Quality in Health Care 1992;1:236-240

Factors influencing default at a hospitalcolposcopy clinic

Gillian Sanders, Carole Craddock, Ian Wagstaff

Newcastle HealthAuthority, Newcastleupon Tyne NE2 lEFGillian Sanders,consultant/senior lecturerin public health medicineCarole Craddock, socialscience researcherNewcastle HealthGeneral Hospital,Newcastle upon TyneNE4 6BEIan WagstaffCorrespondence to:Dr G Sanders, NewcastleHealth Authority, 2-10Archbold Terrace, Newcastleupon Tyne NE2 1EFAccepted for publication9 October 1992

AbstractObjective To identify factors

reducing compliance at diagnosis,treatment, and review stages amongwomen referred with abnormal cervicalsmears to a hospital colposcopy clinic.Design - Retrospective analysis of

sociodemographic data from hospitalnotes of the attendees and defaultersduring one year (1989-90) and prospectivecollection of information by structuredinterviews of a sample of defaulters andattendees during five months (May-September 1990).Setting - One hospital colposcopy

clinic.Patients - 238 women defaulting on

two consecutive occasions and 188attending regularly (retrospectiveanalysis) and a subset of 40 defaulters and24 attendees (interview sample).Main measures - Sociodemographic

data and interview responses aboutattitudes, behaviour, choice, accessibilitycultural understanding, communications,and emotional response.Results - 22(12%) women defaulted at

diagnosis, 24(13%) at treatment, 39(21%)at the first check up after treatment, and84(45%) at the review stage; 19(10%)defaulted from the first check up afterdiagnostic examination revealed no needfor treatment. Age and social classdiffered between the two groups. 181(76%)defaulters were under 30 compared with91(48%) attendees; 14(6%) compared with41(23%) were over 40(p < 0.001). Theproportion of women in social classes 4and 5 was 33%(20/60) for defaulters and21%(25/120) for attendees (p < 0.05) andunemployed was 66%(158/238) and36%(68/188) respectively. 63(28%)defaulters were pregnant compared with11(6%) attendees (p < 0.001). Moredefaulters came from gynaecology orantenatal clinics. Most defaulters (93%)had child care responsibilities and theyknew and understood less aboutcolposcopy. Their explicit reasons fordefaulting included child carecommitments and fear and their implicitreasons lack of understanding,inaccessibility of information, and staffattitudes.

Conclusions - Compliance may beimproved by promoting women'sunderstanding of treatment andencouraging health professionals to

develop a service more sensitive to thevarious needs of women in differentsocioeconomic groups.(Quality in Health Care 1992;1:236-240)

IntroductionThe effectiveness of follow up and treatmentof women with abnormal smear test results isessential if the cervical screening programme isto succeed. Clinicians providing the localdistrict colposcopy service had expressedconcern about the amount of defaulting fromclinics: 3 1O% of 3067 colposcopy appointmentsfor diagnosis, treatment, or review in a 12month period had not been kept.

Published work has concentrated on theorganisation of cervical screening rather thanefficiency and effectiveness of follow up ofwomen with abnormal smear test results.'Posner and Vessey2 and Quilliam3 recentlyindicated that emotional responses such asfear, embarrassment, stigmatisation, anger,and guilt play a greater part in women'sresponse to colposcopy than previouslysuggested. Marteau et al reported very highanxiety levels in women referred forcolposcopy.4 Analysis of attendance at thisdistrict colposcopy clinic by age group hadindicated that defaulting was more likely inyounger age groups, but little else was knownabout the pattern of defaulting or its reasons.This study aimed at identifying the factors

influencing non-attendance of new andreturning patients at this colposcopy clinic sothat necessary service changes could beimplemented to reduce defaulting. It wasapproved by the local ethical committee.

MethodsThe women in the study had been referred tothe colposcopy service, which provided mostdiagnostic and treatment facilities for the localdistrict health authority population of 270 000as well as for referrals from adjacent healthauthorities. The dedicated clinic suite is basedin the gynaecology department of NewcastleGeneral Hospital and provides three afternoonoutpatient sessions weekly. At any one clinicthere are three nurses, a consultant, a registrar,a senior house officer, and a clinical medicalofficer.The research design used both qualitative

and quantitive methods of investigation andincluded a retrospective analysis of socio-demographic information from the case notesofwomen who defaulted and regular attendeesduring 1989-90 and a prospective studyentailing interviews with a small sample of

236

on March 25, 2022 by guest. P

rotected by copyright.http://qualitysafety.bm

j.com/

Qual H

ealth Care: first published as 10.1136/qshc.1.4.236 on 1 D

ecember 1992. D

ownloaded from

Factors influencing default at a colposcopy clinic

women who defaulted at different stages ofdiagnosis, treatment, and review and with asample of women who attended regularlythrough to discharge.The clinical policy was for colposcopic

examination with a cervical smear and biopsyat the first visit; a visit for treatment shortlyafterwards; and visits for checks at four, 10,16, and 26 months. When no treatment wasrequired women were reviewed once at sixmonths.

RETROSPECTIVE STUDYSome defaulting may occur because anappointment coincides with a woman'smenstrual period. Our intention was toidentify other reasons for non-attendance.Hence defaulting from two consecutiveappointments (without an explanation fromthe women) was used to define non-attendance. These women were identifiedfrom a default register for 1989 held by thedirector of public health. A control group ofregular attendees was identified throughdischarge letters sent during 1989 by theconsultant gynaecologist to generalpractitioners. A sample representing twothirds of the default sample population wasrandomly allocated from this source.

Sociodemographic data (including age,marital status, employment status, educationafter age 16 years, occupation of women andspouse or partner, and postcode of residence),

Guide to content of interview

AgeMarital status

Children and their ages

Further educationCurrent employmentHusband or partner'semploymentMain source of incomeInvolvement in regularcervical screening

Previous colposcopic,gynaecological, and obstetricexaminations and assessmentof experienceEvents around examinationleading to abnormal smeartest result; women's feelingsin response

Contact of women with GP,primary health care teambefore referral for colposcopySources of information aboutcolposcopy(people/literature)Details about referral tocolposcopy clinic (waitingtime for appointment,whether appointmentreceived, attendance,accompanied withrelative/friend or

unaccompanied)

Experience of process withincolposcopy clinic (waiting time,explanation by staff, privacy,atmosphere in clinic)Experience of colposcopicexamination and biopsy (doctor'sexplanation of procedure, use oftelevision monitor, other staffpresent, duration, women's feelings,doctor's explanation of clinicalproblems, postal information onresult)Follow up visits (same/differentdoctors, receptionist's role,importance to women,understanding of reasons for followup)

Experience of treatment procedures(laser, cone biopsy, hysterectomy)Experience of pain (in relation totype of treatment, use of analgesia,use of local anaesthetic, symptoms,and during and after procedures)Process of follow up (women'sknowledge, attendance, reasons fornon-attendance, alternative followup for example (GP))General questions about knowledgeof cervical disease, risk factors forcervical cancer, reasons fortreatment

Why other women decide not toattend for colposcopy afterabnormal smear result

source of referral, and pregnancy status wereextracted from the notes for the default andcontrol group and analysed with the statisticalpackage for the social sciences (SPSS). Inaddition the women's addresses were checkedagainst the current address held by the familyhealth services authority (FHSA).

PROSPECTIVE STUDY

A prospective study was carried out tocompare the attitudes and views of defaultersand attendees.During five months starting in May 1990

the names and addresses of women whodefaulted from two consecutive appointmentswere identifed by the receptionist at eachclinic. Women defaulted at four stages:diagnosis, treatment, first check up aftertreatment, and review.A sampling technique stratified default

using the proportions from the four categoriesof the retrospective study. However, allwomen who were identified as havingdefaulted at the diagnosis, treatment, and firstcheck up stages were included, firstly, becauseof the few women in these categories whichrequired all to be interviewed to obtain anadequate sample and, secondly, because theywere seen to be at greater risk of disease. Afurther sample of women who attended theirclinic appointments regularly through todischarge were selected as controls from thosecurrently being discharged from the clinic totheir general practitioners. Women who haddefaulted on one occasion several times duringcare were excluded.Women selected for interview were sent a

letter requesting a home interview by CC.Interviews were structured and based on aquestionnaire which contained both closedand open questions (box). They were taperecorded and their content analysed,according to the following major themes:attitudes, behaviour, communication,accessibility, emotional responses, sensitivity,choice, and cultural understanding. Womenwere asked their reasons for not attending theclinic. Their initial responses were consideredthe explicit reasons for default; other reasonsexpressed later during the interview wereconsidered implicit reasons.Some of the sociodemographic variables

(age, source of income, marital status,employment status, and education after age 16years) and number and age of children wereexamined by x2 distribution. Significance wastaken at the 5% level. Information onpostcode was used to identify each women'slocal authority ward of residence, whichallowed a different estimate of socioeconomicstatus, by using the ward deprivation index.

ResultsRETROSPECTIVE STUDYA total of 251 women were identified in thedefault register; 13 files could not be foundand the study was therefore based oninformation about the remaining 238 womenwho defaulted. Of 236 attendees identified, 21were excluded because they had defaulted

237

on March 25, 2022 by guest. P

rotected by copyright.http://qualitysafety.bm

j.com/

Qual H

ealth Care: first published as 10.1136/qshc.1.4.236 on 1 D

ecember 1992. D

ownloaded from

Sanders, Craddock, Wagstaff

Table I Number (percentage) of women defaulting fromcolposcopy appointments and number selected for interview

Stage of RetrospectiVe Prospective Nodefault studv study ilnterviez-ed

Diagnosis 22(12) 8(8)Treatment 24(13) 5(5) 2First check up* 58(31) 34(32) 21Review 84(45) 59(56) 15

Total 188 106 40

*Includes 19 women (10O",,) who attended for diagnosis butdid not require treatment.

from one appointment and the files of afurther 27 could not be traced. This gave afinal sample of 188 women.Most defaulting occurred at the review stage

or during follow up; however 12% occurred atdiagnosis, 13% at treatment, and 21% at thefirst check up four months after treatment(table 1). A further 1 0% of defaulters occurredat first check up after diagnostic examinationhad revealed that no treatment was required(false positives to cervical screeningprogramme).The addresses of defaulters at the diagnosis

and treatment stages for women who livedwithin the main district health authority forreferrals were checked against the FHSAregister, which was known to be reasonablyaccurate. Within six to 12 months 380 o(22/56)of women were found to have changed theiraddress.Table 2 shows the sociodemographic details

of defaulters and attendees. Defaulters were

Table 2 Sumniarv sociodeniographic data and pregnancy status of defaulters andattendees

Defaulters Attezdcs Signlhficaulccof

No~(0)(/°/°) xT()(O)No diff e-en

Total 238 188Age (years) <20 24(10) 4(2)20-29 157(66) 87(46)30-39 43(18) 56(30) p < 0 00140-49 9(4) 28(15)350 5(2) 13(7)

Social class (own occupation):1 and 2 13(5) 33(18)3 (Non-manual 27(11) 62(33) p < 0.05and manual)4 and 5 20(8) 25(13)

Unemployed or 158(66) 68(36)housewifeNot recorded 20(8)

Marital status:

Single 125(53) 86(46)Married 71(30) 67(36)Separated or divorced 32(13) 30(16)Widowed 2(<1) 5(3) p

Not recorded 8(3)Pregnancy status:Non-pregnant 162(72) 177(94)Pregnant 63(28) 11(6) p < 0*001Not recorded 13(6)

Socioeconomic status (ward of residence):°/o Resident in two 69 42 p < 0 01most deprivedquartiles of city's wards

Table 3 Source of referral for defaulters and attendees. Figures are number(percentages) of women

Total General Family, Antenatal Gvnaecology GeitounnarsNo practice planning clinic well woman medicine

clinic clinic clinic

Defaulters 223 88(39) 15(7) 34(15) 30(13) 56(25)Attenders 186 91(49) 34(18) 7(4) 11(6) 43(23)

younger than attendees (181, 76%, defaultersv 91, 48% attendees aged under 30) and weremore likely to be unemployed or behousewives: two thirds of defaulters did notwork outside the home whereas two thirds ofregular attendees were in paid employment.These results are based on the women's ownoccupation. An analysis based on theirpartners' occupation showed similar resultsbut at a higher level of significance(p < 0 001). Marital status also showed adifferent distribution with slightly more singlewomen among the defaulters than attendees(53% v 46%), though more attendees wereseparated or divorced (16% v 13%). The agedifference of the two groups may explain thesevariations. Pregnancy status was recorded forall but 13 defaulters. A minority of bothgroups were pregnant at the time of referral forcolposcopy but the rate was higher amongdefaulters than attendees (63/225, 28%0v11/188, 60/ ). Analysis of postcode informationfound that a greater proportion of defaulterslived in the two most deprived groups of localauthority wards than did regular attendees.For both groups of women the main source

of referral was general practice followed by thegenitourinary clinic, but a larger percentage ofdefaulters than attendees were referred by theantenatal (I 5% v 4°0 ) and gynaecology or wellwoman (13%0' 6%) clinics (table 3).

Logistic regression analysis of thesevariables suggested that the maindiscriminating variables for attendance anddefault were age, pregnancy, and the women'ssocial class.

PROSPECTI1IVE' STUDl)YThe results of this study were based oninterviews from 40 defaulters, whosedistribution by stage of default was broadlysimilar to that of the larger retrospective group(table 1), and 24 regular attendees. Tracingwomen who had defaulted for interview wasdifficult: out of 88 women traced, 23(250 o)had moved, 17(20%) failed to respond, andeight (9%))o) refused to participate. In contrast,change of address was uncommon in theregular attendees occurring in only 10%(2/24).The characteristics of the defaulters were

similar to those in the retrospective study. Inaddition, information was available aboutchildren and their age. In all, 37(93()Vo)defaulters had children compared with 67% ofattenders; 35(88%) defaulters had childrenaged under 16 years (including 12(30%) withchildren aged under 5 years) whereas less thanhalf the attendees had children aged under 16years and only 17% had children of pre-schoolage. Fifteen (38%) defaulters and three (13%)attendees were single parents living alone withchildren. Seven (29%) regular attendees hadreceived further educational training after 16years of age compared with only one (30/,)defaulter.Table 4 shows the explicit reasons for

defaulting. Problems with child care were

considered to be made worse by the lack ofchild care facilities at the clinic and the

238

on March 25, 2022 by guest. P

rotected by copyright.http://qualitysafety.bm

j.com/

Qual H

ealth Care: first published as 10.1136/qshc.1.4.236 on 1 D

ecember 1992. D

ownloaded from

Factors influencing default at a colposcopy clinic

Table 4 Explicit reasons given for defaulting by 40women at interview (categories not mutually exclusive)

No(%)Fear 14(35)Child care 14(35)Clinic waiting time 4(10)Transport costs 3(8)Forgetfulness 3(8)Attends general practitioner 3(8)

(not arranged by clinic)Pregnancy 5(13)

afternoon clinics. Five women gave pregnancyas their reason for non-attendance. They wereparticularly concerned that the examinationwould result in miscarriage and did not believethe advice given to them by their generalpractitioner. Fear was a contributory factor in14 defaulters: fear of debilitating disease whichmight have been revealed by hospital tests,indicating the women's concerns for theirfamilies; fear of an internal examination(sometimes increased by misinformation andscaremongering by other women); and fear offurther pain associated with biopsy ortreatment. Sixteen (67%) attendees similarlyexpressed fear but they had tried to overcomeit by obtaining information and explanation;15 had sought information from externalsources. Although five had still not overcometheir fear and described terror or severeanxiety, this had not prevented them fromattending.

Implicit reasons were evident in mostinterviews with defaulters and includedmisinformation (from friends); lack ofinformation and understanding of the process;choice of doctor; apparent insensitivity ofclinical staff; and experience of the medicalprocess including pain control, intrusivequestioning (for example, about number ofsexual partners), and issues of privacy.An information booklet sent to all new

attendees was considered beneficial by mostwomen who defaulted at the review stage(12/15) and most of the attendees (16/18 whoreceived the booklet). Women who defaultedin the earlier stages found it less helpful: 11out of the 25 women in the diagnosis,treatment, and first check up stages found ittoo technical, only three of the 25 expressingany positive comments. The attendees alsocommented on the presentation and use oftechnical terms: only four felt well informedabout colposcopy and 15 had soughtinformation and explanations from othersources. Women suggested that a simplerleaflet would be more acceptable and that newinformation about how a woman might feelduring or after colposcopy and treatment anda warning of the effects of laser treatment(unpleasant smell, pain) should be included.

Table S Choice of male or female doctor for general or gynaecological examinations.Figures are numbers(percentages) of women

General examination Gynaecological examination

Defaulters Attenders Defaulters Attenders(n 40) (n = 24) (n 40) (n = 24)

Male doctor - 1(4) _ 1(4)Female doctor 16(40) 5(21) 21(53) 13(54)No preference 24(60) 18(75) 19(48) 10(42)

Most women in both groups did not mindconsulting a male or female doctor aboutgeneral medical problems, but forgynaecological problems over half in bothgroups preferred a female doctor (table 5). Asensitive, understanding approach was theparticular attitude women were looking for ina gynaecological consultation and if this waspresent the women stated that the doctor's sexwas a less important consideration.Most favourable comments on the doctor's

care came from the regular attendees, who feltthat doctors had kept them informed duringthe care process whereas negative commentsabout lack of communication came from thedefaulters. Most women interviewed (bothattendees and defaulters) could not pronouncethe word colposcopy and did not know itsmeaning.

DiscussionWomen defaulting from the colposcopy clinicdiffered in several ways from those whoattended through to discharge. Generally, theywere younger, with more responsibility forchild care, were more often pregnant, and hada lower socioeconomic status. Evidence isgrowing that health screening facilities areused less by people in poorer socioeconomiccircumstances and their knowledge of health ispoorer. This study confirms this observation.Being in paid employment seemed to makeclinic attendance more likely.

Attenders were more likely to seekinformation about colposcopy, resulting inunderstanding the preventative nature of theprocedure. The booklet provided by the clinic,although detailed, was found too technical(despite efforts by the consultant in charge toensure its readability). The importance ofproviding information appropriately fordifferent recipients has been described byBernstein6 and Hogart.7

Discussion of these results with theproviders of the clinical service has focused onpresenting the information in different writtenstyles, use of other media (for example, audioinformation for those who cannot read) andalso by improving the oral communicationskills of all clinical staff, which is particularlyimportant to working class women with whomdoctors may spend less time and give lessexplanation. Cartwright and O'Brien showedthat the average consultation between doctorsand middle class patients is longer because thisgroup of patients ask more questions andexpect more explanation.8 Most womeninterviewed in the prospective study did notknow what colposcopy meant. Posner andVessey reported a similar result.2 Both theamount of information received9 and theamount of explanation provided'0 affectcompliance with medical advice. Personaladvice and explanation supported by writteninformation improves the amount of adviceremembered." In addition, the need forspecific education aimed at pregnant womenwith an abnormal smear test result wasidentified in this study as well aselsewhere.2 12 1 Thus health professionals in

239

on March 25, 2022 by guest. P

rotected by copyright.http://qualitysafety.bm

j.com/

Qual H

ealth Care: first published as 10.1136/qshc.1.4.236 on 1 D

ecember 1992. D

ownloaded from

Sanders, Craddock, WagstafW

the colposcopy service need to see themselvesas part of the screening programme with goodliaison and provision of information to otherreferral points within health care rather than asa discrete hospital service.The role of primary health care teams is of

considerable importance as they are often thesource of the initial referral. However, oncereferred for treatment, women were unlikely tosee their general practitioner about thismedical problem and so advice had to comefrom the hospital clinic. Since more than halfof patients are referred from general practicethis could be the initial point for providinginformation. Personal communication eitherwithin primary care at referral or beforetreatment may improve understanding andoffer reassurance. Also there is the questionwhether all women need to have a hospitalreview especially when this is by a cervicalsmear. It may be more convenient for patientsto attend general practice. However this raisesissues of communication, shared protocols,and the need for mutual trust in clinical followup between medical professionals and clinicalteams. Guidelines for management and followup of women with abnormal smear test resultshave now been agreed within the district andhave been circulated to all primary care teams.However, the hospital consultants still preferto follow up routinely for two years all womenwith cervical abnormalities.The women felt issues such as possible pain

and adverse emotional reactions to theprocedures should be explained in a leaflet.Egbert et al have indicated the beneficialeffects of warning patients about pain aftersurgery; such patients needed less treatment tocontrol pain and recovered more quickly,physically and emotionally, than those whowere not warned.`' Since a substantial numberof women found biopsy painful an effectivelocal anaesthetic or other analgesic treatmentshould be offered routinely beforehand, and isnow accepted as normal clinical practice.

Fear was a key issue in non-attendance butalso was present among attendees. Marteau etal highlighted the anxiety associated withcolposcopy.4 More attention should be paid toreducing patients' stress before theconsultation. An explanation from a nursetrained in communication can result inreduction in the stress experienced bypatients. Posner and Vessey's study pointedout the need for women to receive nursingsupport, reassurance, and a period of recoverybefore being expected to leave the hospital.Greater emphasis on the counselling andsupporting role of nurses in colposcopy clinicsshould be considered.

Since the study was completed clinicalpractice has changed and the introduction ofloop diathermy has allowed both diagnosisand treatment at one visit for some women, auseful development for improving efficiency infollow up and reducing inconvenience towomen.The study highlighted problems in keeping

track of women's addresses. Small numbers ofdefaulters are highly mobile and move from

one temporary address to another. Follow upby health visitors is a possibility, but recentexperience in Liverpool indicates that evenwhen a woman is traced she may still notcomply with medical follow up. lo

Nevertheless, women need to be encouragedto keep the clinic informed about changes ofaddress, through correspondence and inposters within the clinic. In addition, bettercommunication between the clinic and theFHSA over notified recent changes of addressmay help.

Since the NHS and Community Care Actall clinical services are striving to become moreconsumer orientated. The process of servicedelivery is important. Not only should clinicsbe provided at times suitable for women whocare for children but they must be attuned towomen's holistic needs and not just to thepresenting medical problem. The benefits ofcervical screening will exceed the costs(monetary and non-monetary) only wheneffective treatment is taken up by all those withabnormal smear results.

This study identified some of the barriers totreatment which are seen by women andprovides a baseline for audit of the service. Itshows that health professionals have a key rolein reducing the barriers for women andthereby improving quality of care, throughhealth education, promoting women'sunderstanding of the treatment process,process improvements, and modifying theirbehaviour to develop a service more sensitiveto the various needs of women in differentsocioeconomic groups.

We thank the Northern Regional Health Authority. itichfunded this research from the quality assurance budget.

1 Dav NE. Screening for cancer of the cersix. _7 1 pidciolls(C ooi)o11o1iti Health 989;42: 103 6.

2 Posner T, \essev Md. )lrLonohi of LL/ icliL ) -I patit llo'fI ondon: King's Fund, 1983.

3 Quilliam S. Positincc oiivar. Penguin, 19)89.4 Marteau TMd, Walter P, (31les J, Smail M\. Anxieties in

ssomen undergoing colposcop.TBr it7 (bstet (Gv/iaLcol1990;97:859 1.

5 TIosw nsend P, Philimore P, Beattie A. HFLalt/i aL)ddplcriLantioi iiiLbqialitiLs Liil the Ao/thi. London: CroomHelm, 1987.

6 Bernstein BB. A socio-linguistic approach to sociallearning. In: Gould JB, ed. Slo'zLx (!'t//f/i)Lil sci)l/tiL.London: Penguin, 1965.

7 Hogart R. Thi ioC of litt'era.v. Middlesex: Penguin andChatto and Windus, 1984.

8 Cartwright A, O'Brien M. Social class variations in healthcare and in general practitioner consultations. In: StacyM\, ed. S'cliOlov'y'.r 'oog/a11(ngrap/h. Keele: University ofKeele, 1976:22.

cQ Kince J.1 Bradshas P, Lev P. Patients' satisfaction andreported acceptance of advice in general practice Y R (oilGm/i P/Lit 19Q7;25:558 62.

10 Korsch B, ,Gozzi E, Francis V. Gaps in the doctor-paticeltcommunication. .\LL', Eugl!]Aled 1969;280:535-40.

11 Lev P. Toward better patient-doctor communication. In:Bennett A, ed. Coouinnl ication between doctor and paticn/s.London: Nuffield Hospitals Provincial TIrust, 1976:75-98.

12 Michielutte R, Diseker RA, Young LD, Max Wj. Non-compliance in screening follow-up among familyplanning clinic patients with cervical dysplasia. P/CL Mcd1985t;4:248-58.

13 Duncan I. GuidLliees /a/t clinical p/aeti/iLrts a/id pI//granone/f /iaiagLw/iie/t. Oxford: NHS Cervical ScreeningProgramme, 1992.

14 Egbert L, Battit G, Welch C, Bartlett M. Reduction inpostoperative pain by encouragement and instruction ofpatients. .N\etw Fngl Med 1964;270:825-7.

15 Skipper J, Leonard R. Children, stress and hospitalisation;a field experiment. 7 Healtl/i oSL Be//Lie 1968;9:275-87.

16 Carter JM, Ellerbv SE. Increasing the uptake of cervicalscreen testing among Asian Wvomen [letter] B.17I(( 1;302:1 152.

240

on March 25, 2022 by guest. P

rotected by copyright.http://qualitysafety.bm

j.com/

Qual H

ealth Care: first published as 10.1136/qshc.1.4.236 on 1 D

ecember 1992. D

ownloaded from