tP

aptea

Original Research

Validity of the Stroke Rehabilitation Assessment ofMovement Scale in Acute Rehabilitation: AComparison With the Functional IndependenceMeasure and Stroke Impact Scale-16Irene Ward, PT, DPT, NCS, Susan Pivko, PT, DPT, Cert MDT,

Gary Brooks, PT, DrPH, CCS, Kate Parkin, PT, MA

Objective: To demonstrate sensitivity to change of the Stroke Rehabilitation Assessmentof Movement (STREAM) as well as the concurrent and predictive validity of the STREAM inan acute rehabilitation setting.Design: Prospective cohort study.Setting: Acute, in-patient rehabilitation department within a tertiary-care teaching hospi-al in the United States.articipants: Thirty adults with a newly diagnosed, first ischemic stroke.

Methods: Clinical assessments were conducted on admission and then again on dis-charge from the rehabilitation hospital with the STREAM (total STREAM and upperextremity, lower extremity, and mobility subscales), Functional Independence Measure(FIM), and Stroke Impact Scale-16 (SIS-16). Sensitivity to change was determined with theWilcoxon signed rank test and by the calculation of standardized response means. Spear-man correlations were used to assess concurrent validity of the total STREAM and STREAMsubscales with the FIM and SIS-16 on admission and discharge. We determined predictivevalidity for all instruments by correlating admission scores with actual and predicted lengthof stay and by testing associations between admission scores and discharge destination(home vs subacute facility).Main Outcomes: Not applicable.Results: For all instruments, there was statistically significant improvement from admis-sion to discharge. The standardized response means for the total STREAM and STREAMsubscales were large. Spearman correlations between the total STREAM and STREAMsubscales and the FIM and SIS-16 were moderate to excellent, both on admission anddischarge. Among change scores, only the SIS-16 correlated with the total STREAM. All 3instruments were significantly associated with discharge destination; however, the associa-tions were strongest for the total STREAM and STREAM subscales. All instruments showedmoderate-to-excellent correlations with predicted and actual length of stay.Conclusions: The STREAM is sensitive to change and demonstrates good concurrent andpredictive validity as compared with the FIM and SIS-16 in the acute inpatient rehabilitationpopulation.

PM R 2011;3:1013-1021

INTRODUCTION

Stroke is the leading cause of serious long-term disability [1]. Individuals who have survivedstroke frequently are challenged by decreases in voluntary motor response and functionalerformance of activities of daily living [2]. When stroke is complicated by cognitive loss,he survivor may be facing further losses of his or her employment and/or premorbid livingnvironment. Every year, approximately 795,000 people in the United States have a strokend, of those who survive, approximately 185,000 go on to have another stroke [1].

Approximately 87% of all strokes are ischemic, 10% are attributable to an intracerebral

hemorrhage, and 3% are attributable to a subarachnoid hemorrhage [3].

PM&R © 2011 by the American Academy of Physical Me1934-1482/11/$36.00 Vol.

Printed in U.S.A. D

I.W. NYU Langone Medical Center, 400 East34th Street, New York, NY 10016. Addresscorrespondence to I.W.; e-mail: ireneg.ward@gmail.comDisclosure: nothing to disclose

S.P. Graduate Center of CUNY, DPT Program,Hunter College Campus, New York, NYDisclosure: nothing to disclose

G.B. College of Health Professions, SUNY Up-state Medical University, Syracuse, NYDisclosure: nothing to disclose

K.P. NYU Langone Medical Center, New York,NYDisclosure: nothing to disclose

Peer reviewers and all others who controlcontent have no relevant financial relation-ships to disclose.

Submitted for publication January 13, 2011;accepted August 25, 2011.

dicine and Rehabilitation3, 1013-1021, November 2011

OI: 10.1016/j.pmrj.2011.08.5371013

mcdiosn

tbiMscutdi

didfmcFptpo

bcIHtfmaSat

tiSo[3

mlke

upsmdntspp

smuuhdstfbbtii

voirbbcctti

SSrbhaalS

1014 Ward et al VALIDITY OF THE STREAM IN ACUTE REHABILITATION

In stroke rehabilitation, the role of the rehabilitation teamis to facilitate the recovery of motor control and functionalindependence in preparation for the individual’s reintegra-tion into the community from an acute care or rehabilitationdepartment [4]. The present acute rehabilitation environ-

ent, which emphasizes evidence-based practice and isharacterized by rapidly decreasing lengths of stay (LOS),emands that clinicians analyze and justify assessment and

ntervention methods on the basis of objectively measuredutcomes. Therefore, valid and reliable clinical tools thatpecifically reflect the recovery of movement and function areecessary for stroke rehabilitation teams.

Although several assessment tools are currently availableo clinicians, they are used infrequently in clinical practiceecause of prolonged administration time and their complex-

ty of scoring [5-11]. The Stroke Rehabilitation Assessment ofovement (STREAM) scale, Functional Independence Mea-

ure (FIM), and Stroke Impact Scale-16 (SIS-16) are 3 out-ome measures used to document the changes in an individ-al resulting from a stroke. Combined, these scales capturehe effects a stroke has had on the individual across theomains of impairment, function, and disability, 3 domains

n the Nagi disablement model.The theoretical framework of this study was the Nagi

isablement model, which outlined 4 domains of health,ncluding pathology, impairment, functional limitation, andisability [12]. In this study, the model was applied asollows. Stroke was considered to be the pathology. A stroke

ay result in impairment, defined as a decrease in motorontrol and mobility, which was quantified by the STREAM.unctional limitations, represented by diminished ability toerform basic skills related to daily living, were measured byhe FIM. Disability ensued when a person was unable toarticipate within his or her community, the self-perceptionf which was measured by the SIS-16.

At the inception of this study, the Nagi model, developedy Saad Nagi in 1965, was recognized as a foundation oflinical practice of physical therapy [12]. Since then, thenternational Classification of Functioning, Disability andealth (ICF) has supplanted the Nagi model as the preferred

heoretical framework. The ICF uses terms such as bodyunction, activities, and participation, which parallel the Nagi

odel’s terms of impairment, functional limitation and dis-bility, respectively. Therefore, when the ICF was used, theTREAM could be considered to measure body function andctivities, the FIM to measure only activities, and the SIS-16o measure body function, activities, and participation.

The STREAM is a comprehensive outcome measure usedo efficiently examine voluntary movement and basic mobil-ty in patients who have experienced a stroke [2,11,13]. TheTREAM is intended for use in both the clinic and in researchn individuals with mild-to-severe presentations of stroke4]. Minimal equipment is required. The STREAM consists of

0 items distributed among 3 subscales: upper extremity [

ovements (10 items), for example, testing shoulder flexion;ower extremity movements (10 items), for example, testingnee extension; and basic mobility items (10 items), forxample, testing walking for 10 m [11].

Scoring for the upper and lower extremity movementsses an ordinal scale of 0 (unable to perform) to 2 (com-letely normal) voluntary limb movement [11]. A greatercore in these areas represents a greater quality of voluntaryovement. Similarly, scoring for basic mobility uses an or-inal scale of 0 (unable to perform) to 3 (independently andormally performs the skill) [11]. Some of the items areimed, such as walking for 10 m [11]. The test is performed inequence starting from the patient in a supine position,rogressing to a seated position, and ending in a standingosition [4].

The FIM is an 18-item ordinal scale used with all diagno-es within a rehabilitation population and is designed toeasure functional ability regardless of impairments in vol-ntary movement. The FIM is mandated by payers and issed in the equation to estimate LOS in acute rehabilitationospitals in the United States [14,15]. Therefore, it has airect impact on health care reimbursement [14]. It is con-idered most useful for assessment of progress during inpa-ient rehabilitation [15]. The motor FIM consists of theollowing functional skills: dressing one’s lower and upperody, toileting, bathing, eating, grooming, managing one’sladder and bowels, transferring to bed, transferring to theoilet, walking, and negotiating stairs [16]. Scores for eachtem range from 1 (dependent) to 7 (independent) depend-ng on the subject’s performance of the skill [16].

The SIS-16 is a questionnaire, administered to the patientia interview, that addresses patients’ perceptions of theirwn recovery status by assessing their self-perceived disabil-ties [17]. Patients are asked to reflect on the last 2 weeks andank the difficulty in performing the following skills: upperody dressing, bathing, toileting, controlling bladder andowels, standing balance, shopping, performing householdhores, sitting balance, walking balance, transferring bed tohair, walking speed, negotiating stairs, walking distances,ransferring to cars, and carrying heaving objects when usinghe impaired side.[17] The SIS-16’s ordinal scale for eachtem ranges from 1 (not able to do) to 5 (able to do) [17].

In several research studies in which they compared theTREAM with other assessment tools, investigators found theTREAM to possess greater responsiveness and interratereliability and fewer ceiling effects. The STREAM was alsoetter in predicting discharge destination from an acute careospital, functional outcome, and gait speed [13,18,19]. Theuthors of previous reports established high interrater reli-bility, intrarater reliability, content validity, concurrent va-idity, and moderate-to-high convergent validity in theTREAM in patients with subacute or chronic stroke

2,4,11]. This study intended to extend the clinical utility of

1015PM&R Vol. 3, Iss. 11, 2011

the STREAM in an acute inpatient rehabilitation population.The objectives of this study were as follows:

● to investigate whether the STREAM detected a change inmotor abilities of acute stroke patients from admission todischarge in the current acute rehabilitation population;

● to determine concurrent validity of the STREAM with 2outcome measures, the FIM and SIS-16, which commonlyare used in stroke rehabilitation at admission and dis-charge; and

● to determine the predictive validity of the STREAM atadmission with respect to LOS and discharge destination(home or subacute facility).

METHODS

In this prospective cohort study we used the STREAM, FIM,and SIS-16, administered at admission and again at dis-charge, to track clinical changes in individuals with acutestroke who underwent acute inpatient rehabilitation. Theoutcome measures were administered within the first 3 daysof the evaluation period and again just before discharge. Thisstudy was approved by the institutional review boards of theparticipating institutions.

Consent

A combined Health Insurance Portability and AccountabilityAct Authorization Form and Consent Form was distributedto and signed by each patient before he or she participated inthe study. Because aphasia and/or decisional incapacity arepotential consequences of a stroke, the consent form was anaphasic-friendly pictorial consent form that facilitated in-formed consent and authorization for research and for vid-eotaping. In addition, if comprehension was in question, aMini-Mental State Examination was performed during thepatient interview. A minimum score of 16 on the Mini-Mental State Examination identified that the patient was ableto provide consent [20-22]. If a patient scored less than 16,he or she was represented by a proxy, or a caregiver ap-pointed by the proxy, who provided consent and authoriza-tion for research and administration of the SIS-16. The sameproxy or caregiver was used to complete the SIS-16 onadmission and discharge to assure consistent reporting onthe patient’s perceived progress. To ensure confidentiality, alldocuments were stored in a locked cabinet within a lockedoffice of the principal investigator.

Participants

Individuals who were admitted to the stroke program at theacute rehabilitation center with the diagnosis of stroke werescreened for entry into the research study with the use ofhospital databases and medical records. Figure 1 details the

flow of participants through the study. Demographic and

baseline clinical information was collected on each patient.Individuals included in the study were adults who wereadmitted to an acute inpatient rehabilitation department of ametropolitan tertiary-care center in the United States whohad sustained, for the first time, a focal or multifocal isch-emic, unilateral, cerebral stroke. Exclusion criteria for thisstudy were bilateral cerebral, brainstem, and cerebellarstrokes or the presence of other neurological conditionsbefore the stroke.

Training of Raters

Five physical therapists were trained to administer theSTREAM and SIS-16. Training for STREAM administrationincluded reviewing the STREAM instruction sheet and glos-sary of terms, watching and concurrently scoring a demon-stration of an experienced physical therapist using theSTREAM, and conducting the STREAM while supervised bya trained physical therapist. Training for administration ofthe SIS-16 included reviewing a copy of the survey withinstructions, watching a demonstration of an experiencedphysical therapist using the SIS-16, and conducting theSIS-16 while supervised by a trained physical therapist. Useof the FIM was mandatory for the acute rehabilitation facility[14]. The motor section of the FIM was administered by amultidisciplinary team of health care professionals composedof nurses, occupational therapists, and physical therapists.Training included case scenarios, the scoring of which re-quired receiving a passing score on a written examination. AllFIM raters took and passed the examination.

Data Collection

The STREAM and SIS-16 were administered by a pair oftrained raters within the 3-day evaluation period immedi-ately after admission and repeated just before discharge.During STREAM testing, one rater administered the test, andboth raters scored simultaneously. The raters compared anddiscussed the test, item by item, and resolved any disagree-ments by consulting the instruction manual and arriving at aconsensus. In accordance with STREAM guidelines, testingwas completed in the sequence provided and within 24 hoursof beginning the examination. Patients were permitted restbreaks when needed. To ensure objectivity, the paired teamswere rotated and never included the primary treating physi-cal therapist.

Administration of the SIS-16 typically followed adminis-tration of the STREAM. Following SIS guidelines, the SIS-16was administered by way of interview, where the recorderread the SIS-16 questionnaire to the subject as the subjectread along [17,20]. Only one patient requiring a health careproxy had one on file, and that individual completed a SIS-16on behalf of the patient on admission and at discharge. The

FIM was administered independently of the STREAM and

1016 Ward et al VALIDITY OF THE STREAM IN ACUTE REHABILITATION

Figure 1. Flow of participants through the study.

1017PM&R Vol. 3, Iss. 11, 2011

SIS-16 as mandated during the first days of hospitalizationand again 3 days before discharge by a separate set of FIM-certified health care professionals [14]. The motor FIM datawere retrieved from the medical record. Demographic infor-mation and clinical characteristics were collected from theelectronic medical record for descriptive purposes (Table 1)[14].

Data Storage and Confidentiality

All data were compiled on a secure master spreadsheet.Confidentiality of the participants was maintained by the useof a data coding system that assigned a unique number toeach patient. This number was the sole identifier on alldocuments related to the research. Confidentiality was as-sured by storing all documents in a locked cabinet. Only theinvestigators were able to retrieve this information.

Data Analysis

Using results reported in a previous study in a similar popu-lation that demonstrated an 11-point change in STREAMscores during the 5 weeks after admission to a rehabilitationhospital, we determined that a sample size of 33 would beneeded in the current study to detect a similar change at 80%

Table 1. Demographic and clinical characteristics of partici-pants

Gender, N (%)Female 16 (63.3)Male 14 (46.7)

Race, N (%)Asian 1 (3.3)African-American 6 (20.0)White 22 (73.3)Not identified 1 (3.3)

Side of primary involvement, N (%)Left 19 (63.3)Right 11 (36.7)

Location of lesion, N (%)Deep 14 (46.7)Lobar 15 (50.0)Missing 1 (3.3)

Discharge destination, N (%)Home 23 (76.7)Subacute facility 7 (23.3)

Age, mean (SD) 66.5 (13.7)Range 40-90�

Days since stroke onset, mean (SD) 7.8 (3.5)Range 3-15

Predicted length of stay, days,* mean (SD) 18.1 (5.9)Range 8-28

Actual length of stay (days), mean (SD) 23.3 (9.7)Range 7-53

FIM � Functional Independence Measure.*Predicted length of stay is derived from a calculation that includes motor

FIM scores [14].

power at an alpha of 0.05 [13]. Because of challenges related

to recruitment, this study did not achieve the targeted 33participants. However, the post-hoc power analysis, forwhich we used data from the 30 current participants, re-vealed that the study had a power greater than 0.99 to detecta clinically meaningful change in the STREAM from admis-sion to discharge.

Individual FIM task scores were summed to create acomposite FIM motor score. For the SIS-16, each partici-pant’s raw scores were summed and converted to a percent-age score [17,20]. Scores for the STREAM were determinedby ratings for activities that were attempted because someitems may not have been tested for reasons such as pain orinadequate range of motion [4]. Raw STREAM ratings weresummed and divided by the total possible score for itemstested to calculate percentage scores for each of the 3STREAM subscales (mobility, upper extremity, and lowerextremity) and for the STREAM total score. Change scores foreach of the outcomes were determined by subtracting admis-sion scores from discharge scores. For all 3 outcomes, posi-tive change scores indicated improvement.

Demographic and clinical characteristics of the partici-pants were described by the use of means and standarddeviations for numeric variables or counts and percents forcategorical variables. Distributions for the motor FIM, SIS-16, and STREAM total and subscale scores were assessed fornormality by inspecting the histograms and normal probabil-ity plots and with the Shapiro-Wilk statistic. Because distri-butions for total STREAM and STREAM subscale scores werenon-normal, criterion validity of admission, discharge, andchange in total STREAM and STREAM subscale scores wasdetermined by computing Spearman rank correlations withadmission, discharge and change in motor FIM and SIS-16scores, respectively.

Change scores were subjected to the Wilcoxon signed-rank test to determine the statistical significance of improve-ments from admission to discharge. Responsiveness tochange was assessed by calculating standardized responsemeans (SRMs) for the 3 instruments, including STREAMsubscales. The SRM is a measure of internal responsivenessthat is calculated by dividing the mean of the change score, ie,the pre- to posttreatment improvement, by the SDs of thechange score [23]. Expressing change in SD units allowed forcomparisons of instruments that use different numeric scales.Changes indicated by SRM values of 0.20, 0.50, and 0.80 andgreater have been characterized as small, moderate, andlarge, respectively [23].

Associations between baseline (admission) FIM, SIS-16,and STREAM scores and discharge destination (home orsubacute facility) were determined with Wilcoxon rank-sumtests. Spearman rank correlations were used to identify asso-ciations between baseline (admission) FIM, SIS-16, andSTREAM scores and predicted and actual LOS. PredictedLOS, in days, was determined on admission by an algorithm

with several clinical measurements, including the admission

d

0, P

T

L

FM

1018 Ward et al VALIDITY OF THE STREAM IN ACUTE REHABILITATION

FIM score. Actual LOS is the number of days that elapsedbetween admission and discharge. For correlation analyses,Spearman coefficients between 0.25 and 0.50 were consid-ered to be fair, between 0.51 and 0.75, moderate to good, andgreater than 0.75, good to excellent [24]. Statistical signifi-cance was determined at P � .05. All analyses were con-

ucted with SAS v. 9.1. (SAS Institute, Cary, NC).

RESULTS

Thirty patients completed the study (Figure 1). Demographicand clinical characteristics of the participants are displayed inTable 1. Mean age was 66.5 years. Most participants werewhite and had left-sided cerebral involvement. Neither gen-der held a substantial majority. The mean number of dayssince stroke onset was 7.8 and ranged from 3 to 15 days. LOSaveraged 23.3 days, ranging from 7 to 53 days. Most partic-ipants were discharged home, but nearly one-quarter weredischarged to a subacute facility.

Mean scores on admission and discharge for STREAM,FIM, and SIS-16 as well as change scores are displayed inTable 2. Change scores indicated statistically significant im-provement from admission to discharge in all outcomes,including STREAM subscales. SRMs for all outcomes werelarge, with the motor FIM being the largest [23]. SRMs for thetotal STREAM and SIS-16 were comparable, and SRMs forSTREAM subscales were slightly lower.

Table 2. Mean (SD) admission, discharge, and change score

Admission Score

FIM (motor) 36.8 (10.6)SIS-16 36.0 (17.5)STREAM

Total 60.8 (29.9)Mobility subscale 59.7 (24.0)UE subscale 59.3 (39.3)LE subscale 66.5 (32.0)

FIM � Functional Independence Measure; SIS-16 � 16-Item Stroke ImpaAssessment of Movement.*Change score � discharge score – admission score. All change scores �

able 3. Spearman correlations of STREAM scores with admiss

STREAM

Admission

FIM SIS-16

Total 0.7766 0.7073�.0001 �.0001

Mobility subscale 0.6501 0.6451.0001 .0001

UE subscale 0.7489 0.6088�.0001 .0004

E subscale 0.7905 0.5992�.0001 .0005

IM � Functional Independence Measure; LE � lower extremity; SIS-16ovement; UE � upper extremity.

*Values include Spearman rho coefficients and corresponding P-values.

Spearman correlations between admission, discharge, andchange in STREAM scores with respective FIM and SIS-16scores are found in Table 3. Admission and discharge corre-lations were moderate to excellent, and all were statisticallysignificant. Positive correlations indicate that STREAM scoresincreased along with increases in FIM or SIS-16 scores.Among the correlations between STREAM change scores andFIM and SIS-16 change scores, change in the total STREAMcorrelated moderately with change in the SIS-16. Otherwise,total STREAM change scores did not correlate with the FIM,and the separate STREAM subscale change scores did notcorrelate with change scores of the FIM or SIS-16.

As seen in Table 4, all 3 outcome measures were signifi-cantly correlated with predicted and actual LOS. Motor FIMhad the strongest correlation (rho� �0.9438) with pre-dicted LOS, but the weakest correlation (rho � �0.6846)with actual LOS. Of all outcomes, total STREAM and SIS-16had the strongest correlations with actual LOS (rho ��0.7972 and �0.6846, respectively). The negative correla-tions indicated that as outcome scores increased, LOS de-creased. In addition, as displayed in Figure 2, all instrumentswere significantly associated with discharge destination suchthat participants who were discharged home had signifi-cantly greater scores than those who were discharged to asubacute facility. These associations were stronger for totalSTREAM and STREAM subscales than for the FIM or SIS-16.

otor FIM, SIS-16 percent, and STREAM percent scores

harge Score Change Score* SRM

0.5 (11.7) 23.7 (10.1) 2.349.1 (15.8) 23.1 (14.0) 1.65

3.1 (26.1) 12.3 (8.8) 1.401.6 (23.7) 11.9 (14.4) 0.831.8 (32.7) 12.4 (12.9) 0.976.7 (26.5) 10.2 (9.9) 1.03

; SRM � standardized response mean; STREAM � Stroke Rehabilitation

� .0001.

ischarge, and change FIM and SIS-16 scores*

Discharge Change

FIM SIS-16 FIM SIS-16

0.7802 0.7153 0.2535 0.4456.0001 �.0001 .1765 .0136

0.8292 0.7985 0.3055 0.2655.0001 �.0001 .1007 .1561

0.7012 0.5499 0.1277 0.2461.0001 .0016 .5011 .1898

0.6954 0.6371 0.2811 0.1955.0001 .0002 .1324 .3006

em Stroke Impact Scale; STREAM � Stroke Rehabilitation Assessment of

s for m

Disc

65

7777

ct Scale

ion, d

�

�

�

�

� 16-It

Tnawftt

MSS

s

W

1019PM&R Vol. 3, Iss. 11, 2011

DISCUSSION

In this study we investigated (1) whether the STREAM de-tected change in motor abilities of patients with stroke fromadmission to discharge in acute rehabilitation; (2) whetherthe STREAM correlated with 2 outcome measures commonlyused in stroke rehabilitation, the FIM and SIS-16 at admis-sion and discharge and; (3) whether the STREAM had pre-dictive validity with regard to LOS and discharge destination.

Total STREAM and STREAM subscales demonstratedgood responsiveness to change as seen in the large SRMs.Both the SIS-16 and motor FIM also displayed large SRMs,suggesting that all 3 instruments captured clinical change ina population in whom substantial change was expected [25].

The findings also supported several aspects of validity ofthe STREAM in a population with stroke in an acute inpatientrehabilitation department. Concurrent validity was sup-

Table 4. Spearman correlations between admission FIM, SIS-16, and STREAM scores, and predicted and actual length ofstay*

Predicted Lengthof Stay

Actual Lengthof Stay

otor FIM �0.9438 �0.6846IS-16 �0.6743 �0.7953TREAMTotal �0.8011 �0.7972Mobility subscale �0.6361 �0.7423UE subscale �0.7717 �0.7469LE subscale �0.8446 �0.7364

P-values for all correlations are �.0001 except for STREAM mobilityubscale and predicted length of stay, P � .0002FIM � Functional Independence Measure; LE � lower extremity; SIS-16 �

16-Item Stroke Impact Scale; STREAM � Stroke Rehabilitation Assessment ofMovement; UE � upper extremity.*Values include Spearman rho coefficients.

Figure 2. Admission FIM, SIS-16, and STREAM scores by dischar

ilcoxon rank-sum test. LE � lower extremity; UE � upper extremity.

ported by the moderate-to-excellent positive correlationsbetween admission and discharge STREAM scores and therespective scores on the motor FIM and SIS-16. Predictivevalidity of the STREAM was supported by the associationsbetween baseline STREAM scores and discharge destinationand the moderate-to-excellent negative correlations with ac-tual LOS. That change scores between the FIM or SIS-16 andthe STREAM did not correlate, or correlated only weakly,with change in total STREAM and change in SIS-16 was notsurprising. As Husted and associates [23] pointed out, stud-ies in which all participants improved are characterized bydiminished variability in change scores. In the absence ofsubstantial variability (ie, where some participants improved,some declined) correlations between measures of changemay not have been seen [23]. In the current study, improve-ment in all measures was expected because participants werereceiving rehabilitation treatment after stroke and spontane-ous recovery was occurring concurrently. Thus, despite thepresence of significant correlations between admission anddischarge scores, change scores were not expected to vary asmuch between measures.

The admission STREAM was more strongly associatedwith actual LOS than the admission motor FIM (Table 4).This finding was surprising because admission motor FIMscores have been used in the calculation of predicted LOS(which was observed by the strong correlation between ad-mission motor FIM scores and predicted LOS; Table 4) [14].

his finding may be explained in part by the 2-dimensionalature of the STREAM. The STREAM measured performancet both the impairment level and at the functional level,hereas the motor FIM measured performance only at the

unctional level. Also, timed function at admission may addo the predictive value of the STREAM. The STREAM and nothe FIM had items that were actually timed, namely, bed

position. All within-scale differences are significant at P � .05,

ge dis

1020 Ward et al VALIDITY OF THE STREAM IN ACUTE REHABILITATION

mobility, standing tolerance, and ambulation. This may ex-plain why admission STREAM scores were stronger predic-tors of discharge home compared with the motor FIM andSIS-16. These results contributed to the validity of the use ofthe STREAM in monitoring change in the acute in-patientrehabilitation department in individuals who have had aStroke.

Strengths

Included in this study were individuals with ischemic stroke,which represents the most common type of stroke [3]. Also,this study had mechanisms in place, such as an aphasia-friendly authorization and consent form, that made admit-ting patients with communication deficits possible and aprocedure for using a patient’s health care proxy to provideconsent. Participants with either hemorrhagic, brain stem, orcerebellar strokes were excluded because of the diffuse im-pact on movement associated with those types of lesion andthe possibility of reducing interrater reliability. Participantswith a previous history of stroke were excluded to limit theinfluence of learned compensations in movement from pre-vious insult(s), which may impact how recovery is measuredby the outcome measures used in this study.

The use of a patient-driven objective measure (SIS-16) tocapture the participants’ experience enriched the data byoffering the patients’ perspective. STREAM testers were al-ways paired as a measure of improving the consistency ofscoring, and testers were independent from treating clini-cians. In contrast to an earlier study performed in Canada inwhich the authors compared the STREAM with the BarthelIndex, a 10-item survey of functional independence, thisstudy used the FIM, a mandated functional assessment mea-sure used in inpatient rehabilitation hospitals in the UnitedStates [13,14,26].

Limitations

Many patients who were admitted to the rehabilitation hos-pital in which the study occurred did not have an appointedhealth care proxy, which had an impact on the heterogeneityof the patient population recruited for this study becausepatients with profound cognitive and/or communication def-icits were not capable of providing consent and therefore notincluded in this study if a healthcare proxy was not identified.Only one patient had a documented proxy who was able tocomplete the consent and SIS-16 on admission and at dis-charge.

Another limitation was that the STREAM testers were notblinded to the SIS-16. However, the SIS-16 was almostalways performed after the STREAM had been performed andit is a patient, self-reported measure. Patients with bilateralcerebral, brainstem, and cerebellar strokes or other neurolog-

ical conditions before the stroke were excluded from the

study, and therefore the results of this study could not begeneralized to these patients. Although the STREAM wasresponsive and sensitive to change (Table 2), it showed theleast amount of change of the 3 outcome measures. This maybe explained in part by a smaller range of potential changethat is measurable because of the scoring of the STREAM. Forthe STREAM, scoring of each item is determined on the basisof 3 or 4 levels, whereas scoring of each item for SIS-16 andFIM is determined on the basis of 5 and 7 levels, respectively.

CONCLUSION

This study extends the clinical utility of the STREAM inidentifying recovery from a stroke in an acute inpatientrehabilitation department in the United States. Because theSTREAM was a better predictor of actual LOS than the FIM,future studies should identify what impact, if any, the timedcomponents of the STREAM have on the predictive nature ofthe STREAM as compared with the FIM. In addition, futureinvestigators may analyze the relationships of different typesof strokes on testing reliability and the relationship of indi-vidual items in the STREAM, SIS-16, and FIM.

ACKNOWLEDGMENTS

We thank Divya Mathur PT, MPA, Daniella Leone, PT, DPT,Alicia Esposito PT, DPT, MS, and Randi Farkas, MA, CCC-SLP, for data collection; Taly Gofer, PT, DPT, Shira Klein, PT,DPT, Shoshana Nehmad, PT, DPT, Deena Elias, DPT, SophiaShayek, DPT, and Talia Sosnovitch, DPT for literature reviewand data entry; Daniel Labovitz, MD, MS, and Ira Rashbaum,MD, for medical consultation; Martha Sarno, MA, MD(hon),CCC-SLP, BC-NCD, for assistance in creating the aphasic-friendly consent form; and the patients who participated inthe study.

REFERENCES1. Centers for Disease Control and Prevention. Stroke facts. Available

from: http://www.cdc.gov/stroke/facts.htm. Accessed September 8,2011.

2. Wang CH, Hsieh CL, Dai MH, Chen CH, Lai YF. Inter-rater reliabilityand validity of the Stroke Rehabilitation Assessment of Movement(STREAM) instrument. J Rehabil Med 2002;34:20-24.

3. Heart Disease and Stroke Statistics 2010 update [homepage on theInternet]. Available from: http://circ.ahajournals.org/content/121/7/e46.full.pdf. Accessed October 6, 2011.

4. Daley K, Mayo N, Danys I, Cabot R, Wood-Dauphinee S. The StrokeRehabilitation Assessment of Movement (STREAM): Refining and vali-dating the content. Physiother Can 1997;Fall:269-278.

5. Brunnstrom S. Movement Therapy in Hemiplegia: A Neurophysiologi-cal Approach. New York: Harper and Row; 1970.

6. Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient. 1. A method for evaluation of physicalperformance. Scand J Rehabil Med 1975;7:13-31.

7. Bobath B. Adult Hemiplegia: Evaluation and Treatment. 2nd ed. Lon-

don: Heinemann Medical Books; 1978.

1

1021PM&R Vol. 3, Iss. 11, 2011

8. Ashburn A. A physical assessment for stroke patients. Physiotherapy1982;68:109-113.

9. Carr JH, Shepherd RB, Nordholm L, Lynne D. Investigation of a new motorassessment scale for stroke patients. Phys Ther 1985;65:175-180.

10. Gowland C. Standardized physical therapy measurements for assessingimpairment and disability following stroke. Neurol Rep 1991;15:9-14.

11. Daley K, Mayo N, Wood-Dauphinee S. Reliability of scores on theStroke Rehabilitation Assessment of Movement (STREAM) measure.Phys Ther 1999;79:8-19; quiz 20-23.

12. American Physical Therapy Association. Guide to Physical TherapyPractice. 2nd ed. Phys Ther 2001;81:9-746.

13. Ahmed S, Mayo NE, Higgins J, Salbach NM, Finch L, Wood-DauphineeSL. The Stroke Rehabilitation Assessment of Movement (STREAM): Acomparison with other measures used to evaluate effects of stroke andrehabilitation. Phys Ther 2003;83:617-630.

14. Centers for Medicare & Medicaid Services. IRF-PAI Training Manual;2002 [updated January 16, 2002]. Available from: http://www.cms.gov/inpatientrehabfacpps/downloads/irfpai-manualint.pdf. AccessedSeptember 8, 2011.

15. The Center for Outcome Measurement in Brain Injury [homepage onthe Internet]. Available from: http://www.tbims.org/combi/FIM. Ac-cessed September 8, 2011.

6. Rules and regulations. Fed Register 2005;70:47898. Available at:http://www.gpo.gov/fdsys/pkg/FR-2005-08-15/pdf/05-15419.pdf. Ac-cessed October 7, 2011.

17. Duncan PW, Lai SM, Bode RK, Perera S, DeRosa J. Stroke Impact Scale-16:A brief assessment of physical function. Neurology 2003;60:291-296.

18. Hsueh IP, Wang CH, Sheu CF, Hsieh CL. Comparison of psychometricproperties of three mobility measures for patients with stroke. Stroke2003;34:1741-1745.

CME QuestionThe Stroke Rehabilitation Assessment of Movement Scale (STREAM)

a. participation based on mobilityb. limb spasticityc. mobility and voluntary movementd. fine motor control

Answer online at me.aapmr.org

19. Hsieh YW, Wang CH, Sheu CF, Hsueh IP, Hsieh CL. Estimating theminimal clinically important difference of the Stroke RehabilitationAssessment of Movement measure. Neurorehabil Neural Repair 2008;22:723-727.

20. Duncan PW, Wallace D, Lai SM, Johnson D, Embretson S, Laster LJ.The Stroke Impact Scale version 2.0: Evaluation of reliability, validity,and sensitivity to change. Stroke 1999;30:2131k-2140.

21. Folstein MF, Folstein SE, McHugh PR. Folstein Mini-Mental: Mini-Mental State. A practical method for grading the cognitive state ofpatients for the clinician. J Psychiatr Res 1975;12:189-198.

22. Husted JA, Cook RJ, Farewell VT, Gladman DD. Methods for assessingresponsiveness: A critical review and recommendations. J Clin Epide-miol 2000;53:459-464.

23. Portney LG, Watkins MP. Foundations of Clinical Research: Appli-cations to Practice. 3rd ed. Upper Saddle River, NJ: Prentice Hall;2009.

24. Verheyen G, Nieuwboer A, De Wit L, et al. Time course of trunk, arm,leg and functional recovery after ischemic stroke. Neurorehabil NeuralRepair 2008;22:173-179.

25. Mahoney FL, Barthel BD. Functional evaluation: The Barthel Index. MdState Med J 1965;14:56-61.

This CME activity is designated for 1.0 AMA PRA Category 1 Credit™ andcan be completed online at me.aapmr.org. Log on to www.me.aapmr.org,go to Lifelong Learning (CME) and select Journal-based CME from thedrop down menu. This activity is FREE to AAPM&R members and $25 fornon-members.

ost-stroke outcome measure that examines:

is a p