations of nurses withvarious levels of educational background.

It is proposed that there be different standardsof measurement because the same level of compe-tency cannot be expected for nurses with variededucational preparations. What is meant is thatthere must be different standards, different levelsof competency that will be acceptable, not differ-ent standards of measurement by which to ascer-tain those levels.

1 am confronted with these misconceptions re-peatedly in relation to the Slater and Qualpacsscales, in which the standard of measurement is"care expected of a first level stall nurse." Thoseconsidering use of the scales, whether educationor service personnel, ask about changing thestandard of measurement in the events of using

EVALUATION OF QUALITY OF NURSING CARE

the scales to measure competencies of nursing stu.dents, staff nurses, or clinical specialists. They ob-viously are unaware of the meaning of a standardof measurement being "an object which servesto define the magnitude of a unit." They fail todistinguish between two concepts:

(a) the concept of standard, which denotes anaccepted level of excellence and which impliesthe appropriateness of establishing varyinglevels of excellence for various levels of stu-dent or for the nurse staff with variousqualifications,

(b) the concept of standard of measurement,which denotes an object that, under specificconditions, serves to define or represent themagnitude of a unit and that, in turn, per-mits comparable measurement to he made ofsimilar phenomena. It provides the unit ofmeasurement to be used for determining thedegree of attainment of an established or ac-cepted standard.

Subjectivity and Judgment

Another difficulty frequently identified by per-sons considering the use of tools for measuringclinical competence of nurses is concern aboutsubjectivity in measurement. In a way, it seemsremarkable that nurses have so much discomfortabout the matter of subjectivity and "subjectivejudgment" in relation to measurements of clinicalcompetence displayed by nurses. This is particu-larly remarkable in relation to clinical instructorsand supervisors. Their whole effort in clinical in-struction, guidance, and supervision is aimed to-ward assisting nurses to make judgments.

The professional nurse's forte is her clinicaljudgment. In 01 of her clinical supervision, theinstructor or supervisor uses her own clinicaljudgment to determine when a nurse being super-vised is making a correct judgment and when shemay need guidance to improve her judgment.This is accepted with no problem. Now let thesupervisor redirect her conscious focus away fromguidance and instruction and toward evaluationfor the purpose of accounting for the nurse's dis-play of clinical competencea large part of whichis professional judgmentand she, and all abouther, question the goodness of the evaluation shemight do because it is based on subjective judg-

135

127

ment! Her use of clinical judgment is acceptedwithout question when she uses it to instruct.But when used to evaluate the results of herinstruction, it is immediately something less thansatisfactory.

Perhaps it would help if we looked at whatthe words mean. To judge is "to form an ideaor opinion abolit; to think or suppose." Subjectivemeans "of, affected by, or produced by the mind."Actually, subjective judgment is a redundancy:all judgment is subjective; there is no judgmentwithout subjectivity. All nurse actions stem fromjudgments. Professional judgment has basis in factbut would be nothing without the catalyst ofsubjectivity.

Always there is the quest for objective evalua-tion. Here we have blatant examples of impre-cise use of terms: subjective judgment is a redund-ancy, and objective evaluation is an antithesis.

Evaluation is the term usually used by thoseexpressing dissatisfaction, despite the fact thatthey are usually talking about subjectivity in themeasurement aspect of the evaluation process.Possibly some of the discomfort and condemnationoccurs because the denotation of subjectivity is

used instead of that for subjective. The diction-ary definition of subjectivity is "The tendency toconsider things primarily in the light of one'sown personality. Concern with one's own thoughtsand feelings." Whereas subjective means "Of, af-fected by, or produced by the mind." The defini-tion of subjectivity is undoubtedly what leads tothe comment: "They evaluate a staff nurse onwhether or not they like her as a person."

Might nurses be helped if they were reminded ofspecifics of what they are really about? Thatevaluation is a generalization descriptive of a judg-ment based on many measurements. The measure-ments are of the nursing care actions of the indi-vidual being evaluated. The measurements onwhich the evaluative judgments are based are notmeasurements of attributes or qualities of theperson. The person happens to be the vehicle forthe entities being measuredthe performer of theactions. The actions are being measured, not theperfomer of the actions.

Also, nurses might develop confidence from areminder from Abraham Kaplan when he said,"The new treason of the intellectuals is thatwe have shared and even contributed to the cur-

128

rent loss of faith in the power of the humanmind to cope with human problems, faith in theworth of reasoned discussion, faith even in thepossibility of objective truth.- It seems to me thatnurses need to recognise specifically what theirconcerns are and to have the confidence and four-age to know that they are capable oc making in-formed judgments based on observed facts.

Criterion Measures

Another whole area in which nurses neglect touse definitions and precise meanings of words toassist them is in identification of criterion meas-ures appropriate to securing measurements onwhich to base valid and relesant evaluations. Whenfaced with the need to plan for evaluation, theybegin by identifying observations in terms of theoverall objective or purpose of the program to beevaluated. They fail to approach the task at eventhe suhobjectives level, not to mention enteringat the level of constructing precise definitions ofeach crucial word in the subobjectives. They seem

RESEARCH ON NURSE STAFFING IN 110SPrAIS

not to have discovered that detailed specificity ofterms can reveal whether an objective, as stated,actually says what those who constructed it in-tended it to say. They seem not to know that theonly assurance of relevant criterion measures' beingidentified is via detailed and specific definitions.

There I commend to you the process of develop-ing a three-phase definition of the crucial termswhere in the first phase is a general or dictionary de-finition. The second phase is a pertinent generaldefinitionthe general meaning of the term in thecontext of the concern of the evaluators. In de-lineation of the pertinent general definition, thecriterion measures will evolve. The third phase isa for-instance definition, in which an example ofa single relevant measurement is described. Thisprocess of definition ensures a systematic approachto assurance that criterion measures and actualmeasurements to be sought will be relevant to thepurpose for which the evaluation project is under-taken. The crucial terms to be so defined are thosethat stipulate the outcomes of achievements of theprogram to be evaluated.

Failure to Make. the Purpose Explicit

The general concept that evaluation is done forthe purpose of "gathering information on whichto base improvement" is useful as far as it goes. Itcan he used to serve as a criterion on which toscreen potential relevance of observations to bemade. But there are other purposes for t:oingevaluations of quality of nursing care. Among themare quality control, the obligation of account-ability, continuing equation, and determining re-lationships among a variety of potentially relatedvariables. Here .again there may arise difficultiesdue to lack of clarity about meanings of words.

Reasons and purposes for doing evaluations aresometimes confused. Particular measurements mayserve the intents of persons with various reasonsfor doing evaluations. For example, data that willprovide for determining relationships among vari-ables will serve persons whose interest is research.the interests of the therapist. and those of theadministrator. Therefore, failure to identify speci-fically the reasons is not apt to influence thequality of the outcomes of evaluation. 41k.

On the other hand, the purpose for which thedata will be gathered must he made explicit, if

136

data germane to the purpose are to be obtaineaFor example, if the purpose for evaluating out-comes of a new procedure is to determine itseffectiveness in preventing skin breakdown, thedata to be gathered will be rather different fromthose gathered should the purpose be to determinethe numbers and mix of staffing required to per-form the new procedure as recommended for aset of patients.

Other factors alluded to earlier as encompassedin this broad area of difficulty, failure to makepurpose explicit, may be considered under thethree-part framework for evaluation of human ac-tion programs: structure, process, and outcome.When planning evaluation of quality of nursingcare it is CSACIIIIII1 to be aware of what, specifically,is to be measured. What is to be measured, ofcourse, is identified in the declaration of thepurpose for doing the evaluation. It is conceivablethat a particular purpos, would dictate measure-ment of all three components of care: structure,process, and outcome. If this is the case, the plan-ners should he fully aware if the intent, Theymust also approach the task with definitive plans

EVALUATION OF QUALITY OF NURSING CARE 129

for measurement of each of the components. Thisawareness and approach will ensure cognizancethat few, if any, evaluation projects can be ac-

complishecl through utilization of a single tool orprocedure for gathering data.

Lack of Neatness

The impossibility of accomplishing an evalua-tion of nursing care with a single procedure orinstrument of data gathering is part of what Imean by the lack of neatness of the process. Ifnursing care were unidimensional, if there were anestablished standard of measurement for this di-mension, if there were a validated, scaled instru-ment for making the measurements, evaluation ofnursing care would pose few if any difficulties.

But nursing care itself is multidimensional. Thereis no consensus about what many of the dimen-sions are, not to mention consensus about stand-ards or units of measurement for them. Whenevaluation is considered there will be concernsabout the various facets of care, such as the process

of care, the setting and a situation of care, theattributes and qualifications of persons administer-ing care, and the outcomes of care in terms of timeand efficiency and in terms of welfare of thepatient.

These and many more factors make for lackof tidiness and directness in the accomplishment ofevaluation of nursing care. Perhaps the most im-portant element in' dteviation of these difficultiesis the recognition and acknowledgment of theirexistence. The recognition of them and the natureand influence of each should lead to organizationfor orderly approach to the task of evaluation.They should lead to effective and efficient se-

quencing of the components of the total task.

Failure to Recognize the Magnitude and Complexity of the Task

There was reason for elaboration of the othersources of dissalsfaction with evaluation projectsbefore addressing..the failure to recognize the mag-nitude and complexity of the task. Essentially,facets of the other sources as elaborated identifyand describe components of the task. For example,there is the need to make explicit the reasons andpurposes for doing the evaluations, the need forprecise definitions of terms in order to reveal boththe accuracy or lack thereof of the identification ofobjects to be measured and the criterion measuresrelevant to them, and the determination of theprecise data germane to the purpose to be served bythe evaluation.

There needs to be recognition .hat much hardthinking and planning must be done. Many de-

cisions must be made before any data are collected,and many and varied data must be gathered andorganized. In turn, further hard work of thinkingand decisions must be done before evaluative state-ments can be delineated, before there will be in-formation to serve as bases for change, to fulfillthe purpose for doing the evaluation.

Since research in evaluation of quality of nttrs-ins, care must involve those directly' responsiblefor the nursing care of patients, the researcher whois cognizant of the difficulties experienced by thenttises and who assists them to understand andcope with them will advance his investigation andincidentally will assist nurses with an importantadministrative task.

WhereIn Evaluation of Quality of Nursing Care

Variables of primary concern to investigators ofnursing care of patients are those composing thestructure, process, and outcome of patient care as

described by Donabedian. The ultimate concern isthe relationship of all other variables to thoseidentified as outcomes of nursing care. In this

137 ,

130

context, all other variablesstaffing, facilities, philo-sophy, policy, staff performanceare causal varia-bles. Outcomespatient well-being, patient im-provementare effects variables. It is within thisbroad framework that research in evaluation ofquality of nursing care must be pursut

Historically, research in evaluation of nursingcare has progressed along a continuum of theconcrete to the more-or-less abstract. There havebeen successes in development of measurements ofstructure: numbers of persons, time, facilities.Among them are the "how to" techniques andinstruments developed by the Division of Nursingduring the 1950's and early 1960's: how to studynursing activities in a patient unit, how to studynursing service in an outpatient department, howto study nursing supervision, and others.

There have been successes in development ofmeasurements of process: nurse actions, interac-tions with and interventions on behalf of patients.Among these are three developed by the facultyof the College of Nursing, Wayne State University:

(a) Slator Nursing Competencies Rating Scale,for measuring competencies displayed bya nurse,

(b) Quality Patient Care Scale, for measuringthe quality of nursing care received by apatient while care is ongoing.

(c) Nursing Audit, for measuring the quality ofnursing care received by d patient after acycle of care has been completed and thepatient discharged.

These three instruments and suggestions fortheir uses, have been described in some detail inan article in Hospital Topics, August 1972. Andthere have been beginning studies to develop meas-urements of outcomes of nursing care.

There must be continuing work in all areas alongthe continuum. But it does seem to me that wehave reached a point where a concerted effort iswarranted in the area of measurement of outcomes


of nursing care in terms of patient well-being.All of the comments about difficulties in evalua-

tion apply here. Some of us have wrestled with theideas of the magnitude and complexity of the task.We have faced the need to define some of theterms. In these attempts we ran into problems ofmaking explicit the purposes for evaluation: Whosepurposes are to be of concern? During a confer-ence last fall, a group of us recognized and re-conciled ourselves to one difficulty: the task ofmeasurement of outcomes of nursing care will notbe accomplished through development of a singlemeasuring instrument.

Recognition of the magnitude and complexityof the task should be sought, not as an excuse forimmobility but rather as a caution that the taskcannot be approached all of a piece. Progress todate has been made by individuals and groupsisolating single facets of the larger task and pursu-ing information about each. It may be expectedthat continuing work will follow that pattern butalways with the plea for increased communicationsamong the seekers. Communications will promotethe work on individual projects and preclude un-necessary overlap or repetition.

The goal to be sought through research in eval-uation of quality of nursing care must be capabilityof determining relationships among variables ofstructure, process, and outcome. There are availa-ble to us, not finished or perfect but usable, meas-urements of the first two. There are some measure-ments identified as potentially usable for the thirdarea. But it seems to me we need to recognizethe magnitude of the task and proceed in a system-atic fashion by beginning with determination ofcriterion measures of outcomes of nursing care.The criterion measures must be delineated interms of patient well-being.

Work has begun. There is extensive time anddistance to go, but the beginnings have been madeand the direction is discernible.

138

EVALUATION OF QUALITY OF NURSING CARE 131

Definitions

Definition.An arbitrarily imposed descriptionthat allows common understanding.

Evaluate.To ascertain or fix the value or worthof.

To examine and judge.Evaluate implies considered judgment in settinga value on a person or thing.

Measure.To ascertain the dimensions, quantity,or capacity of.To mark off, usually with reference to some unitof measurement.

Evaluation.A generalization describing a judg-ment based on many measurements.

Evaluation.A process by which we gather infor-mation as a basis for improvement.

Quality.A characteristicor attribute of some-thing: excellence, superiority, degree or grade ofexcellence.

Standard.Something used by general agreementto determine whether a thing is as it should he.An agreed upon level of excellence.An established norm.

Measurement.A comparison of a single pheno-menon with a standard of measurement: the re-corded number or symbol that represents themagnitude of the phenomenon in terms of themagnitude of the standard of measurement.

Standard of judgment.An object which, underspecific conditions, serves to define, represent, orrecord the magnitude of a unit.

Unit of measurement.A precisely defined quan-tity in terms of which the magnitude of all otherquantities of the same kind can be stated.

Judge. To form an idea or opinion about (anymatter).To think or suppose.

139

Judgment.The mental ability to perceive anddistinguish relationships or alternatives; the criticalfaculty; discernment.The capacity to make reasonable decisions, espe-daily in regard to practical affairs of life; goodsense, wisdom.

Subjective.Of, affected by, or produced by themind or a particular state of mind.Of or resulting from the teelings or tempermentof the subject or person thinking, rather thanfrom the attributes of the object thought of; as,a subjective judgment.

Subjectivity.The tendency to consider thingsprimarily in the light o5 ont'N own persmatity.Concern with one's own thoughts and feelings,

Objective.Of or having to do with material ob-ject as distinguished from a mental concept, idea,or belief.Having actual existence or reality.Uninfluenced by emotion, surmise, or personal prej-udice.Based on observable phenomena; presented fac-tually.

Measurement.An objective term: the ascertain-ing of a dimension through observation of a phe-nomenon.

Evaluation.A subjective term: the ascertainingor fixing of a value through considered judgment.

Criterion measure.A quality, attribute, or char-acteristic of a variable that may be measured toprovide scores by which subjects of the same classcan be compared in relation to the variable.

Variable.A measurable or potentially me; ura-ble component of an object or event that mayfluctuate in quantity or quality, or that may bedifferent ill quantity or quality from individualobject or event to another individual object orevent of the same general class.


References

(1) Donabedian, Avedis. A Guide to MedicalCare Administration. vol. 11. Medical CareAppraisalQuality and Utilization. Ameri-can Public Health Association, New York,New York, 1969.

(2) Phaneuf, Maria C. The Nursing Audit:Profile for Excellence. New York: Appleton-Century-Crofts, 1972.

(3) Wandelt, Mabel A. The Slater NursingCompetencies Rating Scale. College ofNursing, Wayne State University, Detroit,Michigan, 1967.

(4) Wandelt, Mabel A., and Ager Joel. The

Quality Patient Care Scale. College of Nurs-ing, Wayne State University, Detroit, Michi-gan, 1970.

(5) Wandelt, Mabel A. Guide for the Begr,ingResearcher. New York: Appleton-Century-Crofts, 1970.

(6) Stufflebeam, Daniel L., et al. EducationalEvaluation and Decision Making. Itasca,Illinois: F. E. Peacock Publishers, Inc. 1971.

(7) Wadelt, Mabel A., and Phaneuf, Maria C."Three Instruments for Measuring Qualityof Nursing Care." Hospital Topics, August1972.

140

DISCUSSION OF DR. WANDELT'S PAPER

Dr. Elizabeth HagenProfessor of Psychology

Teachers CollegeColumbia UniversityNew York, New York

Discussion Leaders

Mr. Richard G. GardinerOperations Research Specialist

Hospital Association of New York StateAlbany, New York

Dr. Hagen

In all the definitions that are given, there is

no definition of quality. During the past day anda half, whenever we have talked about quality ofcare, values and value systems have been reflectedalthough no one has made them explicit. Manyof the arguments about what should be studiedas indicators of quality of care arise from conflictsin values. We would do well to make the valuesystems that underlie our selection of criteria forjudging quality explicit.

One problem with giving definitions is that youalways want to start haggling with the definition.I do not want to start haggling. However, I amgoing to use the phrase "criterion variables" be-cause one of the immediate problems is to identifythe variables that are relevant indicators of qualityof care and then to talk about how to appraisethose variables.

We ought to make a clear distinction betweenthe term "criterion variable" and the term "stand-ards." Theoretically a criterion variable can andshould have a range of values from low to high.Standards imply that you are seeking some definitevalue or score on a particular variable in order tosay the performance reflected by the variable is atleast satisfactory or adequate.

The three terms "structure," "process," and"product" that Dr. Wandelt used are creeping intothe literature on evaluation in the health care area.I believe these terms were first used in a paperby Tyler in 1931. He used the three terms to

describe levels of evaluation in education. Hepointed out that most of the evaluation of qualityof education was in terms of structure, the condi-tions under which education takes place, or ofprocess, how the education was being conducted.The product or outcomes of the educational proc-ess, according to Tyler, were rarely appraisedbecause they were believed to be too difficult tomeasure. Tyler criticized people for assuming thatbecause certain structures or processes were pre-sent the desired outcomes must be present. Weknow that the assumed high relationship betweenprocess or structure and outcomes in educationdoes not exist, but in education we still have mostof the emphasis placed on structure and processand too little on outcomes. I am bothered when Isee the health care field repeat the mistakes ofeducational evaluation. Although the three termsmay be useful in classifying variables used to judgequality of health care, they do not provide anadequate framework for studying quality of healthcare.

To make progress in the evaluation of healthcare, perhaps we should stop asking the globalquestion, "What is the quality of nursing care?"and start asking more specific questions. The morespecific question will permit better identificationof the value system being used and will providea better frame of reference for judging the rel-evance of criterion variables than will the globalquestion.

133

141

134

A number of frames of reference can be usedto study nursing care or, if yoo prefer the term,quality of nursing care. One frame of referencecould be what the profession of nursing would likethe practice of nursing to be. If you could identifywhat. professional nurses deem to he "good prac-tice," you could appraise the extent to which thepractice matches these expectations. Or if you couldidentify the conditions necessary for delivering whatis considered "best nursing practice," you couldappraise the extent to which these conditions arepresent.

A second frame of reference that could be usedto appraise nursing care is patient outcome. Toascertain that nursing practice is congruent withprofessional expectations of that practice does notguarantee that desired patient outcomes arc pre-sent. Although no one would argue that the ex-pectations of the profession concerning the desira-ble practices of nursing are set without regard towhat happens to the patient, we have very littlehard data on the relationship between nursingpractice and patient outcomes. To locus too muchattention on practice and too little on patient out-

REsEARCH ON NURSE STAFFING IN HOSPITAL s

comes might result in the methods of practicebecoming ends in themselves or being used more.for the benefit of the profession than for the bene-fit of the patient.

A third frame of reference that could be usedis consumer or patient .expectations of the healthcare or nursing care system. This frame of refer-ence differs from the second one mentioned in theprevious paragraph. In the latter, patient outcomesare judged by the health care professional in termsof what the professional wants to achieve with thepatient in terms of his health or well-being. In otherwords, it is the extent to which the status of thepatient meets the expectations of the health careprokssional.

In the third frame of reference, it is the extentto which the status of the patient meets patientexpectations or the extent to which the servicemeets patient expectations for that service. Patientexpectations may be reasonable or unreasonablein the eyes of the-professional; however, they arevery real to the patient and cannot be ignoredin the appraisal of any health care system.

Mr. Gardiner

I have simply a set of questions that hopefullywill provoke some response. First, is it not likelythat points at which we make measurements ofquality go a long way toward determining whatends up being defined as quality? We can measureit either at the input level of the process or at theoutput level. If it is measured at the output level,are we not measuring the ability of the profes-sionals to carry out their tasks; not whether, infact, the tasks are carried out, but whether theyhave the ability to carry them out?

We have said the psychological social needs ofthe patient are critical toward an evaluation ofquality. I suspect this is equally true of the psycho-logical social needs of the nurses who respond tochanges in the level of staffing. This is, in fact,equally as important in their view as the qualityof care that they provide.

At the output level, we are judging the ultimateeffects in the patient's welfare or well-being. Doeshe recover? Does he recover sufficiently, or towhat degree does he recover? How long does he

stay recovered? It may be that we need to con-sider the relative costs and benefits of taking meastires at any of these points to determine where weought to focus our attention. Where do we wantto begin measuring to define quality?

Why measure? That has been alluded to. Isquality relevant to changes in staff? Harvey Wolfeasked the question and I think it's a good question.At least two things indicate that if we are lookingat quality simply as a reflection of marginal orsmall changes in levels of staffing, maybe the staffingcan be done independently and we do not needto examine quality.

First, under crisis or stress situations (shortstaff), the quality of care delivered to the sickestpatients probably remains a great deal differentfrom the quality of care delivered to the patientswho are not as sick. Do changes in staffing levelshave an effect on the ultimate outcome? Not howthe patient perceives he's being treated or how thenurse feels, but does it have an effect on thefact of recovery? Second, are there not sufficient

DISCUSSION OF DR. WANDELT'S PAPER

episodes of illness that are, themselves, self-limit-ing? That is, will the patient get better despitethe quality of nursing care delivered?

That seems to beg the question of why we wantto measure quality. If we want to measure qualityfor purposes other than staffing, it seems to berelevant. Is it relevant for evaluating the effects ofchanges in staff, at least marginal changes in staff?Moreover, can the quality of nursing care be con-sidered independently of other aspects of quality

135

of care, the quality of medical practice, the qualityof the ancillary services and all the other servicesthat are provided in the health care system? Don'twe have to decide a priori what the cost benefits areor at least estimate them to determine what aspectsof quality we want to investigate and what we wantto measure?

These are all considerations that must be un-dertaken if we are to perform a realistic measure-ment of quality of nursing care.

143

Psychosocial Factors in

Hospitals and Nursing Staffing

Dr. 'Reginald W. RevansSenior Research Fellow and Scientific Advisor

Fondation Industrie-UniversiteBrussels, Belgium

Introduction

In preparing this document I am only too awareof our general ignorance of the psychosocial fac-tors that influence the behavior of those who servecomplex human institutions. After a year as rap-porteur general to a group of management expertsnominated by the governments of the Organizationof Economic Community Development countriesto study the relations between management andlabor across the industrial world, I am particularlyaware of may own ignorance.

In 1964 I concluded that hospitals are institu-tions cradled in anxiety. In consequence, the mosturgent task of nurse administrators and all othersupon whom they rely is to ameliorate this anxiety.No research work since that time has modified thisconclusion, although much has been done to sug-gest how the anxiety may he tempered. Since theanxiety varies significantly from one hospital toanother and even from one ward to another in thesame hospital, it is clear that anxiety is signif-icantly a product of the local situation. It fol-

lows that only changes in the local situation arelikely significantly to affect anxiety, however desira-

137

ble more general measures may be to improveconditions throughout health services as a whole.

One way of changing a local situation is by ad-ministrativt. surgery, transplanting the person incharge. But like surgery carried out in other con-texts, many subtle equilibria may be disturbedthat were neither forseen nor even understood. Ihave therefore relied in this paper upon the blanderresources of administrative and managerial psy-chotherapies. To what extent can we treat theanxiety situations as offering scope for learning?What are the conditions under which the dailywork, however charged it may be with anxiety,may become a process of autonomous development,of learning by doing, of social reinforcement?

I give here two sets of illustrations from ac-tivities in and around London. One set offers theresults, good and bad, of what was called theHospital Internal Communication (HIC) Proj-ect. The other describes a series of discussiongroups among nurses of all grades from matrons tostudents at the King's Fund Hospital Centre.

144

138 RESEARCH ON NUKSE STAFFING IN HOSPITALS

Earlier Studies

A number of us at the University of Manchester,Manchester, England, did research on what beganas a study of nurse wastage, or dropouts amonggirls who were qualifying to become nurses.

Manchester is the largest city in industrial Lan-cashire. When the Industrial Revolution took placeit left behind it in Lancashire 14 towns that, withthe exception of Manchester, are all about thesante size. They are all identical; the same per-centage of people do all different kinds of jobs.The ratable value for head of population, com-munity tax inhabitant is remarkably similar. So wehave 14 separate towns scattered around LancashireMai are more homogeneous than any other 14towns that you can draw anywhere.

We had thus assembled 14 ecologies that aredistinct but identical. In each of these was ahospital. We therefore decided we would examinethe structure of these hospitals, their metabolismrather than their anatomy, to determine how suc-cessful they are in digesting those who enter them,both nurses and patients.

Researchers in the extremely homogeneous en-vironments of the Lancashire towns suggested thathospitals with serious nurse staffing problems haddifficulty in retzining all grades, not student nursesalone. If the senior staff and ward sisters werestable, were the nurses in training. In like man-ner, instability reflected itself down the line fromsenior staff to nurses in training. Again in thishomogeneous Lancashire sample, hospitals withstable nursing staffs seemed to have short lengths ofpatient stay. A stable staff seemed to make forefficient patient recovery.

It was suggested that the hospital factor thatsignificantly accounted for this relation of effec-tiveness to stability was the quality of the humancontacts within the hospitals. These contacts areof two kinds, factual and emotional; both are im-portant. A nurse may need to know somethingabout a patient's treatment and so have need forfactual guidance. She may he afraid to ask forsuch guidance and thus need emotional support.

When we tested these suggestions in the actualstudy we found some interesting results, and Ibelieve subsequent research is beginning to suggestthe results we found there are probably universal.The hospitals we could compare were similar in

145

size and age and set in identical semblance of sur-rounding population. The resulting comparison oflike with !E,e showed conclusively that hospitalsthat could keep their nurses, the student nurses,could also keep their senior nurses. Even theirmatrons stayed longer, and this was also associatedwith a short length of patient stay. The hospitalwith a highly stable nursing staff discharged theirpatients most quickly.

We examined this in many details. We found,for example, that hospitals with short length ofpatient stay had very short waiting lists. Youwould think patients go through the hospitalquickly because many people are waiting at thedoor to take their beds. But that isn't so. Thehospital that discharges patients :quickly has ashort waiting list. As the operations researchpeople know, if the stay time is short the meantime of waiting is small.

This led us to examine very carefully a sampleof five hospitals. We examined them in great de-tail. We asked what ward sisters did. The headnurses, as I think you call them, seemed to us tobe the key people in the system. We examinedthe communication patterns of these hospitals. It .

was clear that the causative variables that ex-plained staff stability and patient recovery werevery closely connected with the intelligibility andfriendliness of the communication inside the hos-pitals. We actually established methods of measur-ing what communication was.

There is not only the dimension of fact butalso perception, anxiety, fearcall it what youlike. My thesis was that if we could improve thegeneral level of communication in hospitals wemight do two things: improve staff stability andreduce length of patient stay. This was the hy-pothesis that appeared in the published report,Standards For Morale. 1 tested it outside of Lan-cashire by examining 15 hospitals all over theUnited Kingdonr. Scotland, London, Yorkshire.We found high concordance among five things.

One was the communication between the wardsister and the power structure; how approachablethe ward sister thought the doctors, the matron,and the administrator were. What the clarity ofcommunication was down from the ward sister tothe nurses and the patients was another considera-tion. We measured both of these in the 15 hos-pitals.

PSYCHOSOCIAL FACTORS IN HOSPITALS AND NURSING STAFFING

NVe also measured the average length of stay ofall qualified nurses in the 15 hospitals. We com-pared this with the mean period of discharge ofsurgical and medical cases for a limited numberof rather simple. straightforward, uncomplicatedcases, like asthma, bronchitis, and pneumonia. Outhypothesis that good communication was in someway associated with staff stability and patient staywas borne out and confirmed when we shifted awayfrom the Lancashire semblance.

All of this appeared in SfundardS For Morale.King's Fund Hospital Centre of London thoughtit a good idea to follow up some of the statementsmade in Standards For Morale. The thought was

139

that improved communications could improve theattitudes of stall and the effectiveness with whichthe hospital is run.

It was the view of the author that the improve-ment of human contacts, both factual and emo-tional, both operational and personal, was bestachieved by action and responsible cooperation inefforts to understand and to improve those con-tacts where they existed. Since the HIC Projectstarted such attempts to improve human experi-ence, the term "action-learning" has been intro-duced. This dual nature suggests that to under-stand a proposition one must be able to apply itin practice.

The HIC Project

NVe persuaded 10 hospitals to join a consortium.The expe.ment design was straightforward. Wechose front each hospital two teams of threepeople. The senior team included the nursingsuperintendent, the chairman of the medical com-mittee (senior doctor), and the hospital secretary-treasurer.

The junior team consisted of an assistant to thematron of some order, a deputy matron or a seniordepartmental sister, the hospital secretary -treasurer, consultant, (but normally he was a reg-istrar, the level below the consultant).

For the senior team, the top three, we ran a 3-day conference. We talked to them about thenature of internal communications in hospitals aswe then understood them. We got together to talkabout their perception of how the hospital . heldtogether as an organism, what the network of com-munication was that turned a large mass of peopleinto a staff. What was it about the hospital thatgave it this quality of nurse stability or length ofpatient stay that discriminated that hospital andother hospitals?

Looking back on it, I can see we spent far tooshort a time working with these people, the topthree people at the hospital. L.4t the junior teamspent a month together in a hotel at Hastings, andthere we tried to go a little farther. We tried togive them a smattering of methodology: how toconduct an interview, to draw a flow process chart,to design and score a questionnaire. They inter-viewed each other to practice their skills of

wheedling information out of people, measuringopinions. And during that period we sent them towork for a week in a hospital other than theirown.

During that week they tried to put into practicesome of the skills, the efforts, the practices, thetechniques that they had learned while at the sea-side and tried to identify within a week whatseemed to be the problems of communicationwithin their host hospitals that gave the hospitalsmost concern.

When they came back we had further confer-ences. Each of the receiving hospitals had beenacquainted by its visitors with what seemed to bethe outstanding problems. The receiving hospitalschose particular projects that they would work onwith their own team. For the next 3 or 4 yearsthese hospitals among themselves selected quite alarge number of projects that they worked on,sometimes with the advice of the teams that hadvisited them in the first place, sometimes with thehelp of a small team that I assembled at Guy'sHospital. Much more frequently they worked withthe help of students of sociology, nurses in train-ing, all kinds of people, university students whointended to be nurses or who were interested inthe sociology of hospitals. A very large number of

people lent a hand with the study of what wasgoing on in these 10 hospitals.

The Problems of PerceptionThis direct invocation of the psychosocial genre

146


of the hospital turned out to be the most evasiveand uncertain departure. A large number of per-sons at all levels of responsibility in the hospitalsand, indeed, in the central research team guidingthe project, confessed that they never understoodwhat it was trying to do. Only when a group ofpersons at a hospital really determined to see whatthey could learn of their problems by energeticallytackling them did the mission of the project be-come clear, both intellectually and emotionally.

Such, of course, is true of all psychosocialphenomena. There must be some leaven to en-ergize the lump. Even reviewers of the books thathave appeared about it say its aims were nevermade clear, despite the fact that these were setforth in writing for all to read. A resume of theaims of the HIC Project is attached as appendix I.Such action-learning projects have since beenstarted in many other places, mostly outside ofBritain. In appendix II the responses to such ef-forts in an American university are analyzed andshould be helpful to those concerned with action-learning in hospitals.

Now, what did this effort actually achieve? Ithink it is fair to say that where from the outsetthe senior team in the hospital believed in thegoals of the project, then the results were highlypositive. The attitudes of the staff were favorablychanged and the project on which they workedachieved much of what it set out to do. Wherethere was no confidence, where there was skepti-cism from the start, then very little happened.Sometimes, in the course of the study, the personsconcerneddoctors, nurses, and administratorswere converted. They saw what these notions. oncommunication really meant when they themselvesbecame engaged in it.

The union of these ideas in studying communi-cation in this sense anti the scientific operationalresearch development have produced quite a re-markable result. Where people had become in-volved and were expected to contribute ideas to

the solution of the problems, two things happened:their attitudes toward the hospital as a whole im-proved, and the system itself improved.

They improved the communication system atthe level of attitudes. They improved. it at thelevel of operations. They found that if people be-can.e operationally involved in attempting to im-prove the system that was giving them anxiety,they could, in fact, improve the system and theycould change attitudes toward the institution.

Psychological Factors

"... the feeling of people concerned with thisstudy has been that a happy hospital is the bestguarantee of a nurse completing her training. Thepractical problem is to achieve this by attention tothe difficulties that beset sisters and the sisters'attitudes to the junior nurses, by providing efficientnursing management, by helping the nurses toorganize an adequate social life, and by arrangingsupport for the nurses when they have emotionalproblems. Some of these are easier to do thanothers, but a start has been made with all ofthem... (1)."

Here, in less than one hundred words, is thesubstance of this paper. More fully, it describes aseries of efforts, made both within hospitals andoutside them, to identify, understand, and resolvesome of the human problems bedeviling the pro-vision of patient care. Whether the results thatflowed from these efforts were promising as wellas meager, or whether they were simply meagerbecause the approach was misconceived, is a matterof opinion. The efforts were either part of theHIC Project or part of a series of monthly dis-cussion groups organized at the Hospital Centrefor nurses from hospitals in and around London.Both of these efforts were strongly supported bythe King Edward Hospital Fund for London andthe first by the Ministry of Health.

Some Illustrations of Psychological Development

"Our results indicated that the attitudes andopinions of the nurses in the ward and those ofnurses in charge of wards are closely connected.Without a doubt, the attitudes of the nursing ad-

ministration, in turn, have some effect on wardsisters and charge nurses. We attempted not onlyto alter procedures and attitudes at ward level butalso to involve the senior nurses more closely,

147

PSYCHOSOCIAI. FACTORS IN HOSPITAI S AND NURSING STAFFING 141

hoping for sonic modification or change in thetipper levels of nursing hierarchy. These nurseswere involved in at least 30 meetings to discussdifferent aspects of their work. These became in-creasingly uninhibited and healthy. At two meet-ings held with H1C team members to discuss thewhole project with especial reference to the seniornurse's involvement. many nurses spontaneouslysaid that the meetings helped them to reflect ontheir own ideas and attitudes. Some gave examplesof how they had attempted to overcome problemsin their particular areas, indicating some changeshad occurred. How far attitudes changed through-out the hospital could he assessed by repeating thesurvey in the future (2)."

We give in this section a few illustrations of howthe involvement in action-learning studies seemedto those at the cutting edge of the hospital system.The examples are in the alphabetical order ofname of the hospital from which they are drawn.The first was written by the matron and herdeputy at Edgware Hospital.t

Accident and Emergency Department

In response to complaints received about theaccident and emergency department, many ofthem apparently justified, the HIC decided to in-vestigate. Finding that the casualty department andthe casualty theatre were controlled by differentheads of departments (one of whom was about toretire) with followup dressings controlled by athird, the team decided to perform an in-depthstudy of the problems, conducted in a way cal-culated to involve all grades of staff in the de-partment.

The team's objectives were as follows:(a) to provide a comprehensive casualty service

despite the physical limitations of the depart-ment,

(b) to establish continuity of treatment and pa-tient care,

(c) to provide an improved practical compre-hensive training fcr student and pupil nurses,

(d) to reduce patients' waiting time to a mini-mum,

Vick. Pauline, and Merchant, Barbara. "Study of theAccident and Emergency Departmert." Hospitals: Corn.

municationJ, Choice and Change, 1972. The following accountis taken entirely from this chapter.

(e) to enhance the public image of the hospital.Following detailed explanations of the survey

by the HIC teams, comments by department staffmembers were solicited and 1,750 of these were re-corded in 38 interviews. When agreement had beenreached as to organization and as to physical orstructural problems, these were referred to the re-sponsible administrative, medical, or nursing pol-icy makers for action. In subsequent meetings withHIC team members, many solutions were contrib-uted by the staff and were put into practice im-mediately. Redeployment of staff resulting fromthis has not only improved patient care but hasprovided an improved training atmosphere forstudent and pupil nurses. Nursing responsibilityhas been more dearly defined, and discussions havebeen initiated concerning medical and nursingpolicy and procedure. A great deal of the successmust be attributed to the fact that the chairmanof the hospital subcommittee gave her supportthroughout the study, attending many feedbackmeetings and several presentations by HIC teamsof other hospitals participating in the project.

There are several outstanding items of interestin this account. First, the impending retirement of

the head of the department offered the oppor-

tunity to change the system. Second, the exercisewas planned both to modify a system and trainthe staff. Third, from the very outset it was seenby the subordinate staff to be supported by those in

established positions of power. Fourth, the mean-ing of the term "communication" became clearonly after vigorous efforts had been made to im-

prove it.

148

Hillingdon Hospital (3)

"My own view, for what it is worth, is that theHIC Project suffered from too much pressure toproduce results by its techniques, when in the hos-pitals there may have been more pressing prob-

lems that had to be solved by other methods. This,

I believe, was the case at our hospital where,during most of the project period, we were con-cerned with the overriding task of completing,equipping, moving into, and commissioning the

new hospital building. It is only since the movehas been completed that we have begun to useHIC techniques to solve a real problem, the oneconcerning nursing staff described by my col-leagues.


"There was a lack of definition of the purposeat the Hastings course. It seemed to be a collectionof interesting lectures with an insufficient centraltheme, I believe that the hospital people on thecourse could have contributed more by cooperat-ing with the research workers in devising ways inwhich communication in hospitals might be im-proved, and I think too much emphasis was placedupon demonstrating the connections between poorcommunications and other short-comings at theexpense of not devoting sufficient time to devisingpractical means of improving matters."

These are entirely fair and constructive criticismsof a poorly tinted effort to use what might in otherconditions have been reasonably successful. Amajor upheaval may not necessarily be the besttime to try out new methods of treating old prob-lems. The opportunity so apparent in the previousexample was missing here, Indeed, the move intonew premises is seen as countereffective. Anotherhospital reported that the same pressure was bothnecessary and constructive.

"One other aspect of the central team's role hasbeen of great importance to all teams. This wasthe pressure exerted on hospital teams to reportback on their activities and progress. Without thisadded stimulus many of us would have done verylittle, especially in the early stages. We still re-member that after the training course at Hastingswe were expected, within a couple of months, tostate our intentions for the future. How manyteams would have any future had we been allowedto proceed at our leisure (4)?"

This, contrasted with the previous extract, em-phasizes the profound differences of perceptionthat can arise between colleagues engaged in sim-ilar missions. The lesson for the psychotherapistis to choose well his occasion. Finally, the objec-tion made by Mr. Bardgett that the Hastingscourse lacked.definition of purpose runs throughmany commentaries. As is shown in appendix II,it is a very general objection. But we note that thestaff , Edgware Hospital grasped what communi-cation meant when they tried to improve it. Thereis, unfortunately, no royal road to explaining itin advance. Our next illustration makes the pointmore clearly.

The London HospitalThis was the only teaching hospital among the

10 in the HIC Project. It is marked out as havingits own research units, including one to promoteoperational research within the hospital itself. Itis a unit with a wide reputation and, for example,is lending its computer to the researchers into theservices for the mentally handicapped.

Mr. Michael Fahey, the deputy house governorof the hospital and author of most of the sectionof the report concerning the London Hospital, atthe outset shared Mr. Bardgett's confusion as tothe objectives of the project.

"The hospital first became aware of the WCProject through the King Edward's Hospital Fundfor London. It is difficult to say what members ofthe team expected of the project, .since its purposewas not entirely clear. It might well be that thiswas because they were not able to attend the twobriefing sessions Meld at the beginning of the proj-ect. The mystery was not noticea!,te dispelled dur-ing tite progress of the month's course at Hastingswhere a similar confusion was apparent. Membersof the team formulated a working hypothesis thatthe aim of the project was to instill methods bywhich to perceive and study problems within alarge organization, and that the success of this at-tempt would be evaluated. In that event, this hypo-thesis proved to he not too far from the truth."

At least one person has given a lot of thoughtto the implications of these lines, and their lessonha5 been well marked fir future applications ofaction-learning. The project chosen at the LondonHospital aimed to improve the emergency admis-sion of patients.

It is of great interest that the London Hospitalshould have tested the hypothesis underlying' theentire project, that responsible involvement in theidentification, analysis, and treatment of perceivedproblems should lead to changes of general atti-tudes towards working in the system manifestingthe problems. In the words of the hospital's ownreport, "There were highly significant changes inattitude in all those areas where alterations to thesystem had been made. In marked contrast, in theone area where no changes had been carried out. . . opinion remained remarkably stable."

North Middlesex Hospital

The report of Mr. Alderson, the assistant groupsecretary at this hospital, shows the difficulties

149

PSYCHOSOCIAL FACTORS IN 1-10SPITAIS AND NURSING STAFFING

that confront those vho ny to stimulate action-learning projects for the first thne %vit bout beingable to illustrate their goals with visible evidenceof success elsewhere.

"Perhaps the outstanding feature of both ourearlier studies was the general lack of enthusiasmon the part of the staff concerned, comb:ned witha certain amount of skepticism and even suspicion.. . . Perhaps what we need, to achieve greater staffparticipation, is the completion of a survey show-ing tangible proof of success, for it is on resultsthat are of obvious and lasting benefit that themajority of staff judge our efforts in this field. . .

"So although our use of HIC methods has notbeen very extensive, we are not entirely disen-chanted, but would prefer not to undertake morethan we can cope with, even though the help ofexperts is at our disposal. As to the future, wehave hopes of being able to continue to do some-thing to meet staff wishes in the catering field, andthere are plans for a small departmental projectupon which our thoughts have not yet fully crystal-lized.

"Beyond this, we shall continue along the pathalready mapped out, but perhaps with diminishedenthusiasm and no longer any great belief in thetheory that a problem has only to be evaluated tobe solved."

NIr. Alderson was not alone in his views. Dr.G.S.A. Knowles, the senior medical member ofthe team, had equal difficulties.

"I must admit that I joined the project withconsiderable skepticism. The stream of verbiageemanating from the central team overwhelmed meboth quantitatively and qualitatively, and it didnot matter how often I read it, I still did notunderstand it properly: the individual words hadmeaning, but when strung together, they com-nunicated little. . . .

"As far as I am concerned, there have been fewpositive results to get excited about so far. Staffof all grades tend to be uninterested or suspicious.Unless there is some money available to make thenecessary changes that are suggested by an investi-gation, further effort is a waste of time."

Dr. Knowles speaks here for a large number ofhis colleagues, for many senior nurses and, pre-sumably, also for a majority of hospital adminis-trators, The problem of convit.7ing senior mem-bers of the staffs of institutions, without immedi-

145

ately relevant evidence, that their everyday opera-tions can be seen as learning projects may proveinsoluble. Without such belief no opportunity canbe engineered to give the hypothesis the more-than-favorable launching that it will always de-mand.

Several ls Hospital

We have quoted from the report of this hospitalat the head of this section. Mr. D. J. Dean stun-marites the total experience. "Not all of those as-sociated with the project felt able to contributeor become involved to any great extent. Much ofthe workload fell on one or two members togetherwith a university student. For the efforts of peopleinvolved in such a project to be successful, it is

essential that they not only feel themselves to be ateam but are seen as a team by the hospital as awhole. For this reason the initial stages of prep-aration are of crucial importance.

"If this form of mahagement is to succeed, thehospital managers must be fully conversant withthe ideas behind HIC and be involved fully inany projects that are undertaken."

The physician superintendent of this hospital,now at Rochester, New York, helps to clarify whatthe HIC Project should have done and very muchin terms of psychosocial factors in nursing.

"Communication occurs in two principal forms:(a) Propositional facts and propositions are

transmitted,(b) Emotional feelings and attitudes are trans-

mitted."I hope the Hospital Internal Communications

Project will highlight for readers the importanceof people timethe way people's opinions arelistened to and dealt with or ignored the way theirfeelings are ruffled or respectedthus producingthe enthusiastic and willing worker or resulting inthe sour deadbeat.

"Object time, on the other hand, is time that isspent dealing with things, and although accountants, carpenters, and committees can deal withmaterial things easily, they concern a separate

dimension of hospital function and, when thechips are down, the least important one.

"If the simple platitude that tr 2.s matter andpeople count is again emphasizf u and if a morereasonable approach to the happiness and well-

144

being of staff and patients is observed, then thisstudy will be most worthwhile and my own .supportfor it well rewarded."

West Middlesex Hospital

The team at this large hospital tackled moreproblems than did the teams at any others andmade'a %Aim effort to get to the heart of theirserious problems of nurse staffing.

"At the inception of the HIC Project we wereundergoing a testing time over certain nursingproblems, and like everyone else we were short ofnurses. This provided a powerful stimulus toameliorate the lot of junior nurses, a first move webelieved, in reducing nurse wastage. . . . Thetopics discussed by the nurses emphasized theirpreoccupation with the hospital environment;they fell into six main categories: lack of sociallife and activities, defects of communication,

personal problems, professional matters,patients, and simple administrative deficien-cies. . .

"Interesting manifestations were observed in thegroups. Young girls in their first year of trainingdiscussed in a mature fashion defects and deficien-cies in the hospital that affected their lives, andalthough they .felt strongly about some of theircomplaints, they discussed them with moderationand insight, showing an ;-4pressive perception ofthe hospital structure and of the feelings of theirsuperi. rs.

"Or. the other hand, sisters and consultants, eventhose taking the nurses' discussion groups, weresee unapproachable Outside the groups. It was

A Question

During the 1960's the interest in the humanproblems of British hospitals was heightened bythe publication of a number of critical books,some of which were widely read. Following one ofthese, Sans Everything, a group of people at TheHospital Centre of the King Edward Hospital Fundfor London agreed to stage a st ties of discussiongroups for nurses from both general and psychiat-ric hospitals in and around London. The aims ofthese were suggested in the following paragraphfrom a report of the East Anglian Hospital Board.


evident that during their training many nursesrapidly developed feelings of resignation andhopelessness about the possibility of changing any-thing, or of obtaining redress, and these feelingswere principally responsible for our scheme notbeing conducive to personal counseling.

"At the discussion meetings, the feeling of peopleconcerned with this work had been that a happyhospital is the best guarantee of a nurse completingher training. The practical 1...oblem is to achievethis by attention to the difficulties that beset sistersand to sisters' attitudes to the junior nurses, byproviding efficient nursing management, by help-ing the nurses to organize an adequate social life,and by arranging support for the nurses whenthey have emotional problems. Some of these areeasier to do than others, but a start has been madewith all of them."

Conclusion

Such, then, are some perceived results of theHIC Project. They suggest that, given the interestand support of those in positions of authority, thenurses and other staff within the hospital systemmay work together to improve both the function-ing of that system and their attitudes toward theirwork as a whole. We shall see that the concept ofautonomous learning, or of self-development fromthe group examination and discussion of one'sown problems and progress, whatever attention ithad attracted in the past, still deserves to be moredeeply developed as a form of psychosocial ther-apy.

of Attitudes

"It may not be an over simplification to statethat the care of patients in hospitals . , hingeson the two main factors. The first is the nature,quality, and amenities of the accommodation uteyoccupy, and the second is the morale of the staff,involving as it does their personal attitudes towardspatients in their care. the latter the more im-portant, but inevitably it is influenced by theformer."

We may observe in passing that all researchersinto the subject have revealed highly significant

151

PSYCHOSOCIAL FACTORS IN HOSPITALS ANI) NURSING STAFI ING 145

differences, among otherwise comparable hospitals,in the measures of their nurses' attitudes towardstheir work. It is the existence of these differencesthat suggested the HIC Project. Cannot a processof organisational learning be set in train so that,in the hospitals with the less favorable attitudes,something is done to correct the conditions out ofwhich they arise? The group at The HospitalCentre was less ambitious and confined itself toorganizing a series of monthly discussion groups,eventually spread over 4 years, that gave hospitalnurses of all ranks the opportunity to discuss theirproblems and to give each other support inameliorating or enduring them.

The original invitation to join the group hadin it these words: "There can be little doubt thatthe attitudes of nurses towards their patients andtowards those with whom they work are an im-portant element in the standards of patient careachieved. Very little is known about the nuncs'understanding of their patients' needs, nor do weknow enough about their reaction to the day-to-dayproblems of ward administration as it affectspatients and staff."

The illiscussions set out to develop both understanding and knowledge. To what extent they suc-ceeded is hard to define, as responses were morethan contradictory. A rigorous attempt to measureattitudes during and after one series of seven dis-cussions showed highly significant changes for thebetter. A ward sister who did not cven attend themeetings but read of them in the nursing presswas so inspired that she set up in her own wardher own discussion group. She had been 12 yearsin charge of an acute male rnedli.41 warn of 26beds staffed by two consultants, three housemen,and about 15 full-time and part-time nurses, aux-iliaries, and domestics.

She said, "In the Kings' Fund Reports I wasparticularly impressed that frank discussion couldmodify attitudes. I decided to initiate ward meet-ings. The first problem was time. A ward sisterhas little time on her hands: this is the first at-

titude to throw out. One has to make time. Want-ing to do it means one has to find the time. Havingmade up nrte's own mit-id, then it is possible tocarry it out. I have proved this."

As I see it, this is the key to all attitude changeand to all learning. Our problem is to know whatit is about this particular sister and her relation

to the hospital that precipitates in visible socialaction the slower changes detectable, but notobviously so, in the analysis of attitude changereported below.

Our ward sister concludes: "I would not like togive the impression that discussions of this typecure all illnesses of the ward, nor that they dramat-ically alter our attitudes towards each other andthe patients. But one thing I do know; discussionand frank appraisal like this can make a consid-erable dent into this problem of attitudes. . . Itis worthwhile continuing, and I feel that anyward sister would gain if she were prepared togive the time, energy, and thought to initiatingsuch a group on her own ward. I warmly recom-mend it to anybody really concerned about people,ether patients or visitors, or whether our col-

leagues who work alongside us each day."This woman once more demonstrates what some

of us have long known and what a few in the HICProject actually learned: that one can solve prob-lems both if one wants to solve them and if onerecognizes that, in so emotional a situation as ahospital ward, one may in oneself be a substantialpart of the problem.

The fundamental question remains: Why dosome senior nursing staff think as did this wardsister, and why do others regard not only theHIC Project but the discussions organized at theHospital Centre as all a waste of time? Even more,why do some regard the discussions as destructive?

The following extract from the Hospital Centrerecord shows how difficult it may be for somesenior nursing staff to change their attitudes: "Thisdebate . . . was the culmination of 'several re-quests for students to speak up; the result waspersonal attack rather than reasoned debate onthe points the students had raised. . . That itwas upsetting there is no doubt. How much socan be gauged by a principal tutor apropos astudent with things to say: 'Empty vessels makethe most sound. I have often found that the mostvocal student verges on the psychopathic.' "

1.5Z

The CaKteasts Revealed

It would be hard to find two attitudes moreapparently opposed than those of the ward sisterwho was convinced of the value of frank discussionmerely by reading of the Hospital Centre's efforts

146

in the nursing press, and those o{ the -principaltutor, that guiding light for the future, who iden-tifies a willingness to ask questions as a mark ofthe psychopath. Yet the contrast appears againand again and at many stages.

The group discussions at The Hospital Centreleave us in no doubt about the immense range ofpersonalities involved. It is as if some nurses areso accustomed to authoritarianism and depend-ence that no efforts of a psychotherapeutic kindcan ever change them, at least not in the par-ticular hospital settings into which they havegrown. The point is made by Dr. Tom Caine inhis summary of the first discussion series.

"We have seen that the differences are in termsof attitudes to discipline and organization, atti-tudes to the required degree of personal involve-ment with the patient, attitudes to a formal versusan informal approach, attitudes to free communi-cation and a questioning of the fundamental scien-tific status of work, including the value of theformal diagnosis. We are all aware that thesequestions form the cornerstones of the training ofnurses. A nurse trained in the beliefs and attitudesof one type of nursing, indeed of one institution,may find that she is required completely to reversethem on transferring to another.

"My own feeling about the conference on pa-tient care and our analysis of the data is that wehave thrown up fundamental problems withoutyet finding the answers. . . . All patients awl ellstaff are people and not just diagnostic categoriesor nursing grades. The full recognition of this ispossibly the most therapeutic attitude of all. Per-haps the most important things are to find outhow we lose sight of this truism, how we canrelearn it and, having done so, what we can doabout it,"

All I feel able to add to this is that Dr. Caine'sobservations apply not only to nurses in Britishhospitals but to all the other hunian groups, inall five continents, with whose action orientededucation I am now concerned. As Henry Wallaceobserved 25 years ago, this is the Century of theCommon Man. Whatever his sex may be, he isat least demanding to be treated as an individualperson.


Nursing Discussion Groups and AttitudinalChanges

At the end of the first monthly series of dis-cussion groups, lasting 7 whole days in all andattended by 82 nurses, Dr. Toni Caine, PrincipalPsychologist at Claybury Hospital, undertook amodest evaluation of their effects. He prepared,from a content analysis of topics raised in suchgroups, a questionnaire with five broad .categories.

One of these could be described as "Getting in-volved with the patent." Its terms were "A nurseshould take care not to show too much interestin patients' deeper problems in order to avoidgetting involved . . . . Patients should be discour-aged from developing feelings towards staff mem-bers. . . . Staff being too friendly towards patientsmakes for poor discipline on the ward. . . Pa-tients should not call nurses by their Christiannames."

Another man section tested the nurses' viewsof their relations with the doctors, A third sec-

tion was on discipline, cleanliness, and efficiency.About two-thirds of the total number of nurseswho went through the seven monthly discussionswere able to complete the questionnaire twice. Thesecond test was given several months after theyhad returned to their hospitals.

Score comparisions were rigorous. The usualfive-point scale of strongly agree, uncertain, dis-agree, and strongly disagree was used, but in seek-ing change of attitudes only reversals were marked.In other words, changing from strongly agree toagree was not considered sufficient indication thatthere had been any shift of attitude at all, onlydisagree to agree or vice versa would be admitted.There was a highly significant (below 0.1 percent)change' in attitudes toward more involvement withthe patients and less subservience toward thedoctors. Indeed. for both classes of nurses, generaland psychiatric, there were improvements in at-titudes in all of the five broad categories recognizedin the questionnaire and based on the contentanalysis of the discussions.

153

PSYCHOSOCIAL FACTORS IN HOSPITALS AND NURSING STAFFING 147

Appendix I

The Aims and Intentions of the HIC Project

1. Normally the demands. for order and economyoblige hospital staff to develop procedural rulesfor overcoming operational problems. If no preciserules exist it is expected that those in authoritywill use precedent or judgment to settle whateverneed may arise. Even if the powers and duties ofstaff have not been set forth in detail there willexist within every hospital an expectation as towho should handle the emergent problem.

2. In such conditions staff tend not to examinethe origins of problems, only to allocate responsibil-ity for dealing with them or to attribute blameshould they not be dealt with. Thus there is littletrue learning, certainly no education, the firstobjective of witich is to encourage the ability toask those questions most likely to lead to the solu-tion of a problem. Education aims to develop thegeneral capacity to inquire, to show people how toLet about finding out. Since, it seems, the principaldefect of many hospitals is a lack of awareness oftheir internal problems, education in this senseis of great importance. The project aims to provideit.

3. We are, as I see it, primarily trying to helpthose in charge of hospitals and of their servicesat all levels to use better their own resources inidentifying and tackling their working problems.We choose this approach both because we believethat these are the greatest resources that the hos-pitals have available and because we also believethe hospitals possess at all levels a considerablefund of latent goodwill and ability. Without ex-ception, all research shows that, given the organ-ized opportunity of working through a problem ofconcern to them personally, those who serve evenin the most stressful conditions and who may haveseemed most incapable of coping with them canrise to unexpected heights of insight andresolution.

4. Any first project should on this account he!wen as a growth point, as a kindergarten in coop-eration. It should involve few people, say up to 10,

' Rem+. R. W., Hospitals: Communication, Choice andChange, 1972, pp. 130-132.

154

quite significantly, in collecting data and in dis-cussing the meaning of the data and their rel-evance to the tasks of those involved with them.Beyond this committed core there should he asupporting force of partisans, not deeply involvedbut willing to give help when asked for it andthus ready to be involved should the project beextended into their territory. Beyond these, allother employees in the hospital should at leastknow that the project is going on.

A well designed project should have the follow.ing qualities:

(a) It should demand the cooperation at all stagesof a group of persons whose tasks are inter-locked. A study assigned to one or two indivi-duals, to produce a report for others to imple-ment, is useless. The choice of project, itsdepth and scope, the methods of study to beused, the forms of coordination between in-dividuals, the nature of any report to othersoutside the group, the methods of puttinginto effect the findings of the 'project, and soforth, are all essenti:-.1 subjects for agreementbetween the team leaders and those whosework will be woven into the study.

(b) It should demand not only cooperation at allstages within the group but also an appeal tosources of information or ideas outside thegroup. These may at first he limited to purelytechnical affairs, such as asking for help fromthe WC team on methods of interview or ofdata classification. Or it may turn out that theworking group, in order to make real prog-ress, will need to share the experiences ofanother hospital or even to inquire about thetreatment of similar problems in industry.For this reason the hospital team should, ata fairly early stage, be in touch with any localcollege of technology of commerce with a de-partment of management or administration.

(c) It must be limited in time; from the outsetthere should he agreement as to when par-ticular stages of the work should he com-pleted. Once such a project is started the

148

interest aroused can be so great that anywaiting time may lead to disillusion and hos-tility in the very places where enthusiasm andcooperation have once been generated.

(d) It should, when completed, suggest anotherproject, elsewhere in the same hospital butinvolving largely different persons.

5. In work on the project, the framework ofauthority should be temporarily thrown away.Those involved in it are faced with a new situa-tion; how to define the need, how to collect in-formation how to work with others, how to measureprogress ancl, above all. how to employ their ownabilities. It is as if, by stripping away the hypotheti-cal framework of authority, we are trying to under-stand how individuals think and act as the personsthey are. They are being given an opportunity,like children whose teacher has suddlnly left theclassroom, to compare notes among themselvesas to how a problem can best be solved.

6. The project should teach each who workson it at least the following:

(a) to observe how the group respond to theoriginally understructured situation,

(b) to question what behavior on his part willbest help himself or the group,

(c) to act, whether in words or otherwise, of hisown volition, as in collecting data or in try-


ing out an idea that seems to him to contri-bute to the definition or solution of theproblem,

(d) to sense the effect of his words or actions:he learns how this behavior has affected othersand thus can test the adequacy of his self-understanding.

(e) to generalize from this learning: in the proj-ect he will begin to sense the accuracy orotherwise of his own impressions and of thoseof others; he will see how he and othersrespond to different attempts to influence andto be influenced; :.ow differing goals anddesires, if contrasted and evaluated freely,lead to learning and to improved under-standing of communications.

7. It follows that to be successful the groupsmust be highly informal and the imposition of thetraditional framework of authority upon theirproject must be avoided. I have little fear, nowthat I know some of the senior staffs of the hospitalspersonally, that such imposition will deliberatelyoccur once experience of open discussion has beengained. Nevertheless, its removal must be accom-plished by slow degrees; senior staffs must gainconfidence in free exchanges before subordinates.In this I feel both the HIC and the hospital teamshave a role first to learn and then to enact.

- PSYCHOSOCIAI. FACTORS IN HOSPITALS AND NURSING STAFFING

Appendix II

An Experiment in Active T earning

In 1971 I was approached by the dean of theBusiness School at Southern Methodist University,Dallas, who, with the support of somebut aminorityof his faculty, was endeavoring to changethe principles upon which the school was run. Inthe words of Roger Dunbar and John Dutton, twoof the dean's active supporters:

"At the School of Business Administration,Southern Methodist University, vigorous effortshave been made to educate students for an activerole in society. As a step in this direction, allundergraduate course requirements but one havebeen abolished. Thus stw'ents now must designtheir own learning prc.,am instead of fulfillinga course sequence dictated by the faculty. In thisway an attempt has been made to transform theschool itself into an environment which challengesthe student to act effectively. One required courseremained as a setting in which to force studentsto consider what is involved in designing one'sown learning experience. This requirement ap-peared necessary to the faculty in order to assiststudents to adopt a more Active style of learningbehavior, for students at the school tended to be-have passively in learning situations, waiting for in-structors to generate material rather than seekinginformation for themselves."

The experiment soon ran into some of the prob-lems of the HIC Project: a lack of intelligibilityfollowed by resentment. In the words of its or-ganizers:

"From the instructor's point of view, the purposeand the design of the course were clear enoughand reflected their beliefs as to what was necessaryfor student learning to occur. Specifically, for stud-ents to learn, they would have to be active andresponsible for their own behavior. However manystudents felt lost and confused with this unfamiliarstructural design. Some responded with curiosity,exploring what could be done in the new situation.But others became passive, sullen, or even violentlyangi y because they could not understand what wasexpected of them. These disaffected students seldomfound helpful the extensive written handouts dis-tributed in the course concerning the purpose,

156

149

structure and available options of the course.. . . "Fromm's (1941) discussion of freedom may

be used to help understand the variety of studentresponses to the course. Fromm distinguishes be-tween freedom from constraints and freedom todevelop facilitating structures that allow individualachievement. The absence of many of the expectedinstructor demands provided freedom from manyof the usual course constraints. HoWever, studentsoften showed little freedom to develop new be-havioral responses to utilize the new freedom fromexpected instructor demands. Some got angry be-cause they did not perceive the instructors as will-ing to tell them what to do and a number remainedantagonistic throughout the semester. Others be-came angry because they could not reconcile thefreedom that did exist in the course with anyconstraints whatever; they may have felt that therequirement that they examine. the consequencesof their own choices contradicted the idea offreedom to make choices.

"Still others, originally skeptical, eventually sawthe purpose of the course. Frequently, these in-dicated they only regretted it had taken so longto become aware of the objective. Finally, those whowere already highly motivated to learn reportedthat as a result of the course they were evenmore enthusiastic about their work at school. Insummary, responses to the course appeared to varygteatly and the instructors believe they probablyreflected whether students were able to formulatenew behavioral responses to cope with new de-mands of the learning environment."

Nothing could describe our experiences withthe HIC Project more faithfully. M at S.M.U.,moreover, there were members of the staff whowere unable to grasp this idea of freedom to un-dertake one's own learning.

Two Extremes of Students

In the HIC Project no effort was made to dis-criminate between two types of hospital staff asthe S.M.U. studies discriminate between acceptors


and rejectors of the new learning program. Someof the central team were awareas, indeed, isshown here in the main paperthat some of thehospital staffs had more difficulty titan othersits grasping the notion of self-directed studies.But at S.M.U., where conditions for quantitativeevaluation were possible, a deliberate search wasmade to identify the extremes among the 412students who took the course. Of these there werechosen 32 manifest acceptors and 32 manifest re-jectors of this new method: the discriminatingcriterion was the persistence with which they waitedfor guidance from the teacher and abstained fromexperimenting with their new methods. The mani-fest acceptors neither waited nor abstained; themanifest rejectors did both.

It was then shown, as a result of a survey in-volving a sample of 238 students of which the twosets of 32 are therefore about one quarter, thatthere were categorical differences between thetwo sets. Among the manifest acceptors there weresignificant positive correlations between learningand efforts to assess the feelings of the other stu-dents and of the teacher. Learning among the mani-fest acceptors was thus coupled with an awarenessof others in the group. Among the manifest re-jectors there was no such correlation: instead there

was a significant positive correlation between learn-ing and a desire to be noticed by the teacher.

In the light of our HIC Project experience, theseresults are very plausible. Both the Severalls andthe Middlesex Hospitals stress, moreover, that theattitudes of the junior nurses ultimately reflectthose of their seniors (a result also reported inStandard For Morale). It therefore looks as if ourfirst task is to persuade those in the senior postsof the hospitals to act as manifest acceptors. Any-thing effective in the way of change will demand,at all times, the support of those in authority.

In the many programs of institutional changewith which I have been concerned since the HICProject, I have been more than careful to involvethe coalitions of power responsible for the partici-pating institutions. Not enough attention was givento this in launching the HIC Project. In our latestattempt, to stage a development program for thenationalized industries of Syria, we have not onlysecured the personal interest of the Prime Ministerbut allocated to him an operational role and a time-table to go with it. Since the program was designedat his invitationhaving heard of its success inanother Arab countrywe may he confident that,should our efforts fail, it will not be for lack of highlevel interest.

Impact of Computerized Information Systemson Nurse Staffing in Hospitals

Donald W. Simborg, M.D.Chief, Clinical Systems Development

Johns Hopkins HospitalBaltimore, Maryland

Introduction

To describe the impact of a technological de-velopment such as computerized information sys-tems, it is necessary to define and understand theproblem area to which it is applied. Nurse staffingin hospitals is a broad and complex topic. Forthe purposes of this discussion, nurse staffing willrefer to the following two problems:

How many of each type of nursing personnelare needed on inpatient nursing units?How can the assignments of personnel be madeto fit those needs?

I would like to discuss what impact computer-ized information systems might have on these components. To answer the first question, how manyof each type of nursing pesonnel are needed oninpatient nursing units, four items of informationare required:

(1) What kinds of tasks need to be done by thenursing staff?

(2) How many of each task are there to be doneon any given shift on any given nursing unit?

(3) Which nursing personnel type should do eachtask?

(4) How long does it take to do each task?

From these four items of information, we can de-termine the nursing staff needs on any nursingunit at any time such information is available,

To answer the second question, how can theassignments of personnel be made to fit thoseneeds, we must have the means to provide theitems of information sought in the first questionon a timely basis and must then have the abilityto optimize the distribution of available staff according to those needs. This optimization can oc-cur on a short-term, dynamic basis, or on a long-term basis. The advantages and disadvantages ofeach approach will be mentioned later, I wouldlike now to discuss how all of this might be ac-complished and what role a computer system mightplay.

We have not solved the nurse staffing problemat Johns Hopkins. Howeve: we have partiallysolved a related problem that I think will providesome basis for solving the staffing problem. I wouldlike to digress for a while to describe some of thework we have been doing the past year.

Approximately 2 years ago we became inter-ested in the general problem of management of

151

1b8

152

an inpatient nursing unit. What is the process bywhich the nursing stall carry out their duties, howwell does this process work, and what might bedone to improve it? In general, nurses do twocategories of tasks: those that the doctors tell themto do and those that they decide themselves needto be done. Let us first look at what the doctorstell them to do; i.e., the carrying out of physicians'written orders.

At our hospital. the nurses use the traditionalkardex system of carrying out orders. There areabout a dozen different kinds of orders each car-ried out in a different way. As an example, figure 1shows what happens when a physician writes anorder to have a patient weighed on a periodicbasis; e.g., every other day. The basic steps arethe writing of the order, the initial transcription,the written communication onto some workingdocument, the execution of the order by thenursing staff, and the communication of the re-sult back to the physician. Figure 2 shows thissame process for a vital sign order; i.e., the orderto measure the pulse, temperature, respiration rate.and/or blood pressure at some given interval.Figure 3 shows this process for an intravenousfluid order.

The point of these figures is simply to illustratethe complexity of our manual "system" of carryingout orders. Of the 12 categories of orders westudied, no less than 13 different documents areused to keep track of these orders. One naturallyasks, "Why was the system designed this way andhow well does it work?" It is clear that the systemwas not designed at all. Rather, it evolved. Itevolved over 60 to 70 years' with changesfrequently by different people in different decadesto meet certain expediencies.

In answer to the second question, how well doesit work, we made some observations. During a 2-week period on one acute inpatient medical ward.all physicians' orders were examined. Each orderwas read and evaluated, the transcription wasread, the written communication onto variousdocuments was checked, and on site observationswere made to determine whether or not the orderwas tarried out. This was done 'for the nonmedica-tion orders only.

Figure 4 summarizes the results of these studies.Of the 1,592 observations made, about 15 percentof the orders were not carried out. In each such

, 159

RESEARCH ON NURSE. STAFFING IN HOSPITALS

instance, an attempt was made to determine theexact point on the flow chart of that order that ac-counted for the failure. As well as could be de-termined, 14 percent of the problem could be at-tributed to the physician's original order, 4 percentto transcription of the order, 22 percent to writtencommunication of the order subsequent to theinitial transcription, 58 percent to the execution(actually nonexecution) of the order by the nursingstaff, and 2 percent to notification errors. (Noti-fication refers to those specific orders in whichthe physician requests to be notified of a certainfinding if it occurs.)

Another study performed was a traditional nurs-ing activity study in which random observationswere made to see what nurses did with their timewhile on the ward; i.e., when they were not sup-posed to take a break. Our results are similar toother such studies and are shown in tal,le 1. Anurse spends about 30 percent of her time at thebedside and about 40 percent of the time doing in-direct care activities such as paperwork, supplyfunctions, and preparing medications.

From these observations of the problems it waspossible to determine some requirements for theirsolution. This process of determining functionalrequirements to solve problems is 9 creative orsynthetic process, whereas everything' prior to thispoint was an analytical process. Thus, as in anycreative process, there is tremendous latitude as tothe types of solutions that could be derived, andthere is considerable judgment as to the initialsolutions to try.

For each problem identified in the analysis phasea number of possible alternative solutions wereconsidered. The details of this process will be re-ported elsewhere. The results of this syntheticphase were a set of 14 functional requirementsfor a new system. These are shown in figure 5. Letme call your attention to requirement 6, to allocatenursing staff dynamically according to needs. Thisis the subject of this entire conference, yet notethat it is only one of 14 requirements we had forsolving the problems of managing our wards. Thisparticular requirement is related primarily to oneof the problems found in the analysis phase, thefact that 58 percent of all orders not carried outwere due to the lack of execution by the nursingstaff.

One reason the nursing staff do not execute

IMPACT OF COMPUTERIZED INI.ORMATION SYSTEMS ON NURSE STAFFING IN HOSPITALS 153

Figure 1.Flow chart of weight order.

Weights

Order

BEST COPY AVAILABLE

Reviewed byRN

Pink ticket Treatmentif special kardex

Weighed byspecial

Recorded inspecial chart

160

To weight bookeach morning

Nursing staff weighseach morning as perweight book

Recorded in weightbook by nursing staff

Transferred to graphicsheet by ward clerk

154

BEST COPY AVAILABLE


Figure 2.Flow chart of vital sign order.

Treatmentcard file

Order

TPR Sheet

Page made in Nursing staffBP book reviews andpink ticket(s) performs

Physiciannotified ifnecessary

Reviewed byRN

Pink ticket forBP , specialTPR ,frequentTPR

Nursing staffrecords on TPRsheet and inBP book

Ward clerkgraphs

Pink ticketfor notifica-tion

JTo bedside forspecial nurse

Special nurseperforms

Special nurserecords onspecial sheet

T=Temperature, P=Pulse, BP=Blood Pressure, R=Respiration,

RN=Registered Nurse

161

IMPACT OF COMPUTERIZED INFORMATION SYSTEMS ON NURSE STAFFING IN HOSPITALS 155

Intravenous fluids

Transcribedto 4-partI .V. ticket

Gold copyremains onward

CheckedPRN(infrequent)

Thrownaway in1-3 days

Figure 3.Flow chart of intravenous fluids order.

Order Reviewed by RN 1

Call to.I .

therapy ifstart needed

BEST COPY AVAILABLE

3 parts toI .V . therapy

Yellow copyfor billingI .V . therapy

tA .M nursepreparesI .V. work-sheet

I .,/ .s re-ported innursingreport

Occasionalorder tokardex

White copyfor adminis-tration I .V.therapy

Solutionsent tofloor (ifno additives)with pinktick& label

Pink copyto pharmacyfor additives

Bottle hungby nursingstaff

Solutionsent tofloor withadditivesand type -written label

.Imisharrirsomrrippii

16 .0

Patientbilled

Bottleregulatedby nursingstaff andI .V . therap

BEST COPY AVAILABLE156

Figure 4.Physicians' orders study1,592 OnWard Operational Observations

234 FailuresI5%I. Physician -14%2. Transcription-4c',3. Written communication-22%4. 1.1xecution-513%

5. Notification-2%


Figure 5.Functional requirements of a new system to carry out physicians' ordersI. Provide a legible. convenient summary of the physicians' current orders in a place where he is likely to look

at them daily.2. Provide a rot iew process at the input phase of physicians' orders in which an RN verified the 'input.3. !lave physicians write orders on renumbered order sheets. Provide a mechanism to account for every order number,4. Eliminate manual communication.5. Schedule nursing tasks so as to reduce peak workloads.6. Allocate nursing staff dynamically according to needs.7. Provide a single document from which nursing staff can see at a glance all tasks needed to be performed at any given hour.A. Prot ide a mechanism by which the head nurse can monitor what tasks need to he done in the coming hours as well as

what tasks were not carried out during the previous hours.9. Indicate notification orders in the place where the specific patient recording must be made.

10. Make all recorded patient observations at the bedside at the time of the observation.II. Avoid transcription of recorded patient observations.12. Summarise all recorded patient observations in a convenient format in a single location for the physician to review

at the time of morning rounds.13. The nets. system should require less nursing staff administrative time to maintain than the present system.14. The costs of the system must be appropriate for the benefits provided.

16J

IMPACT OF COMPUTERIZED INFORMATION SYSTEMS ON NURSF STAFFING IN HOSPITALS 15i

orders is that they are too busy, hence the needto staff appropriately. There are many other tea._sons why nurses 10 not carry out their tasks thatwould not be solved by appropriate staffing, hencethe need for other requirements. I would like to callymn attention to several other of the require-ments on the list, particularly 7, 8, 13 and H.We needed to provide the nurses with a simpleway to know what tasks were scheduled to be doneat any given hour of the day. This is again re.lated to the execution problem, since some ordersare not carried out because people forgot and othersbecause some.people may never have become awareof the need to carry out the order in the firstplace. Also, there is no easy mechanism for thehead nurse on the floor to monitor or superviseher staff. Finally, any new system that we implementshould reduce nursing. staff indirect care time, orat least not increase it. And, of course, the hospitalmust be able to afford the costs of the system.

With these and other requirements in mind,a new system to carry out physicians' orders wasdesigned and implemented on one of our acutemedical nursing units. This is a computer basedsystem that I will describe only briefly. This sys-tem meets all of the requirements listed exceptthat we do not yet staff our floors dynamicallyaccording to needs, nor do we reschedule nursingtasks. We feel that we have the means to developa dynamic staffing system based on what we are nowdoing.

Figure 6 is a diagram of our current system.Twice a day we produce output documents. This

output is listed in figure 7. Again, 1 do not wish todescribe the entire system but just that part whichmay be relevant to nurse staffing. The most im-portant information in this regard is illustratedby one of our output documents called the "TeamAction Sheet." An example is shown in figure 8.Our unit consists of three nursing teams each withapproximately. 10 patients. For each hour of theday for each team, the action sheet shows alltasks that must he performed on any of theirpatients at that hour as determined by the orders.

This output document, as well as all others weproduce. was not designed to solve a nurse staffingproblem. The primary goals of this system wereas follows:

(a) reduce errors in transcribing and communieating physicians' written orders,

(b) reduce failures in carrying out appropriatenursing tasks,

(c) reduce nursing staff administrative time.

We evaluated the system on the one nursingunit by repeating the studies I showed you earlieron a floor with the traditional kardex system.Table 2 summarizes these repeat studies. As youcan see, each of the original goals of the systemhas been met.

Now does all of this relate to the problem ofnurse staffing per se? Can this system and othercomputer information systems have any effect onthat? To answer these questions, let us look againat the questions which we said earlier neede:! tobe answered.

Number of Each Type of Personnel Required

Kinds of Tasks To Be Done

We can see From the output shown in figure 8(and some other output documents not shown)what kinds of tasks nurses do related to physicians'written orders. The nursing staff, however. performother tasks or duties not related to the care of thepatient such as bed changes and patient hygiene.Other tasks that the nurses initiate are specificfor the particular problems of their patients.

We have looked at these tasks and have instituted a procedure in which the nurse makesan assessment of the patient's nursing needs and

1 64

develops a nursing care plan as usual. In addition,however she writes a set of nursing orders in theorder book just as the physician does so that re-minders will appear on the Team Action Sheetto perform those tasks as well as the physiciangenerated tasks. Table 3 is an exmple. If thepatient has a nasogastri tube in place, the nursewill write the older "NC Tube Care." This willautomatically generate the message shown in table3 at the time indicated.

These messages and times were derived by thenursing staff as minimum procedures consistentwith good nursing care. Nlodifiation of the "stand-



Figure 6.General flow chart of new system.

Physician writesorder on pre-numbered ordersheet

Carbon copycollected by

Information Clerk

Orders transmittedfrom paper tapeover telepnoneto computer

Original copyremains inorder book

Orders c.ciF.ed bycomputerprogram

Order enteredonto terminal

Unit Clerkmanuallyaction sheetif necessary

Punched papertape produced

Copy ofaccepted ordersgiven to nursefor verification

Errors returnedto InformationClerk

I

jCorrectionsretransmittedto computer

Hard-copy recordof input producedon teletypewriter

1101.Output documentsproduced bycomputer

Figure 7.System output documentsI. Action Sheets

A. By TeamB. Special PatientC. By Function

I. Diet List2. Utility Room List3. Weight List4. Specimen List5. Intake and Output Sheets6. Patient Lists7. Tetnperature.Pulse Sheet

II, Standing Order Sheets .

i63

IMPACT OF COMPUTERIZED INFORMATION SYSTEMS ON NURSE STAFFING IN HOSPITALS 159

Figure 8.Team Action Sheet.

BEST COPY AVAILABLE

ACTION SHEET FLOOR 03 TEAM A 03/06/72

10 00P 03/06 ilIR I I I I

SlOREOERBE I I I I I

.....1---- I10 OOP 03/06 IDRESSINGsIippo ICLEAN IIN CNAIR I

GLOBE,HERSE IL ANKLE 113 KIN IL FooT I I

____1011_21___I_________ILHISDUILI

IRESP

10 OOP 03/06 lop IRESDJOHNSON,RUS I I

I I

I I

I

I

I

I

I

I10 OOP 03/06 'CHEST PT 16LOW6LovEIJONES,LOCKS I IS MIN

J

I

I I

I t

I

I

t

I

I

1--I10 OOP 03/06 (WALK I

KANE,JoHN WITH I

(HELP I

1 I

I I

I 1

I

I

1

I

I

1

10 01P 03/06 16P IRESPKTNG,CHARLE I I

I i I

I I

I I

I

I

I

I

I

I

10 OOP 03/06 IIPPe I I

KINO,CMARLE I I I

I I I

I

I

I

I

I

1 I

I

I

I

10 OOP 03/06 IAMIULATE IMANrIvENTILAWSEN,R000 MTH I I

.-----IASSISIABC

I

I

I

I

I

10 OOP 03/06 IIPP6 I I

ROONSIROLAN 12CC NS I

Ira5

I

I I

I

I

I

1

10 OOP 03/06 19P IRESP I

SIRONG,LAWR I I

I

I

I I

I I

I

I

A

I

I

I10 00P 03106 I I

I

1

I I

I I

t t

I

I

1

I

I

1

10 OOP 03/06 I I

I

I I

I

I 1

*URINE SCATI I

I

I

i

I

I

I

I

I

I

I

I

I

I

I

IMIO

11 OOP 03/06 1 I

OLOSE,NERBE 1 I

1 1

11 00P 03/06 I

6LO6E,NERIE!

111113 CR I

'GREATER I

I I

11 00P 03/06 I I

I

1 ---__----1

I I

I I

I

I

I

I

I

I

166


ard" formats can be specified at the time the nursewrites her orders. The combination of nursingorders and physicians' orders does not include allof the tasks done by nurses, but it does cover thegreat majority of the patient related variable tasks.

Number of Each Task To Be Done

This has been the most difficult aspect of de-termining nursing needs in the pastthe quanti-fication of the workload. In the system I have de-scribed, we are informing the nurses, on an hourlybasis for each nursing team, which tasks need tobe performed. It is a simple process for the com-puter to add up these tasks according to any cate-gories desired. Thus we have captured in the com-puter, prior to arrival of the nursing staff, a quanti-tative estimate of a major part of the nursing staffworkloadindeed the part related to the most un-predictable variablethe specific need of eachpatient.

The estimate is not based on patient classifica-tion. It is not based on patient diagnosis. Rather,it is based upon the most appropriate determin-ant of nursing staff needs for each patientthespecific configuration of tasks needed for that p-tient regardless of diagnosis or broad classification.It does not require any extra nursing time to pro-vide. The basic input is unrelated to nurse staffing.It is related to the actual process of patient care,

Assignment

One approach would he to measure these nurs-ing staff needs daily for a long period of timeand determine the range and mean of nursingstaff needs for each nursing unit for each shift.Since we are not yet performing a nurse staffingfunction, we have not yet writt:n the program todo this. However, just out of curiosity, we wrotea program that simply adds up the total numberof tasks done on our experimental nursing uniteach day. I ran this program on a sample of ratherrandomly selected days and the summary of the out-put is shown in table 4.

For the sake of discussion, let us assume thatthe number of staff needed on each of these daysis directly proportional to the number of tasks tobe performedan obvious over simplification, One

167

thus ensuring a greater probability of accuracy ofinput. This workload determination can be cal-culated for each nursing unit, nursing shift, nurs-ing team, and even hour of the day if necessary.

Type of Personnel for Each Task

This must be a nursing decision and will notbe done by an information system. However, insituations in which the workload to be done ex-ceeds the number of personnel available, the com-puter could be used to specify which tasks shouldbe done by which personnel (assuming that manytasks can be done by more than one personneltype) so as to maximize the amount of workactually done. More will be said about this later.

Time Required for Each Task

This will vary from situation to situation for agiven task but, generally speaking, measurementshave been made on the variety of nursing tasksso as to arrive at reasonable average tildes. Havingdefined the tasks to be-done, quantitated the tasks,associated each task with a personnel type, anddetermined an average time to perform a task,one then has a statement of nursing staff needs.After doing these calculations for each shift oneach nursing unit one can then consider staffassignments.

of Personnel

way to staff this unit would be to pick some arbi-trary level of preparedness; e.g., to be able to carryout all of the tasks 80 percent of the time. Thisis more or less the way we staff most hospitalstoday. By definition, we are understaffed a certainamount of the time and overstaffed a certainamount of time. The computer system just de-scribed could be used to staff in this manner, theonly difference would be that we would know whatthose proportions of under and overstaffing were.I am not sure knowing this has much value..

Suppose, however, that we have two or moresimilar nursing units. Let us further suppose thatthe day-to-day variations in nursing staff needs arenot completely synchronous between these two ormore units. That is, although the general load of

IMPACT OF COMPUTERIZED INFORMATION SVsTFNIS ON NURSE STAFFING IN HOSPITALS 161

Table 1.Nursing activity study BEST COPY AVAILABLE

Nursingpersonnelcategory Direct care Indirect cate Other

Percent Percent PercentAll 31 38 33

Nurse ... 26 44 30

Technician 45 23 32

Aide _..___ .._. 32 33 35

Table 2. Summary of repeat studies on,Aperimental floor

ObservationTraditional

systemExperimental

...R1

Transcription error rate ____ 3.4% 1.7%Communication errors _____ _ I9/week 3/weekFailure rate in carryingout orders 14.7% 5.8%Direct care activities(all personnel) _....- 31% 38%Indirect care activities(all personnel) __. _ 36% 28%

Table 3.Sample nursing order: "NG Tube Care"

OrderTimes of

Message on action sheet message

NG tube care Check patency NG tube Q4HMouth and nasal care Q8HChange adhesiv tape QDEmpty NG tube drainage Q1214

Notify house officer ifbloody .rainage QI2HRecord NG drainage QI2Hvolume

Table 4.Total number of nursing tasks on onenursing unit on selected days

Date Total tasks

7/30/71 608

10/29/71 1,109

10/30/71 1,I88

11/11/71 957

11/12/71 748

11/15/71 652

11/21/71 527

12/9/71 516

12/10/71 797

12/11/71 611

12/15/71 804

12/19/71 627

12/30/71 2733/18/72 8474/11/72 397

Range 273-1.188Mean _ 677

162

patients will vary roughly in parallel between theseunits, day-to-day variations in the particular kindsof patients will lead to unequal workloads on theseunits. Having the ability to measure these workloadson a real-time basis as our system may be ableto do, a flexible or dynamic staffing system basedon this ability to monitor workload is possible.Thus, although the total number of nursing staffpersonnel is fixed, their distribution among thenursing units will vary from day to day.

In addition to providing a real-time measureof nursing staff needs, can the information systembe used to help in distributing the staff? Withthe help of Dr. Judith Liebtr:.n, Assistant Professorof Public Health Administration, we developedthe following model. Suppose that we prefera nursing aide to perform task (i). If an aide isnot available, a licensed practical nurse (LPN)could also perform this task as a second choice andfurthermore, a registered nurse (RN) could do itas a third choice.

Let us then define a coefficient C, such thatC-3 if an aide performs task (i), C.2 if an LPNperforms it, and C. 1 if an RN performs it. Let usmake these judgments for each task. Associated witheach task (i) will be a tinting constant t, related tohow long it takes to perform task (i). If we knowhow many times (n,h) task (i) must he done on agiven shift on a given nursing unit (k), then for anytask (i), nikt, tells us how long it will take to do allof task (i) on nursing unit (k). Let Xuk be theproportion of time that personnel type (j) performstask (i) on nursing unit (k). If we always assign themost preferred personnel type (j) to perfrom task(i), then for that (ij) combination Xiihm 1. For anyother (ij) combination for the same task (i) Xiik100.Then 7E nikt,Xijk is a measure of the total

hours of each personnel type (j) needed to ideallystaff nursing unit (k).

Summing for' all nursing units gives the totalamount of each personnel time needed for all

169


the units. If it so happens that we have avail-able at least enough of each type personnel ourproblem has ended. But unfortunately, the exactmix of personnel needed for all the units willnot correspond exactly to what is available everyday. Thus we can utilize the information con-tained in the coefficients C for each task (i)for each personnel type to modify personnel as-signments. Thus if task (i) can be done by anyone of the three personnel types (j), each onehaving a different value of the coefficient Cii,then I n CijniktiXijk is maximum if the person-nel (j) performing task (i) is always the mostdesired (i.e., with the highest value of CIO .

It is possible that on a given day the values ofnu, will be so high that there is no solution tothe problem assuming we wish to carry out everytask. We are then forced to assume that somethingless than 100 percent of all tasks will be performedand solve the linear programing equations forthis arbitrarily reduced value of nil,. We can decre-ment all n,,, equally by the coefficient F and deter-mine the highest value of F that will allow theequation to be solved. Or we can assign prioritiesto the tasks and in the event that the total work-load on all the units together exceeds the totalavailable staff, no matter how they are distributed,certain tasks will be eliminated first.

This in actuality is what happens every day inour hospitals in an unplanned manner. In eithercase, we will always be able to find some propor-tion of the total workload that can be done bythe available staff and we will be able to optimizethe distribution of that staff to perform that partof the total workload. This optimization can thenhe suggested to the nursing administrator prior toarrival of the staff for that shift.

Many of the ideas expressed here are not original.However, the possibility that we can actually carryout some of these ideas because of informationsystems under development places them in a newperspective.

IMPACT OF COMPUTERIZED INFORMATION SYSTEMS ON NURSE STAFFING IN HOSPITALS

Summary

The impact of computerized information systemson nurse staffing in hospitals occurs in the areasdescribed below.

They will permit an accurate assessment ofnursing staff needs on inpatient nursing units.They will be able to monitor these needs ona realtime basis so they can he used in timeto make judgments about staff assignments ona shift-to-shift basis.They will provide the nursing administratorswith a statement as to the ideal number ofeach personnel type needed to staff eachnursing unit for each shift. The nursing ad-ministrator can use this information as shepleases. '

They will suggest to the nursing administratoran optimal distribution of her available staffto meet the needs. Furthermore, they can indi-cate which tasks the optimization assumes willnot be done and who it assumes will be doingthoseJasks that will be clone.

What other impact might these systems have?Let us return to table 2 in which we showed areduction in nursing staff indirect care activities.In addition to helping assign the staff to carry outthe tasks, an information system will reduce thetypes of tasks to be done by the nursing staff inthe first place.

Finally, there is one other important factor re.lating to nurse staffing. Once we narrow the tasks

down to those that truly inns performed bynursing personnel, the physicians ar still free toorder as many tasks to be performed as they feelnecessary. In the past, little attention has been paidto what orders doctors write. Recently, however,people are interested in the fact that in seeminglycomparable situations, different physicians orderdifferent services. This can be reflected in the nurs-ing tasks they order. When people began to auditlaboratory test ordering practices of physicians, itappeared that at least in some cases there was overutilization of the laboratories.

With an information system such as the onedescribed here, it is now possible to quantitateand audit the physician utilization of nursingservices. Perhaps over utilization can be detectedhere as well, and appropriate feedback to thephysicians may affect these practices.

To give an overall answer to the question.What is the impact of computerized informationsystems on nurse staffing in hospitals?, 1 say

this:They will help determine which tasks will beperformed for patients in various situations.They will help perform sonic of the tasks, re-lieving nurses of nonnursing functions.They will permit measurement of nursing staffneeds on a timely basis and provide recom-mendations on the appropriate distribution ofavailable nursing staff to meet these needs.

DISCUSSION OF DR. SIMBORG'S PAPER

Discussion Leader

Miss Marlene SlawsonResearch Associate, Operations Research Unit

Division of Community PsychiatrySchool of Medicine

State University of New YorkBuffalo, New York

Miss Slawson

Dr. Simborg's paper has presented some thoughtprovoking ideas, and I should like at the outsetto state my own position: I am not opposed tothe philosophy of the use of computers in cer-tain hospital settings, but I feel we should movewith caution in certain areas.

I will comment on several points of the paper.Dr. Simborg indicated that 58 percent of the fail-ure of tasks noted in the study resulted fromsimple nonperformance. The first question is,

Were these all really nursing tasks or were theynonnursing tasks with the failure on the part of

. the laboratory personnel rather than nursing?Second, Was the failure of the task related to re-sources such as the lack of equipment? Third, Wasthe nonperformance of the task related to thequestion of the nurse, whether or not it reallywas a valid task to be done? He indicated thatone of the reasons the nursing staff does not exe-cute orders is that they are too busy, hence theneed to staff appropriately. Maybe an alternativeshould be to staff more efficiently with what staffis available.

Dr. Simborg indicated also that the computer-ized information system will help perform someof the tasks, relieving nurses of nonnursing func-tions. My question here is, Can another kind ofsystem accomplish the same result for less overalltotal cost? His basic premise is that automateddata processing techniques can be used to enhance

he development of quality performance by proc-essing orders and by increasing performance ofnursing care. What does the addition of a com-puterized information system add that other man-ual systems cannot? Does a computerized informa-tion system increase quality? Does it decreasecost? Does it increase efficiency? Does it increasestaff satisfaction? I did not find presented anykind of evidence that comparative studies havebeen done to indicate that an automated dataprocessing system as proposed by Dr. Simborgcan present quality data to answer the abovequestions

I feel the system proposed here today has cer-tain limitations. While it can simplify the presen-tation of a complex idea, it can also oversimplifyit. And one of the constant problems faced bymany researchers in this area is that in the studyof human behavior there is the necessity of trans-lating complex relationships, concepts, and ideasinto some type of a simpler representation ofform.

I agree very much that it is necessary to con-tain a conceptual description of the system. Thisincludes taking the institutional needs and patients'needs, coupling them with the resources, both ofstaff and equipment, and then evaluating themin terms of quality execution of patient care, pa-tient satisfaction and personnel satisfaction. Thesystem described was based upon physicians' or-

DISCUSSION OF DR. SIMBORG'S PAPER

drib as presented by nurses themselves in termsof their own description. Alternative input thatcould enter into the model are those tasks per-formed by the institution in terms of protocol andpolicy.

For example, patients receive baths that arenot specifically ordered by the physician. Thereare tasks performed as a matter of hospital policy,and tasks related to psychosocial enhancement ofthe patient; these should also be included in theoverall development of a model. Tasks in thelatter group include such processes as alleviatingpatient anxiety and conducting preop psycho-logical preparations. These components do notreally come into the realm of specific tasks butare important in allocating staff and resources.It is well perhaps to have some type of allocationof time in the model for crisis intervention, un-foreseen events that occur on the floor, new ad-missions, and or emergencies.

We should move with caution. When some of

the variables for success have been met, a systemthat will summarize the existing knowledge willprovide an explanation for observed events. It

will also predict the occurrence of future rela-tionships that will he most helpful to the nursingprofession. This will essentially integrate the phys-ical system, human organization, and the informa-tion system into an efficient and consistent overallsystem. I do not feel that provision of an auto-mated system is the answer to every problem in a

165

hospital, but it may serve as a useful tool in ad-ministrative and clinical care.

From 1967 to 1969 I was involved in a costanalysis and quality of care study, and at thattime I was not aware of any successful attempts,through the uses of different changes in staffingpatterns, to reduce the cost of patient care. Maybein the total overall operational costs of the hospi-tal this could be accomplished, but I have yet tosee a hospital that reports to its patients, "We aregoing to charge you $50 less a day because you arereceiving less nursing care than other patients orbecause you are on a restricted diet."

At that time I proposed that we try to look fora more equitable system for charging patientsbased on the amount of nursing care the patientreceived. A more equitable system could be ar-rived at with the use of an equation that includesthe following components:

(a) fixed amount for hotel costs;(b) cost of individual nursir care calculated on

tasks to be done for the patient, type of per-sonnel required to carry out the task, salaryper hour, and equipment used;other items shared by all patient charges suchas services of ward clerks.

There is much to be done in this area. Whetheror not automated data processing systems are thevehicle for accomplishing better patient care re-mains to be investigated using criteria establishedfor overall success.

(c)

172

Concept Modeling in Hospital Staffing

Dr. Lillian M. PierceProfessor, School of Nursing

The Ohio State University HospitalColumbus, Ohio

In a recent symposium, Morris described theprocess of model development in the followingwords: "The process of model development maybe usefully viewed as a process of enrichment orelaboration. One begins with very simple models,quite distinct from reality, and attempts to move in

evolutionary fashion toward more elaborate models

which more nearly reflect the complexity

of the . . , situation."This seems harmless enough, yet it is of some

importance to point it out explicitly. . . . The at-tempt to begin immediately with a rather richmodel may become a serious source of frustration.

Starting simply gets things moving and thustends to relieve some of the tension. It does, how-ever, require a certain amount of poise or 'guts'to back off from a complicated problem and hcgin

with a simple conceptual structure. It requires one

to deliberately omit and distort certain aspects of

the situation and to knowingly commit the sinsof suppressing' difficult considerations and subop-

timizing (1)."Models are abstract represent".tions pf a set of

empirical phenomena. The successful model notonly "must be isomorphic with its domain of ap-plication (2)" but also much correspond to a for-

mal system or validated theory for that set ofphenomena (3). Willer says of models, "A modelcontains a rationale and nominally defined con-

167

cepts which are structured in the form of a mech-

anism. A model is cumbersome and complex incomparison to the purity and parsimony of aform,' system. . . . There is a one-to-one cor-respondence between the relationships developedin the mechanism of the model and those of theformal system. . . .

"The terms and relations of the formal system

were derived from the concepts and mechanism

of the model, respectively, and neither the con-cepts nor the mechanism are part of that formal

system. The purpose of the model is to generaterelationships . . . . The formal system is limited

to statements of relationships in the simplest pos-sible form in order to make them easily testable

(4)."Three kinds of models are generally described

in the literature: analogue, iconic or typological,

and symbolic. Churchman says "an analogue modelemploys one set of properties to represent some

other set of properties which the system beingstudied possesses (3)." Willer says "analogue mod-

els are constructed by allowing some set of proper.

ties, structure and (or) process A to stand for

the properties, structure, and (or) process of thephenomena being studied, X This is most com-monly done when the A properties are betterknown and more familiar than X (4)." The majoradvantage of the analogue model is the known

168

reference or starting point. However, the inher-ent weakness of "borrowing" rather than construct-ing the rationale and the mechanism may renderthe model ineffiective, particularly if the analoguesare not isontorphic with the phenomena beingexplored.

Iconic or typological models are generally scalarmodels of direct similarity to the subject theyrepresent. The mechanism of the iconic modeldepends upon the number of similarities betweenthe model and reality. As the model increases inabstraction the similarities decrease, and the modelis rendered useless. Iconic models are strongest atlower levels of abstraction. If properly constructedthe iconic model is isomorphic with the phenom-enon. If the transformation from reality to modelis only scalar then isomorphism is assured.

When a set of empirical phenomena is repre-sented directly by a series of nominally defined con-cepts, the rationale for such representation is

iconic. However, unless the concepts that representthe phenomena are structured in a mechanism ofsome kind, the end result can be nothing morethan a conceptual scheme with little, if any,capacity for prediction. The effectiveness cr suc-cess of iconic models depends upon abstracting amechanism from the phenomena and then usingthe mechanism to order or connect the concepts.

The symbolic model is more formal in its con-struction than either the analogue or iconic model.Additionally, the symbolic model has a strong butnot rigid mechanism. As Willer says, "Symbolicmodels are constructed by the meaningful inter-connection of concepts. Models of this sort aresymbolic in that (1) their general rationale con-sists of allowing a set of connected concepts tosymbolize a set of phenomena, and (2) theirsymbols or concepts are the source of their mech-anism (1)."


The construction of a symbolic model begins .

with the nominal definition of concepts and thedevelopment of rationally consistent assumptions.It is not unusual for concepts to be theoreticallyas well as nominally defined. The structural andfunctional relationships of the concepts may bederived from their theoretical definition. Explicitassumptions, in addition to the implicit assump-tions of theoretical definitions, may be introduced.

The meaning of the concepts and the relationsbetween them provide the rationale for the sym-bolic model. If the construction of the model be-gins with the definition of concepts and the de-velopment of assumptions, the rationale will evolve.On the other hand, it is not unusual to begin withan explicit statement of the rationale that thenbecomes the basic assumption for the model. Aseach concept is operationally defined and itsmeaning made clear, it leads to further definitionand clarification of related concepts.

Although there are no formal rules for modeldevelopment, it is wise to keep in mind Morris'admonition to start simply and "make haste slowly"toward the more elaborate model that reflects thecomplexity of the situation. The primary requi-site in all instances is extensive knowledge of thephenomena for which the model is being con-structed.

There are, of course, certain phenomena suffi-ciently circumscribed that expert knowledge froma single discipline is adequate. However, the phe-nomena with which we are dealing is of suchcomplexity that, to date at least, attempts to con-struct models have been most successful when theknowledge from several different disciplines hasbeen brought to bear on the problem. The re-mainder of this paper discusses one particularmodel.

Development of a Patient Care Model

Introduction

The project discussed here is part of a longrange research program of the Systems ResearchGroup at Ohio State University. The primary ob-jective has been the development of methods forinvestigating complex man-machine systems. That

174

part of the program dealing with health systemshas been directed toward the development of ameasure of patient care and a model of healthcare systems relating the measure to health sys-tem behavior.

When one speaks of a measure of patient care,the listener generally tends to think of a scale, a

CONCEPT MODELING IN HOSPITAL. STAFFING169

checklist, or some other such instrument. Themeasure of patient care developed (luring thisproject is not a "tool kit." It is a descriptive modelat the tactical, or bedside, level. It is descriptiveof the interactions of the individual patient withvarious members of the health team and reflectspatient response to system .behavior.

The early phases of the work were characterizedby multidisciplinary organization. Representativesof the various disciplines investigated those aspectsof the problem that were traditional with them.For example, the sociologists looked at such things

as status relationships and work performance, thepsychologists examined personality variables andpatient progress, and the operations researchersstudied problems of inventory and resource alloca-

tion. Fruitful as these activities were, they still didnot attack the major problem: how well is thehospital doing its job? The degree to which the

research concepts and variables were commonto the various disciplines was limited.

Conceptual Framework for the Model

The later phases of this study have been inter-disciplinary. Unifying concepts were selected, andthe research has been conducted in a conceptualframework common to representatives of all thedisciplines involved. Cybernetic concepts providedthe framework for the development of the pa-tient-care model (5,6,7,8). These concepts havebeen criticized as being mechanistic; i.e., rulingout the richness of variable human behavior. Theyare, in fact, based on the behavior of living or-ganisms (7). Beginning with Bernard's disturb-ances and responses of the internal and externalmilieu (9) and with Cannon's ideas of homeo-

stasis (10), the concepts have been revised andextended by advances in information theory, com-munications, and computer technology (11).

The basic concepts of cybernetics are informa-tion and communication, regulation and con-trol. Von Bertalanfry describes cybernetics as

"theory of control systems based on communica-tion (transfer of information) between system andenvironment and within the system and control(feedback) of the systems function in regard toenvironment (/2)." Weiner, who coined the term,defines cybernetics as control and communicationin the animal and machine (6). Ashby calls cyber-

netics a theory of machines that treats not thingsbut ways of behaving. It does not ask, What is thisthing? but, What does it do? It .is thus essentiallyfunctional and behavioristic. (11). Cybernetic con-cepts are familiar to a number of disciplines suchas medicine, physiology, psychology, and sociology.

The cybernetic model is primarily concernedwith communication and information flow in com-plex systems. The model feedback, control, andregulation has applicability for biological and so-cial systems as well as the engineering problemsto which it was first applied. Every system is sub-ject to forces that make it necessary to adopt tochange in the environment, internal or external.The homeostasis, dynamic equilibrium, or steady

state model is essentially a process of balancingthe components of a system in response to change

in the environment. The system can adapt throughfeedback information to a wide variety of situa-tions (13).

The patient care model we have developed isa descriptive model at the tactical level. It is

descriptive of the interactions of the individualpatient with all members of the health team. Thepatient is viewed as an information source com-municating with members of the health teamthrough a variety of sensory input channels. In-formation about the state of the patient is ob-

tained by the nurse or physician or other healthteam member who observes patient behavior.

Comparisons are made between the actual stateof the patient and the desired state. Action is takento reduce any difference that may exist betweenthem. The rationale for the model is that thepatient care system is adaptive, relying ou ;,,for-mation feedback to cope with disturbances fromthe internal or external environment (ii), andthat the purpose of the system is the reestablish-

ment and raaintename of homeostasis (or dynamicequilibrium) for the patient (15,16,17).

The patient care model was developed by firstconceptualizing an integrating model based oncybernetic concepts (17) to serve as a theoreticalpoint of reference for the collection of data aboutthe phenomena that the model represents, pa-tient-health team interaction. These data wireused to refine and extend the model to its presentform (figure 1). The system or subsystem can be

described as a series of functional componentswith the following behaviors.

170

Memory

Control

XLt

YL_t

XLy


Figure 1.Patient care model.

Ys1CA(Ys-4A) (X ,,,X

1=11111w Regulate

Environment

V (11,1

wmass.01.

Monitor

176

InformationSource

CONCEPT MODELING IN HOSPITAL. STAFFING

The control function consists of specifying thelimits within which the patient is to he maintained(symbolized as YR). The limits are based on knowl-edge about the individual patient as well as medi-cal and nursing knowledge in general. Limits aresubject to change as the patient progresses

through the course of his hospitalization or as newinformation is obtained.

The .function of the memory unit is to storeknowledge and information. The content of thememory unit is constantly updated. The patientrepresents the signal source transmitting informa-tion about patient state (symbolized as VA) to themonitor where tF'c information is observed andrecorded or displayed. The function of the com-parator is to compare the actual state of the pa-tient as observed by the monitor to the desiredstate specified by the controller, and to note anydifferences between them (symbolized as Y8 VA).

The ...egulatory function consists of selectinganti carrying out the appropriate action [symbol-lized as (X,, Xr,)] necessary to reduce the dif-ference between the actual and desired patientstates. The patient, as a process, responds t' in-puts from both the environment and the regula-tor.

Collection of Data for the Model

To develop a model that accurately describesthe behavior of the system in terms of the pa-tient's response to the allocation of hospital re-sources, it was necessary to observe the interac-tion of the patient and various members of thehealth team in a variety of settings throughoutthe hospital. Since the system is adaptive, it is

not possible to predict a priori when interactionsor changes will occur.

To obtain the comprehensive data needed todevelop the model, patients were observed con-tinuously (24 hours a day) from admissionthrough discharge. We explored several methodsfor collecting the data. During the earlier workon this phase of the project, we collected data inthe operation room by visual observation and man-ual recording of events supplemented with taperecording. The time span for those observationswas relatively short, the number of variableswas relatively few, and the situation was con-trolled by the anesthesiologist and the surgeon.

171

We considered electronic recording equipmentbut found it to be impractical when the patientwas ambulatory. We also considered video record-ing. This method is excellent for collecting con-tinuous data in well controlled settings such asthe operating and recovery rooms or the intensivecare unit where cameras can be ceiling mountedand remotely controlled and where the time spanfor observations is relatively short-4 to 8 hours.But the high cost for equipment and technicalassistance and the low reliability of the equip-ment when used continuously over long periods,as well as the limited range of the camera's scopewhen the patient is ambulatory, made it impracti-cal for our purposes.

We finally decided that nonparticipant observa-tion with manual recording of data in the formof a narrative description of events would give uswhat we needed. We decided to use nurses tocollect the data, for several reasons. First, observa-tion is considered an essential part of the nursingprocess. Second, the hospital would be a familiarenvironment and nurses would be less concernedwith their surroundings than would observersfrom another discipline. Finally, nurses would beable to identify various personnel interacting withthe patient and the procedures being carried out.

We developed a manual of data collection de-si3ned to orient the observers to the project andindicate what was to be observed and how theobservations were to be recorded (18). The entirephilosophy of the data collection was that of the"now generation, tell it like it is. We used avariety of techniques to prepare the observers, in-cluding the use of videotape and practice sessionsin various hospital settings. We also used video-tape and dual observations to determine interob-server reliability.

The observers were formed into six memberteams, with each member of the team responsiblefor the same 4-hour block of time each day dur-ing the patient's hospital stay. Observations werebegun when the patient was admitted to thehospital and continued until discharge. Whereverthe patient went the observer also went and re-corded, as objectively as possible, a narrative de-scription of "who did what to, for, or with thepatient and how the patient responded." Tenpatients were observed. The length of stay rangedfrom 9 to 24 days. With observations on a 24-hour

basis, approximately 3,000 hours of time seriesdata describing system behavior were collected.Five complete sets of data are being used.

The data, recorded as narrative description, hadto he reduced to a format suitable tor «mpteranalysis. Again using the conceptual Iramework toprovide the guidelines, we developed a coilingformat that allowed numbers on punclicards tobe translated into sentences. The sentences canbe read as "at this time, in this location, this jw-son is doing this to, for or with the italic:al, andthis is the patient's response" or "at this time, inthis location, this is the patient's state and this isbeing done, by this person."

Analysis of the Data

The analyses of the data are described in thefinal report for this project. The description ofthe programs for analysis conies directly from doltreport (19).

With the capability of computers. the systemsresearcher is able to deal with the complexity ofMultivariate systems. However, it is necessary tospecify the final output and prepare the necessaryprograms to generate the specified output. So farwe have developed four basic programs for dataanalysis.

If the model is descriptive of system behavior,one of the first questions to he answered is, whodoes what for patients, when, with what resources,and how does the patient respond? A programwas developed to cross-tabulate the following:

(a) resourc_ and patient variables with time,(h) patient with resource variables,(c) resource with resource variables,

(d) patient with patient variables.The yield from this program has been a series

of frequency histograms that demonstrate patient-resource interactions such as the frequency withwhich various levels of personnel carry out variousactivities during each day of hospital stay. Scrutinyof these functions is expected to suggest hypothe-ses that can be tested, either in the real worldor with data generated by the model.


Because tabular data are difficult to interpret,a graphic time series plot was developed. Theplot is designed to demonstrate the concomitantvariation of patient and resource variables. Theoriginal plot of 21 variables has been expanded toallow up to 120 variables to be plotted simultane-ously. The plots can be examined to determinethe state of the system when situations that re-quire nursing intervention arise (20).

The third. basic problem is a search program.Having specified, or located, a critical patient vec-tor, such as pain, the data are searched prior toour following the occurrence of such a vector todetermine patterns of resource allocation that fol-low or precede it. Analysis of the data in thisfashion makes it possible to determine to whatextent resources are allocated on basis of pa-tient information and to will; extent otherconsiderations are more important (2!).

The three previous programs deal with the in-dividual patient. They describe the interaction ofthe patient with the system environment in time.To generalize from the individual patient, it isnecessary to sum the results across patients. Thisis a function of what Ashby calls the R x Dtable (11) that describes the consequences of asystem response to a disturbance in the environ-ment (see figure 1).

The R's represent resource allocation in re-sponse to the D's, disturbances such as pain. Forexample, if pain is a disturbance (D) and theadministration of Demerol a regulatory action(R). the resulting state of the patient would befound in the cell at the Rrow, Dcolumn inter-section. As more data are obtained, the cells canhe filled in. If there are statistical regtdatives theprobabilities of various patient responscs, to regu-latory-disturbance can be determined.

The patient care model has not been developedto that stage earlier identified as the ultimategoal of the model builderthe mathematicalmodel. It has, however, been developed to a stagethat allows its to examine the patient's responseto system behavior, to determine who is doingwhat for patients, when, with what resources andwith what responses.

CONCEPT MODELING IN HOSPITAL. STAFFING

References

(1) Mc..ris, William T. "On The Art of Model-ing," In Ralph Stogfill (ed.), The Processof Model Building in the. Behavioral Sci-ences, Columbus, Ohio.

(2) Black, Max. Models and Metaphors: Stud-

ies in Language and Philosophy. Ithaca,New York: Cornell University Press, 1962,p. 238.

(3) Churchman. C. West; Ackoff, Russell L.;

and Arnoff, E. Leonard. Introduction ToOperations Research. New York: JohnNViley and Sons, Inc., 1957, p. 157.

(4) Willer, David. Scientific Sociology. Engle-wood Cliffs, New Jersey: Prentice-Hall, Inc.,1967, pp. 18-19.

(5) Beer, S. Cybernetics and Management. New.York: John Wiley and Sons, Inc., 1960.

chapter 1.(6) Weiner, N. Cybernetics. New York: John

Wiley and Sons, Inc., 1948.Ashby, W. R. An Introduction to Cyber-netics. New York: John Wiley and Sons,Inc., 1956.Guilband,' W. T. What is Cybernetics? Lon-don: Heineman, 1959.Bernard, C. An .!ntroduction to the Studyof Experimental Medicine. New York:Dover, 1957.

(10) Cannon, W. B. The tf. isdom of the Body.

New York: W. N. Norton ik Company, 1932.(11) Ashby, W. R. Design for a Brain. New

'York: John Wiley and Sons, Inc., 1960.(12) Von Bertalanffy, L. General System Theory.

New York: George Braziller, Inc., 1968.

(7)

(8)

(9)

173

(13) Pierce, Lillian. "General Systems Theory:An Overview." Paper presented at 2ndNursing Theory Conference, University ofKansas, Department of Nursing Education,1969.

(14) YOvits, M. C., and Cameron, S. (ed.) SelfOrganizating Systems. New York; Perg-

amon Press, 1960.(15) Howland, D. "Approaches to the Systems

Problem." Nursing Research, 12: 172-176,Fall 1963.

(16) Howland, D. "A Hospital Systems Model."Nursing Research, 12: 232-236, Fall 1963.

(17) Howland, D., and McDowell, Wanda. "TheMe. surment of Patient Care: A Conceptual

Framework." Nursii Research, L3: 4-7,Winter 1964.

(18) Pierce, Lillian; ;.: r, Nancy; and Lara-bee, Jean. Data ction Manual. Adap-tive Systems Re ..arch Group, Ohio StateUniversity, Columbus, Ohio.

(19) Howland, D.; Pierce, Lillian; and Gardner,M. The Nurse-Monitor in a Patient-CareSystem. Report No. RF1651, Adaptive Sys-tems Research Group, Ohio State Univer-sity, Columbus, Ohio, (in preparation.)

(20) LaRue. R. "Mediplot." RF1651 Wp64,(on file) Adaptive Systems Research Group,Ohio State University, Columbus, Ohio,1969.

(21) Reid, R, "Computer Program for Back-ward/Forward Search of Data." RF1651 WP.62, (on file) Adaptive Systems ResearchGroup, Columbus, Ohio, \1969.

DISCUSSION OF DR. PIERCE'S PAPER

Discussion Leader

Dr, Wanda E. McDowellAssociate Dean For Research

School of NursingUniversity of MichiganAnn Arbor, Michigan

Dr. McDowell

Throughout this conference, concerns havebeen expressed regarding the lack of model de-velopment, the need for a theoretical frameworkto direct inquiry,, the disappointing progress inthe development of measures of care, and thefailure to make the utility of nursing explicit.Implicit in all these are further concerns regard-ing decision making, patient outcomes, and allo-cation of resources.

Dr. Pierce's paper represents an approach to allthese concerns. The key elements in the adaptivesystem model are communication and control, Asit is presently composed, it is descriptive, notpredictive. We do need to know the outcomesand consequences of various regulatory strategiesin the management of patient condition, and wedo need to be able to predict these outcomes.Otherwise the approach (the model) is not goingto be utilized by practitioners, by administrators,or others in the health care system. It should beuseful to researchers, also.

There has been further criticism that the modelhas been used only in hospital settings. But theattempt originallyand I think it still pertainshas been to move out of the hospital settinginto the community and the home, wherever thereis a consumer of care. In a series of three paperson nursing research and the development of thesystem model produced in 1964, there was con-cern for the various settings in which the con-sumer of care finds himself and the classes of

174

variables that must be attended to and studied.The model is especially helpful in selecting

critical variables for nurse action or for nurseintervention. I think we have to consider whatthe nurse attends to, what the physician attendsto, and certainly what the patient attends to be-cause he, himself, is a regulator of his condition.We have to complete this cycle and consider con-sequences for the patient. I chose the recoveryroom as a source of data for my doctoral dis-sertation. In that setting, there is a circumscribedperiod of time in which the patient is captive,so to speak, and the classes of variables that oneattends to are fairly limited. But if you think ofthe physiological and psychological variables ofconcern there, then as the patient returns to theunit following recovery (from anesthesia) othervariables come into focus. We are concerned withsocial variables and economic variables, for ex-ample.

These variables mange over time, too, asdo the goals of the subsystem. One thing thatshould be added to figAre 1 of Dr. Pierce's paperis that time is a critical variable that thus illus-trates the dynamic component of the model. An-other very crucial element is that the focus ison the patient. Of primary concern, too, is theinterdisciplinary effort required to tackle thisquestion, this problem. There has been criticismof the model and a lot of argument about it, and

140

DISCUSSION OF DR. PIERCE'S PAPER

someone made the comment earlier, Why is it

that each one has to develop his own model? Why

can't we use a model, redesign it, refine it, andbe critical in a constructive kind of way? Unless

175

we test available models in the real world wewill delay inquiry and interaction and any kindof professional, scholarly attempt to get a handle

on the concerns we have been discussing.

Summary, Synthesis, andFuture Direction

Dr. Mary K. MullaneProfessor, College of Nursing

University of IllinoisChicago, Illinois

Our conference started with a discussion ofstaffing studies. based on an engineering modelof nursing performance. In the studies reported,nursing activities were measured along productionindustry modes: standards, standard times, worksampling, etc. These' are very useful to the degreethat nursing is a sequence of tasks to 'be per-

formed within given limits of time.Criticism of limitations of the engineering

model was evident in the discussions followingthe formal presentations. Some suggested the en-gineering model is faulty in that the tasks ap-proach does not reflect nursing responsibility foremotional support, for patient instruction, forcoping with nonmedical and nonhospital patientconceths, and for family management.

Illustrations of nursing performance in the"psychosocial" facets of patient care and supportwere frequently cited in papers and discussion atthis conference. Nurses in hospitals have not yetexplicitly defined these facets of rare, and I thinkwe nurses are fully cognizant that such definitionmust be ours. It cannot be delegated. Definition,description, and then measurement cannot behoped for except through our own efforts.

One criticism of the engineering modality is

177

that it can measure only what is being doneovertly. It cannot measure what nurses have notdefined. It cannot Net ideals or standards of ex-cellence. Measurement means some type of defini-tion and quantification of what can be physicallydiscerned.

Another criticism of the engineering model de-rives from the construction of discrete, specific

time per activity from work and time studies; i.e.,a mean, a point estimator without indication oflevels of confidence. Without confidence intervalsbeing-made explicit, nurse staffing based solely onan engineering model will be suspect not only bynurses but by administrators and economists pre-cisely because the flexibility of confidence intervalsis not built into staffing estimates. If the notionof confidence interval in nurse staffing estimatesdues not apply, can engineers and statisticians makeus understand why it does not?

Patient classification schemes/were presented andctiticized during the confere4e, but time is per-haps more similarity in patient schemes than wearc prepared to admit. Questions of their validityand reliability have been raised frequently, yetseveral speakers reported coming to comparableconclusions. I refer you to the Canadian series

178

that was reported for us by Phyllis Giovannetti. Itseems evident that patient classification scheme asa base measure of nursing workload is generallyapplicable despite the concerns expressedin-ability or unwillingness or impropriety of general-izing among hospitals that often seem comparablefrom the outside.

Hospitals seem comparable from the usually reported data size, whether or not they are volun-tary, etc. But they vary widely internally. Theyvary as to style and capability of their generaladministration. They vary about medical therapeu-tic capabilities and ambitions, the adoption of fash-ions, whether they are cobalt bombs, organ im-plants, or transplants. They vary as to internalorganization. They vary as to allocation of responsibility and of costs to nursing. Each of thesevariances has a major effect upon nursing care andnursing administration. Dr. Gerrit Wolf's report ofhis peer review study pointed this out.

The fact that the engineering model alone ispresently incomplete does not suggest to me itshould be replaced with another model. Rather,progress reported in the literatute suggests its con-tinuing usefulness ;f engineers, nurses, hospitaladministrators, and trustees are sophisticated aboutwhat it can and what it cannot do and are pre-pared to apply it with full understanding of theproper concerns and reservations of each of theparties.

Human services have all of the problems, humanproblems, that production industry has. But theseare compounded by a product that in itself is

human and therefore must be described and meas-ured in human terms. Industrial design is mucheasier to deal with than nursing care design. Nurs-ing supply and performance and procuction arecharacterized both by human inputs and humanoutcomes, Describing and measur;ng them inhuman terms (and engineering modes can behumanized) is the only real way, however difficultthis may be.

John Griffith introduced the bargaining model.He did not suggest substituting the bargainingmodel for the engineering model; they can com-plement one another. A bargaining model wouldmake explicit the administrative relationships be-tween trustees and administrators concerned prop-erly with costs, and the professionals, nursing in-cluded, concerned with quality of professional


performance. The bargaining model makes con-structive advocates out of each, placing them inopposition to negotiate apportionment of hospitalresources to achieve both hospital efficiency andprofessional effectiveness; i.e., quality patient careat a tenable cost.

Gerrit Wolf's peer review report reinforces theneed to attend to administrative relationships andthe bargaining skill of all parties to it. Negotiationgoes on constantly between nursing administrationand other departments, and perhaps the processand content of such negotiations would shed lighton nursing load and its changes, nursing utilization,and so on.

The bargaining concept leads inescapably toadvocacy. Improving patient care through thismode requires that both administrators and profes-sionals accept negotiation as good, healthy, anduseful. Attitudes will need to be changed andsophistication in the modes and processes of bar-gaining must be developed. Successful bargainingrests upon negotiation from strength, mature un-derstanding of power, and the development andconstructive use of it.

One wonders what would be revealed in thestudy of the amount of nursing power, its sources,present and potential, and its tun, present andpotential. Can patient care in hospitals really beeither efficacious or efficient with nursing unableor unwilling to become skilled advocates for it?Application of the bargaining model raises manyuseful questions, power being only one of the moreobvious.

Dr. Flagle and Mrs. Giovannetti asked in dif-ferent ways whether it is not now time to approachnursing staffing and its host of related issues throughsystems analysis. Mrs. Giovannetti's paper containsa scheme for looking at the entire health caresystem, nursing included. Dr. Flagle suggests fun-damental questions about the objectives of thehealth care system. Specification of objectives ofthe health care system and the measurement oftheir achievement are difficult. Perhaps that ap-proach for some researchers could help ration-alize the nursing subsystem in hospitals.

The Pierce report adds both weight and portentto that approach. Often in the discussion, factorsin the climate in which nursing is practicedmentioned, and their effect upon nursing staffing,nursing utilization, and nursing effectiveness were

183

SUMMARY, SYNTHESIS, AND FUTURE DIRECTION 179

alluded to, but their identification and systematicin depth investigation lie ahead.

Dr. .14,:de Hone spoke of the effect of the varietyof medical styles upon nursing staffing. Experi-

CObSerVerS of the hospital scene know -onecannot completely study nursing or get definitiveanswers except within the framework of the givenhospital's medical style, medical capability, medi-cal ambition if you will, medical modes of practice.The ranges of therapeutic: capability present orgrasped for has a critical effect upon nursing staff-ing. Medical milieu simply must be studied to afforda complete understanding of the problem. Thesame is true for the hospital's style of generaladministration. Research into these relationships isurgently needed.

The most frequent question raised in discussionperiods at this conference has been that of qualityof nursing care and its measurement. Though beginning attempts at definition and measurementare under way, the major work lies ahead and itspriority after this conference would certainly seemto be very high. Perhaps ways can be found todocument and synthesize what work has been cloneand subsequently to convene persons knowledge-able in investigation of the question and in ap-plication of modes of documenting quality innursing care given.

My assignment for this summary included thecharge to make explicit some of the "straws inthe wind" of the future. Clearly, greater atten-tion to the economics of health care is mademandatory by the growing resistance to continuingrise in costs, John Griffith referred to the inclusionof health care costs under the Federal Price Com-mission; references to the cost benefit ratio areincreasingly common in literature and professionalconversation.

The emphasis of this conference on the urgencyof identification and measurement of quality inhealth care delivered is uncles .cored by develop-ments like the Commission on the Quality ofHealth Care as proposed in the Kennedy bill, andthe refinement of standards and criteria of theJoint Commission on Accreditation of Hospitals.Professionals who are making health care the fieldof their career endeavorsphysicians, nurses, ad-ministrators, engineers, social scientistsmust cometo grips with an operational, measurable definitionof quality of health care, specifically nursing care.

Our habit is to describe the ideal. Although theideal is certainly important, experience suggestsone can work toward it only by delineating andpursuing first what is feasible and rational. Oneof our speakers introduced the notion of -moderateutility," the concept I am advancing here. Is itpossible to identify critical indices of qualityrather than concluding that we have no quality atall unless we measure everything imaginable?

Another point of future direction lies in therefinement of patient classification schemes. ()nes-tions of validity and reliability remain, but thegeneral applicability of patient classificationschemes suggest their usefulness. As Mrs. Giovan-netti pointed out, the question is not whether theyare good or had; the question is how do we refinethem so as to make them more widely useful.What other measures of patient load are there?

Surely the future will bring greater study oforganizational and operational management ofnursing services. Systems analysis and design holdgreat promise, in my view. So too does the conceptof management by objectives. Horizontal and veni-cal relationships and controls exercised by nursingand upon nursing need to be studied. Long over-due are studies of the effect of hospital adminis-tration and of the effect of medical practice uponthe' organization, objectives, and character of nurs-ing load and nursing care outcomes.

We gave patient satisfaction no attention in this.conference. It may be that the, fashion of studyingpatient satisfaction has passed. It may be that weknow all we need to know about ii. I am notfamiliar with what we know about it, but I wasstruck by the absence of discussion of patient satis-faction. Indeed, I drew an inference from discus-sion that patients might be thought incapableof judging the quality of care. 1 do warn againstthis inference, because we are moving more andmore to the notion of satisfied customers in a

population becoming more sophisticated. and morevocal about health care.

Study is needed of the porposes and processesof surveillance of nursing performance: of super-vision, the supervisory process in nursing: of con-trols upon or within nursing; of the skill andsophistication of nursing administration and theconsequences of their presence or absence.

Almost 20 years ago I studied nursing adminis-tration in hospitals in Nlichigan. I was startled to

.04

180

learn then that the nursing department budgetexceeded the budget of almost 40 percent of theindustries of the State of Michigan listed in theManufacturers' Directory. Few recognize themagnitude of the operation nursing administratorsare responsible for or the educational and personalqualifications crucial to adequate management ofpresent day nursing services. Directors of nursingare often appointed because of tenure in otherposts in the hospital or acceptability to the medicalstaff, future will be stormy indeed whereversuch practice continues.

I recognize that our conference was directed to-ward staffing in hospitals, and hospitals will con-tinue to hold a crucial place in the health caredelivery system. Rut the future, in my view, willsee greater development outside of hospitals thanin them. Will the modes for rationalizing, measur-

4


ing, and justifying nursing care and service re-quirements we are developing now be applicableoutside of hospitals? What has been or will belearned in hospitals that, if appropriately applied,will prevent public health nurses from "reinvent-ing the wheel?"

If the already discernible movement away fromindividual fee for service medical practice towardHealth Maintenance Organizations and othergroup or corporate modes or practice actually pro-ceeds, then I fail to see how medical services canescape analysis and insistence upon quality assess-ment and staffing justification, as nursing pres-ently is experiencing. What a public-gain it wouldbe if the modes developed in response to appro-priate pressures on the nursing profeision werefound to hue fransfors.Wlity zo.medicinel

Appendix ,

BEST COPY AVAILABLE

Task Force Reports

TASK FORCE TOPIC: Future Directions for Research on Nurse Staffing in Hospitals

In this context, the ask Forces are asked toA. Identi(v areas of needed research:

1. identify factors that may facilitate researchin these areas,

2. identify specific obstacles hindering re-search in these areas..

B. Determine needed resources:1. skills, and how they may be developed,

Task Force No. 1Miss Elinor D. Stanford, LeaderMiss Deane B. Taylor, ReporterDr. Virginia S. ClelandDr. Charles D. Flag leDr. Richard C. JelinekDr. E. tartly JacoDr. Stanley JacobsMiss Angie KammeraadMr. Stanley E. SiegelMiss Marjorie Simpson

Task Force No. 2Dr. Maria Phaneuf. LeaderMiss Marylou NkAthie, ReporterDr. Alexander BarrMr. Richard G. GardinerMrs. Phyllis GiovannettiMr. John GriffithMiss Dolores M. LeHotyMiss Ruby M. MartinDr. Donald SiniborgMiss Bernice SzukallaDr. Ger-it Wolf

2. funding, and appropriate sources,3. other.

C. Determine priorities for research.D. Explore means of coordinating work of re-

searchers and utilizers: ways to improve com-munications.

E. Recommend nature of further conference on ,nurse staffing research.

The Task Forces

Task Force No. 3Dr. NOTH poulin, LeaderPr. Ber044 AO Pt.11!ss Marie R. ennedy

Pr. 1-,4140Dr. Reginald It evansMiss Jessie M. Scott

Mr. PlarcP, Jr.Dr. Fran A. SloanDr, 1141-Al A. WandeltDr. Harvey Wolfe

Task Force No. 4Mrs. Margaret Sheehan, LeaderMiss Pamela Poole, ReporterDr. Mario BognannoMrs. Margaret EllsworthMiss Verna GrovertDr. Doris Bloch'Dr. Elizabeth HagenDr. Richard HowlandDr. Eugene LevineDr. Wanda E. McDowellMiss Marlene R, Slawson

181

186


REPORT OF TASK FORCE NO, 1

Leader

Miss Elinor D. StanfordDirector, Nursing 1 esearch Field Center

Division Jf NursingNational lmitutes of Health

San Francisco, California

Recommendaticy is

A. Future Research Studies1. A study to establish a theoretical base of

knowledgrof the .4r5nrArag relationship."What do we mean by "care" and "caring?"Based on the above applied research to thendetermine,a. Where should care be given or provided?b.. Who should provide various levels of

care? (responsibility and accountability)c. Careful delineation between the:tarious

caring roles in the delivery of health caresystems.

2. Various disciplines need to join force,. indesigning studies including economists, in-dustrial psychologists, sociologists, statisti-cians, industrial engineers, and well quali-fied nurse researchers.

3. Studies to look at various. staffing mixpatterns. The nurse clinician(s) role andwill/anion should definitely be in these mixpatterns. Their contributions to the qualityof patient care need to he identified and thecost analyzed within inpatient and ambula-tory settings.

4. Studies of nursing practice in other coun-tries, more cross cultural studies. What canwe learn from other health care systems?

5. What are the factors that make for delega-tion or assignment of tasks in multipatientsettings: i.e., the question of specializationand division of labor vs. versatility?

(Personal suggest ion from Nliss Taylorskilledhilt .e researchers should form an independentconsultant firm, like private practice, and organize a

Reporter

Miss Deane B. TaylorAssociate Professor

State University of New YorkBuffalo, New York

team of interdisciplinary consultants to really go inand work with a group in a hospital or university toidentify tltp problem, prepAre The proposal, themethodology, etc. Perhaps the nurse would needsome financial support to help her get started for ayear or so until services become known to nurseswho would like to get into research but have nothad doctoral preparation.)

b. Obstacles to research:(1) fragmented roles of auxiliary staff

being locked in by union contracts,(2) ethnic demand for career opportuni-

ties in the health service field,rigidity of present professional andadministrative people to studies oftheir functions and roles,

(4) vested interest groups who are threat-ened by possible findings of studiesfear of loss of jobs, incompetency todo job, job descriptions not clear souncertainty as to role and functions,inadequate preparation of those whowill be involved in the study in anyway results in false data and thwart-ing by personnel.

Skills Needed and How These May Be De-veloped1. Doctoral preparation of adequate numbers

of rttrse researchers. Most now arc in ad-ministration and teaching wit?, no time to

'become really involved in research or evento assist nurses in other than graduateprograms (master level), which really is notenough preparation.

B.

1S';

(3)

(5)

APPENDIX: TASK FORCE REPORTS

2. Nurses who are interested in research shouldwork with an ongoing research study per-haps as apprentices. This would have toinclude writing a propos.ii, etc.

3. Interpersonal skills, interviewing and datacollecting experience, preparing data foranalysis and writing reports.

4. Funding to do needed research was notdiscussed by the group.

C. & D. These topics were not discussed by thegroup.

183

E. A Working Conference to Develop a MasterPlan on Strategy for Research in Nursing1. Participants should do "homework" before

getting together.2. The conference should follow soon after

this (present) conference3. Papers should be prepared and sent to

participants early, and authors should at-tend but only to discuss salient points raisedby participants that would be pertinent tothe development of the master plan.


REPORT OF TASK FORCE NO. 2

Leader


Dr. Maria PhaneufProfessor of NursingCollege of Nursing


Recommendations

A. & B. These topics were not discussed by thegroup.

C. Determine Priorities for Research1. Research should concentrate on clinical

practice with emphasis on the need todevelop quality control.a. Develop a study of patient oriented

structure cutting across professionalstratifications,

b. Examine the care process related topatient r,luirements as opposed to tra-ditional examination of the care processas related to professional categorization.

2. This process can be implemented by es-tablishing research teams throughout thecountry. The research teams should be

Reporter

Miss Marylou McAthieRegional Nursing ConsultantNursing Research Field Center

Division of NursingNational Institutes of Health

San Francisco, California

multidisciplinary and should include, as aminimum, ,indostrial engineers, psycholo-gists and nurses.

I). Explore Means of Coordinating Work of Re-searchers and Utilizers1. Any approach to researchers and utilizers

should be multidisciplinary.2. The providers of care must be involved.3. Organizations such as hospitals, associations,

insurance carriers, should be included.4. Consultants and resource people must be

trained.

5. An educational process should he estab-lished that would prepare potential im-plementers to use results of research,

1N9

APPENDIX: TASK FORCE REPORTS 185

Leader

REPORT OF TASK FORCE NO. 3BEST COPY AVAILABLE

Reporter

Dr. Bernard FerberDirector, Bureau of Research Services

American Hospital AssociationChicago, Illinois

Dr. Muriel PoulinBoston University School of Nursing

Boston, Massachusetts

Recommendations

A. Identify Areas of Needed Research1. General considerations (guidelines)

a. Nurses must assume the initiative forresearch with ,the -:mistance of otherdisciplines.

h. Decisions to initiate and implementresearch are made by the power struc-ture.

c. Costs must be considered.d. There is a gap between nursing staff

(providers of nursing service) and upperechelons. How do nurses break into thecoalition of power?

e. All relevant staff must be involvedin research endeavors. Involvement isneeded to obtain the necessary coopera-tion and to satisfy various needs.

f. What do we want to accomplish byresearch activiti?Nursing is presently functioning in anacceptable manner.

h. How do we learn new thin is?i. In studies to improve mining care, can

we offer improvements?j. Are studies getting to the right people?k. Do we have to be purists?1. teally good work (research) is not

published. Applied research is not pub-lished. How is this rectified?

rn. All inputs (disciplines) must he con-sidered, as all affect output.

n. Who will do research? Where do we getnurses to do research? Individuals on theunit must be involved.

o. What are the capabilities of hospitalstaff to do research. Larger hospitals

g.

190

have greater capability (groupings ofsmaller hospitals) .

p. Team approach maximizes implementa-tion possibilities.

2. Areas of needed -researcha. Recruitment and retention of nurses

require satisfaction on the job.b. Why do staffing levels vary, both within

a hospital and between hospitals?c. How can nursing be improved?d. Identify criterion measures of outcomes

of health care. Then identify input ofnurses to achieve desired outcomes atvarious points of time.

e Measure and test nursing inputs. Whatis it that is going to make a difference tothe patient?

f. Study staff substitutability. (How canstaff be substituted?)%hat are the boundaries (responsibili-ties) of providers (staff) of care? Whatis the interaction between staff providingcare?

h. Study organizational structure in whichcare is given. How do people act withinthe system?

i. / Study system/organizational constraintsthat preclude giving patient the treat-ment required.Study and clarify the expanded role ofthe nurse; make a clearer definition ofnurse input.

S. Factors that may facilitate researcha. T. am approach: interdisciplinary with

strong nurse input.

g.

186REST NWY AVAILABLE

b. Utilization of individuals in setting to hestudied. Individuals involved in collect-ing data should be aware of aims, pur-poses, and objectives of the research.

c. Increase in number of faunally and welltrained nurses who can do research.

d. The creation of an awareness of theneed for research.

e. The value system of the nurse educator(related to item 4) .

4. Obstacles that hinder researcha. Insufficient funding, limited sources of

funding.b. The hospital administration (adminis-

trator) does not fully appreciate theresearch nurse. Education is indicated.

c. Nurse attitudes: the immediate demandsof an operational system have first (top)priority, pressures of an operating sys-tem.

d. The self-recording mechanism.e. Value system of nurse educator.1. Separation of nurse educator, adminis-

trator, practitioner.Less than full commitment of peoplewho are involved in the research activity.

R. Needed Resources1. Skills

a. Develop skills of nurses to be researchcoordinators, liaison persons, alsocharged with interpretation of the re-search effoi t.

b. Develop skills in assessment and processof nurses at practicing level.

c. Nurses must know organizational struc-ture of hospital: how hospital works;what prevents action; what solicits re-sponses.

d. Continuing program to develop newresearchers.

2. Fundinga. Sources should be developed in addition

g.


to Federal grants. There must be greaterefforts at joint funding. joint fundingcould be between Federal agencies, Fed-eral and private sources, etc.

b. Research activity should be included inthe hospital budget. Hospitals shouldinclude research in their budget as wellas look for supplement funding.

c. The Division of Nursing should makeresearch in staffing one of its program

d. Health system should put dollars asidefor research.

e. There should be Federal funding tosupport new researchers (recent grad-uates) . This could be accomplished byfacilitating the number of grants to such

f. Some help should he available to aid newresearchers (those with limited researchexperience) to obtain research grants.

C. & D. These topics were not discussed by thisgroup.

E. RecommendationsFuture Conferences1. More nursing administrators should attend

fut,tre conferences.2. Nat.-ow the focus of future conferences

(fewer topics).3. Future conferences should present more

studies with nursing input and data, moreemphasis on studies done by nurses.

4. Should the current meeting have only in-volved nurses? There was a feeling thatnurses must coordinate and articulate theirown ideas before a meeting involving multi-disciplines. Nurses must decide their ownfunction (role) before having a meetingsuch as this conference.

5. Staffing is not the p,oblem. Effectke utiliza-tion of people should be the focus as itrelates to outcomes (identify nursing out-comes) .

APPENDIX: TASK FORCE REPORTS

Leader

BEST COPY AVAILABLE 187

REPORT OF TASK FORCE NO. 4

Mrs. Margaret SheehanNurse Consultant

Division of ,NursingNational Institutes of Health

Bethesda, Maryland

Reporter

Miss Pamela PooleNursing Consultant

Department of National Health and WelfareOttawa, Ontario, Canada

Recommendations

A. & B. These topics were not discussed by thisgroup.

C. Priorities for Research1. Directly related to nurse staffing in hospitals

a. Further pursuit of Miller-Bryant studieson the mix of personnel with particularemphasis on cost-benefit analysis.

b. Can we demonstrate the unique contri-bution of the RN; e.g., the psychosocialcomponents of nursing as opposed to thetask analysis approach?

c. Studies which foCus on the patient; i.e.,what is he generating that we have toattend to? What kinds of decisions needto be made? What kinds of appraisalsneed to be made?

d. How can we improve staff productivity?e. How much is spent, regarding effort and

cost and time, on the control of qualityof care?

f. What are the costs of on-the-job trainingwhere turnover is high?

g,... Development and study of models ofhow th' .43 could be.

ii. The effect of technology on nurse staff-ing.

i. More multidisciplinary approaches tostaffing studies.

2. Indirectly related to nurse staffinga. Studies of director of nursing; i.e., inno-

vative, styles of leadership, etc.

b. Analysis of conflict relationships be-tween the director of nursing and ad-mieditration and trustees on the basis ofissues; i.e., what issues produce conflict?

c. International comparative studies ofnurse staffing in hospitals.

D. Coordination of Work Researchers and Uti-lizers

1. Communicationa. Clearing house for studies; e.g., Division

of Nursing, AN F.b. Encourage reporting in professional

journals.c. Establish a list of people interested in

nurse staffing, such as this group, andkeep in touch on an annual basis withwhat they are doing. Circulate this list ofcurrent activities.

d. Regional meetings of researchers andutilizers on the. problem of nurse staffing.

E. Nature of Further Conferences1. Narrow the focus.2. Provide lead time to prepare oneself for

more free discussion and reflection on prob-lems to be discussed. Distribute papers inadvance.

3. Maintain the multidisciplinary approach. .

4. Maintain the Task Force format.5. Small interest groups might be organized in

the evenings on a voluntary attendancebasis.

*U.$. GOVERNMENT PRINTING OFFICE: 1913 0-4119-134