DESPITE IMPROVEMENTS IN THEIR general physical

Transcription

1 1616 ORIGINAL ARTICLE Test-Retest Reliability, Validity, and Sensitivity of the Chedoke Arm and Hand Activity Inventory: A New Measure of Upper- Limb Function for Survivors of Stroke Susan R. Barreca, Dip PT, BA, Paul W. Stratford, MSc, PT, Cynthia L. Lambert, BSc, PT, Lisa M. Masters, MSc, OT, David L. Streiner, PhD, C Psych ABSTRACT. Barreca SR, Stratford PW, Lambert CL, Masters LM, Streiner DL. Test-retest reliability, validity, and sensitivity of the Chedoke Arm and Hand Activity Inventory: a new measure of upper-limb function for survivors of stroke. Arch Phys Med Rehabil 2005;86: Objectives: To estimate the test-retest reliability and validity of the Chedoke Arm and Hand Activity Inventory (CAHAI) and to test whether the CAHAI was more sensitive to change in upper-limb function than the Impairment Inventory of the Chedoke-McMaster Stroke Assessment (CMSA) and the Action Research Arm Test (ARAT). Design: Construct validation process. Setting: Inpatient/outpatient rehabilitation facilities. Participants: Stratified sample of 39 survivors of stroke: 24 early (mean age, 71.4y; mean days poststroke, 27.3) and 15 chronic (mean age, 64.0y; mean days poststroke, 101.7). Intervention: Regular therapy. Main Outcome Measures: Intraclass correlation coefficients (ICCs), receiver operating characteristic (ROC), standard error of measurement, and correlation coefficients. Results: High interrater reliability was established with an ICC of.98 (95% confidence interval [CI],.96.99). The minimal detectable change score was 6.3 CAHAI points. Higher correlations were obtained between the CAHAI and the ARAT and CMSA scores compared with the CMSA shoulder pain scores (1-sided, P.001). Areas under the ROC curves were as follows: CAHAI,.95 (95% CI, ); CMSA,.76 (95% CI,.61.92); and ARAT,.88 (95% CI, ). Conclusions: High interrater reliability and convergent and discriminant cross-sectional validity were established for the CAHAI. The CAHAI is more sensitive to clinically important change than the ARAT. Key Words: Arm; Cerebrovascular accident; Hand; Outcome assessment (health care); Recovery of function; Rehabilitation by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation From the Orthopedic and Rehabilitation Services, Hamilton Health Sciences, Hamilton, ON (Barreca, Lambert, Masters); School of Rehabilitation Science, Mc- Master University, Hamilton, ON (Barreca, Stratford); and Kunin-Lunenfeld Applied Research Unit, Baycrest Centre for Geriatric Care, Toronto, ON (Streiner), Canada. Supported by the Ontario Ministry of Health and Long Term Care through the Heart and Stroke Foundation of Ontario Initiative. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the author(s) is/are associated. Reprint requests to Susan Barreca, McMaster University, Box 2000, Station A, Hamilton, ON L8M 3Z5, Canada, barreca@hhsc.ca /05/ $30.00/0 doi: /j.apmr DESPITE IMPROVEMENTS IN THEIR general physical mobility, 1 many survivors of stroke continue to experience great difficulty in regaining functional use of their affected arms and hands. 2-5 Although impairment measures, generic health status measures, and client satisfaction surveys may be used to evaluate upper-limb recovery, 6 functional assessments are considered key in evaluating therapy aimed at improving how survivors of stroke use the paretic upper limb in their daily lives. 7,8 Validating a scale is a process whereby one determines the degree of confidence one can have in inferences made about people based on their scores from that scale. 9 An important aspect of validity is a measure s ability to detect the magnitude of change over time in a person or a group. It is this latter type of measure that is required to evaluate arm and hand functional recovery in survivors of stroke and the effectiveness of therapeutic intervention programs designed to foster this recovery. One of the most commonly used upper-limb assessments 2 is the Action Research Arm Test (ARAT), 10 a shortened version of the Carroll test that was developed in This measure, which consists of 19 movements that are grouped into 4 subtests, was never explicitly developed for the stroke population. 10 Four of the items are impairment-oriented items (eg, touch top of head, hand behind head), and the remaining items are contrived functional items (eg, pick up ball bearings of various sizes between the thumb and each digit). In studies that used the ARAT to assess poststroke upper-limb function, a floor effect has often been observed. 2 Finally, none of the existing scales, including the ARAT, take account of patients preferences for upper-limb tasks. Conversely, the Chedoke Arm and Hand Activity Inventory (CAHAI) was methodically developed to include relevant functional tasks (appendix 1), meet the World Health Organization s definition of activity, 12 and be sensitive to clinically important changes in upper-limb function. Face, content, and factorial validity of the CAHAI have been reported elsewhere. 13 In brief, 751 items were generated from survivors of stroke, their caregivers, and the literature. 13 Twenty-six potential tasks were administered to 145 survivors of stroke participating in active rehabilitation to assess the frequency endorsement, factorial loading, and administration feasibility of those items. 13 A balance between statistical analyses and clinical judgment was used to select the final items. The CAHAI consists of 13 real-life functional tasks that reflect (1) the domains deemed important by survivors of stroke; (2) bilateral activities; (3) non gender-specific items; (4) the full range of normative movements, pinches, and grasps; and (5) the various stages of motor recovery poststroke. Accompanied by a detailed instruction manual, the CAHAI may be completed in approximately 25 minutes. Because the internal consistency (.98) of the measure and the single-item factor loadings (range,.76.96) are high, the CAHAI appears

2 CHEDOKE ARM AND HAND ACTIVITY INVENTORY, Barreca 1617 Characteristic Table 1: Patient Characteristics to be assessing a single construct that is, upper-limb function. 14 The primary purpose of this study was to estimate test-retest reliability, validity, and sensitivity to change of the CAHAI. A secondary purpose was to test whether the CAHAI was more adept than the impairment inventory of the Chedoke-McMaster Stroke Assessment (CMSA) 15,16 and the ARAT 10 at distinguishing change in survivors of stroke who had an acute onset with mild to moderate motor impairments (where we expect to see considerable change) from that in people who had chronic, severe motor impairments (where we expect to see no change). The ultimate goal when assessing sensitivity to change is to distinguish among patients whose health status has improved, deteriorated, or remained stable. 17 Our hypotheses were as follows: In the null hypothesis, there would be no differences in the abilities of the CAHAI, CMSA, and the ARAT to distinguish change in patients with acute onset of stroke and mild to moderate motor impairments from change in patients with chronic presentation. For the alternative hypothesis, the CAHAI would be more adept than the CMSA and the ARAT in distinguishing change (1-sided, P.05) in patients with acute onset and mild to moderate motor impairments than in patients with chronic, severe deficits. METHODS Value Sex (F/M) 19/20 Stratification (acute/chronic) 24/15 Days postonset Acute (Q)* 27.3 (3 63) Chronic (Q)* (81 342) Age (y) Acute (Q)* 71.4 ( ) Chronic (Q)* 64.0 ( ) Dominant/nondominant 19/20 Affected side (R/L/bilat) 22/16/1 Infarct/hemorrhage 32/6 ICD-9-CM code Neglect (present/absent) 22/17 Hemianopsia (present/absent) 8/30 Depression (present/absent) 7/32 Proprioception (absent/impaired/normal) 5/14/18 NOTE. Values are for n 39 unless otherwise noted. Abbreviations: bilat, bilateral; F, female; ICD-9-CM, International Classification of Diseases, Clinical Modification, 9th Revision; L, left; M, male; Q, quartile; R, right. *First and third quartiles. n 38. n 37. Participants A prospective cohort of 39 participants with a complete first episode of stroke (table 1) met the study s inclusion criteria. Stroke survivors were excluded if they (1) had a lower motoneuron injury, (2) had a preexisting disability of the arm or hand apart from the stroke, (3) lacked sufficient stamina to participate, or (4) achieved a combined CMSA score greater than 11 (because there would be no room for change [improvement] over time). Measures Chedoke Arm and Hand Activity Inventory. The CAHAI is a functional measure with 13 items that are assessed using a 7-point quantitative scale, similar to that of the FIM instrument, 18 to give a total raw sum of 91 (minimum score, 13) that may be easily converted to a percentage. Action Research Arm Test. The ARAT is a partial functional/impairment measure with 19 items that are assessed using a 4-point ordinal quantitative scale (0, no movement possible; 1, movement partially performed; 2, movement performed but abnormally; 3, movement performed normally) to give a total raw score of Interrater reliability is.99, and test-retest reliability is Chedoke-McMaster Stroke Assessment. Developed from the original work by Brunnstrom, 20 the CMSA is an impairment measure with 6 subscales (recovery stage of the arm, hand, leg, and foot; postural control; shoulder pain). Each dimension is scored using a 7-point ordinal scale for a maximum total score of 42 and a minimum score of 6. 15,16 We examined the 2 subscales that classify poststroke arm and hand motor recovery into 7 separate stages for a maximum total score of 14 and a minimum score of 2 (1, flaccid paralysis; 2, spasticity is present and felt as a resistance to passive movement; 3, marked spasticity but voluntary movement present within synergistic patterns; 4, spasticity decreases; 5, spasticity wanes but is evident with rapid movement at the extremes of range; 6, coordination and patterns of movement are near normal; 7, normal movement). Interrater reliability is.97, and test-retest reliability is Design In a review of different designs and analytic strategies for assessing change scores, Stratford et al 17 recommend a strong design that makes use of the clinical history of the condition of interest. A key component of this design is to have 2 or more cohort groups of survivors of stroke whose health status, based on previous evidence, is expected to experience various amounts of change. From the stroke literature, we identified key prognostic indicators as (1) the initial arm and hand motor impairment and (2) time poststroke. 3,5,21-27 Consequently, we stratified participants into 2 groups based on these known prognostic indicators for small or large amounts of change in poststroke arm and hand function over time. Group 1, the acute/mild-to-moderate group (n 24), consisted of individuals who were (1) 8 weeks or less postonset of stroke and (2) had a combined initial CMSA arm and hand score between 7 and 11. Group 2, the chronic/severe group (n 15), consisted of people who were (1) 3 months or more postonset of stroke and (2) had a combined initial CMSA arm and hand stage of 5 or less. Written informed consent was obtained from the participants or their designated caregivers. Treating occupational and physical therapists were trained to 85% accuracy in the administration and scoring of the CAHAI and the CMSA in a half-day workshop proven to be effective. 28 Two research therapists familiarized themselves with the administration of the ARAT and developed written instructions. The treating therapists administered the CMSA and the CAHAI at baseline and 2 to 6 weeks later, depending on the clinical pathways for survivors with mild, moderate, and severe impairments. The research therapists also administered the ARAT at baseline and follow-up. Both treating and research therapists

3 1618 CHEDOKE ARM AND HAND ACTIVITY INVENTORY, Barreca Time Points CAHAI (range, 13 91) Table 2: Descriptive Statistics ARAT (range, 0 56) CMSA Arm-Hand Sum (range, 2 14) CMSA Shoulder Pain (range, 1 7) Initial values Acute (n 24) Chronic (n 15) Follow-up values Acute (n 24) Chronic (n 15) Difference (follow-up initial) Acute (n 24) Chronic (n 15) NOTE. Values are mean SD. were blinded to each other s assessment results. A random number table 29 was used to avoid an order effect in the administration of the CAHAI and the ARAT. To determine interrater reliability, a treating therapist (rater 1) and a research therapist (rater 2) administered the CAHAI within 36 hours; the ordering of the therapists administrating the CAHAI was balanced. We chose the time period of 36 hours to minimize confounding variability due to change within the participant. Because this was not an intervention study, participants received their normal rehabilitation programs as determined by their therapists. Validity Because no criterion standard exists for upper-extremity functional status, a construct validation process was applied. 30 Construct validation involves formulating theories about the attribute of interest and then testing the extent to which the measure provides results consistent with the theories. Our theory was that measures designed to assess a similar attribute should correlate more highly than measures intended to assess different attributes. Specifically, we theorized that the CAHAI should correlate more highly with the CMSA arm-hand sum and with the ARAT than with the CMSA shoulder pain score. The former aspect of this theory is referred to as convergent cross-sectional construct validity, and the contrast in correlations between the CAHAI s correlations with similar measures compared with the CMSA shoulder pain score is known as discriminant validity. 30 Sensitivity to Change (Longitudinal Validity) We also applied a construct validation process to examine the CAHAI s ability to detect change. First, we theorized that the CAHAI change scores should correlate more highly with the CMSA and ARAT change scores (ie, convergent construct longitudinal validity) than with the CMSA shoulder pain score. A known group design 17 was applied to test the hypothesis that the CAHAI was more adept than the CMSA and the ARAT at distinguishing change in patients with acute, mild/moderate impairment presentations from patients with chronic, severe impairment presentations. Analysis Descriptive analysis. The means and standard deviations (SDs) were calculated for all measures. Reliability. We applied a 3-way analysis of variance (ANOVA) with main effects only to estimate the reliability as expressed by an intraclass correlation coefficient (ICC). The factors were patients, raters (2 levels: treating therapist, research therapist), and occasions (2 levels: day 1, day 2). The ICC was calculated as the patient variance divided by the sum of the patient, rater, occasion, and error variances. 30 The standard error (SE) of measurement was calculated as an estimate of the CAHAI s reliability expressed in the same units as the original measurement. 31,32 Confidence in a patient s score at the 90% level was estimated by multiplying the SE of measurement by the z value associated with a 2-tailed 90% confidence level (ie, z 1.65). Minimal detectable change (MDC) at the 90% level was estimated as follows 33 : SE of measurement Validity. We assessed cross-sectional construct validity by correlating the CAHAI s scores with the CMSA s and the ARAT s scores. Discriminant validity was assessed by contrasting the previously mentioned correlations with the correlation between the CAHAI and CMSA shoulder pain score. Because the observations for all measures were performed on the same sample of patients, the data were considered to be dependent. Accordingly, Meng et al s method 34 for comparing dependent correlation coefficients was applied. Sensitivity to change (longitudinal validity). We performed 2 analyses to assess sensitivity to change. The first was identical to the correlation method described above except that change scores were assessed. The second analysis was used to test the hypothesis that the CAHAI would be more adept than the arm and hand impairment stages of the CMSA and the ARAT at distinguishing change in survivors of stroke with acute presentations (mild to moderate motor impairments) from those patients with chronic presentation (severe motor impairments). Receiver operating characteristic (ROC) curve analyses were performed, and the areas under the curves were compared using the method for dependent data described by Hanley and McNeil. 35 RESULTS Descriptive Analysis Table 2 provides a summary of the descriptive analysis. Reliability The 3-way ANOVA produced an ICC of.98 for the CAHAI. Because the rater variance was zero, this factor was removed and the analysis was repeated using a 2-way ANOVA. The ICC 2,1 36 was.98 (95% confidence interval [CI],.96.99). The SE of measurement was 2.8 CAHAI points (95% CI, ). The 90% CI for a patient s score at a single point in time was 4.6 CAHAI points and the MDC 90 was 6.3 CAHAI points. MDC 90 means that 90% of stable patients will display random

4 CHEDOKE ARM AND HAND ACTIVITY INVENTORY, Barreca 1619 Table 3: Cross-Sectional Validity Correlation Coefficients (N 39) Measure CAHAI CMSA Arm-Hand ARAT CMSA Shoulder Pain CAHAI.81* (.66.90).93 (.87.96).47 (.18.68) CMSA arm-hand sum.89 (.80.94).87 (.76.93).55 (.28.74) ARAT.93 (.87.96).92 (.85.96).52 (.24.72) CMSA shoulder pain.39 (.08.63).43 (.13.66).40 (.10.64) *Upper right diagonal: initial assessment correlations and 95% CI. Lower left diagonal: follow-up visit correlations and 95% CI. fluctuations of 6.3 CAHAI points or less when assessed on 2 occasions. Validity Table 3 shows the cross-sectional correlations among measures for all participants (n 39). The values above the diagonal provide the correlations at the initial assessment and those below the diagonal report the correlations at follow-up. Higher correlations were obtained between the CAHAI and the CMSA and ARAT scores than between the CAHAI and the CMSA shoulder pain scores (1-sided, P.001). This supports both the convergent and discriminant cross-sectional validity of the CAHAI. Sensitivity to Change The correlations between the CAHAI and CMSA arm-hand sum and between the CAHAI and ARAT change scores (table 4) were substantially greater than the CAHAI s correlation with the CMSA shoulder pain change score (1-sided, P.001). This supports the convergent longitudinal validity and the longitudinal discriminant validity of the CAHAI. ROC curve analysis was performed to examine each measure s ability to distinguish between the 2 groups of individuals who were theorized to change by different amounts, and to test whether the CAHAI was more adept than the CMSA and the ARAT at distinguishing between these 2 groups. ROC curve analysis quantifies a measure s ability to distinguish between groups as an area under the ROC curve. An area of 1.0 indicates the measure is perfect at discriminating between people in the 2 groups, and an area of 0.5 indicates the measure does no better than chance alone. The areas under the ROC curves were as follows: CAHAI,.95 (95% CI, ); CMSA arm-hand,.76 (95% CI,.61.92); and ARAT,.88 (95% CI, ). Statistically significant differences were observed between the CAHAI and CMSA arm-hand sum (z 2.23; 1-sided, P.013) and between the CAHAI and ARAT (z 1.76; 1-sided, P.039). This finding supports the alternate hypothesis that the CAHAI is more adept at distinguishing change than the CMSA arm-hand sum and the ARAT. Table 4: Sensitivity to Change Correlation Coefficients Change Scores CAHAI Change CAHAI CMSA arm-hand sum.52 (.24.72) ARAT.86 (.75.92) CMSA shoulder pain.29 (.55 to.03) DISCUSSION Before the development of the CAHAI, outcome measures of upper-extremity function often incorporated unilateral nonfunctional tasks, lacked theoretic constructs and interval scaling, or were not validated solely for the stroke population. The CAHAI was conceived to overcome these shortcomings by combining survivors preferences with items generated from the literature and experienced clinicians into a meaningful and relevant evaluation of upper-limb function. This study provides ongoing validation of the CAHAI, a new measure for assessing functional upper-limb recovery, by specifically investigating interrater and test-retest reliabilities, cross-sectional validity, and sensitivity to change. Because we wanted to test the reliability of the administration of the CAHAI, we did not videotape the initial assessment. Rather than restricting the sources of variance for the interrater reliability component of this study to raters and eliminating within patient variation by having all raters view the same performance by each patient, we elected to implement a design that included real-world sources of variation. Accordingly, treating therapists and research therapists assessed patients independently. Because different combinations of treating and research therapists assessed different patients, the interrater variance reflects the extent to which treatment therapists systematically differed from research therapists. That the variance estimate was zero indicates there was no systematic difference between treatment and research therapists. Random variation between raters is included in the residual error term. The relatively small variance associated with occasions (1.4) indicates that patients did not change significantly over the testretest portion of this study. Moreover, support for the CAHAI s reliability is found in the magnitudes of the ICC 2,1 (.98) and the SE of measurement (2.8 CAHAI points). Given the study design, these values represent an estimate of reliability applicable to a clinical situation where one therapist assesses the patient on one day and a second therapist assesses the patient at a subsequent point in time. Because intrarater reliability will always be greater than or equal to interrater reliability interrater reliability cannot exceed the square root of intrarater reliability the estimate of reliability presented in this article represents a conservative estimate for the clinical situation where a patient is always assessed by the same therapist. In summary, the establishment of the CAHAI s high reliability with 2 different raters administering the measure within 36 hours benefits researchers and program evaluators for whom the same therapist may not be available to perform the follow-up assessment. A second goal of our study was to provide further support for the validity of the CAHAI. This was accomplished by investigating the measure s convergent cross-sectional validity and cross-sectional discriminant validity. We hypothesized that survivors with less impairment that is, higher CMSA stages would be more likely to have increased functional use in their paretic upper limbs. Table 3 supports this strong association, showing high CAHAI and ARAT score correlations with the CMSA combined arm and hand

5 1620 CHEDOKE ARM AND HAND ACTIVITY INVENTORY, Barreca impairment scores. As expected, there was also a high correlation between the ARAT and the CAHAI (r.93), which supports convergent cross-sectional validity. Furthermore, both measures showed lower correlation with the CMSA shoulder pain scale (CAHAI.47, ARAT.52), and this provided evidence of cross-sectional discriminant validity. By its high correlation with the ARAT, the CAHAI shows that it is measuring what is intended that is, upperlimb function, the principal outcome of interest for therapeutic interventions. However, because of its close correlation with the ARAT, one may question introducing a new measure. The CAHAI, grounded in measurement theory, was carefully developed specifically for the stroke population. By incorporating upperlimb tasks that survivors of stroke deemed meaningful, the CAHAI shows its apparent ecologic validity and relevance to daily life tasks and considers other theoretical constructs, such as normative upper-limb movements of manipulation, reach and grasp, non gender-specific tasks, degree of motor recovery, and bilateral tasks. 13 This type of evaluation contributes to an important clinical process that establishes a baseline, guides treatment decisions, and gives clients and their families a realistic estimate of functional outcome. Therefore, it is important that the measure be responsive to clinically important changes in arm and hand function. The third component of our study was to compare the CAHAI s sensitivity to change with that of several competing measures. This was accomplished by applying a known group construct validation process. Our theory was that acute survivors with mild to moderate motor involvement would show greater change over the study period than survivors with longer-standing, more severe involvement. It should be stressed that the goal of this aspect of the study was to evaluate the CAHAI s ability to detect change and was not to quantify the minimal clinically important difference. Accordingly, our analysis focused on the ROC curve area and not on the cutoff point that best discriminated between the 2 groups, because there was no reason to believe that this represented the minimal important difference. In contrast, van der Lee et al 19 used the reliability change index test that only considers stable patients. Ten percent of the total ARAT score was arbitrarily chosen, with van der Lee using the results of her reliability study to confirm that a change of 5.7 points was an indicator of what she referred to as minimally clinically important difference. 19 However, the difference van der Lee found was, in fact, the amount of minimal detectable change that can be detected under a given set of assumptions. Instead, we preferred an approach that is applicable to the clinical setting where change in functional status may be examined in survivors who were stable and those who changed over time. This methodology is not associated with additional effort or cost but yields greater information. 37,38 The MDC 90 of the CAHAI was 6.3 points in stable patients approximately 7% of the scale range. Not only did our study support the longitudinal concurrent and discriminant validity of this new measure, but more importantly, the CAHAI proved more adept than the ARAT in distinguishing change in upper-limb function. Scales, such as the CAHAI, that are able to detect clinically important change have 2 potential advantages: (1) improved efficacy of clinical trials (by reducing sample size requirements) and (2) theoretically, a more valid evaluation of therapeutic effects. 37,39,40 At a time when there is constant pressure to justify health costs and demonstrate treatment efficacy, the introduction of a new stroke functional outcome measure that is more sensitive to clinically important change should be welcomed. The CAHAI complements the CMSA and offers researchers and clinicians a comprehensive package for assessing motor and functional recovery in patients after a stroke. CONCLUSIONS A priority of future research of the paretic arm and hand will be to ascertain effective treatment interventions and to identify those survivors of stroke who have the potential to change and benefit from intensive therapy. The success of such research will be aided greatly by a valid measure that is most sensitive to clinically important change. As well, a critical step in convincing clinicians of the need to incorporate patient preferences into a measure is to document the advantages of such a measure in contrast to the traditional approaches for evaluating therapeutic interventions for the paretic upper limb. By directly comparing 3 scales in the same patient population, we found that the CAHAI permitted a more accurate and relevant measure of clinical change. APPENDIX 1: CHEDOKE ARM AND HAND ACTIVITY INVENTORY SCORE FORM 1. Total assist (weak U/L 25%) 2. Maximal assist (weak U/L 25%-49%) 3. Moderate assist (weak U/L 50%-74%) 4. Minimal assist (weak U/L 75%) Activity Scale 5. Supervision 6. Modified independence (device) 7. Complete independence (timely, safely) Affected Limb: 1. Open jar of coffee holds jar holds lid Score 2. Call 911 holds receiver dials phone 3. Draw a line with a ruler holds ruler holds pen

6 CHEDOKE ARM AND HAND ACTIVITY INVENTORY, Barreca 1621 Affected Limb: 4. Put toothpaste on toothbrush holds toothpaste holds brush Score 5. Cut medium consistency putty holds knife holds fork 6. Pour a glass of water holds glass holds pitcher 7. Wring out washcloth 8. Clean a pair of eyeglasses holds glasses wipes lenses 9. Zip up the zipper holds zipper holds zipper pull 10. Do up five buttons 11. Dry back with towel reachs for towel grasps towel end 12. Place container on table 13. Carry bag up the stairs Total Score NOTE. For a copy of the administration guidelines, contact the author. Abbreviation: U/L, upper limb. References 1. Flick CL. Stroke rehabilitation. 4. Stroke outcome and psychosocial consequences. Arch Phys Med Rehabil 1999;80(5 Suppl 1):S Heart and Stroke Foundation of Ontario. Management of the post stroke arm and hand: treatment recommendations of the 2001 consensus panel. Toronto: HSFOM; Available at: /ClientImages/1/PostStrokeArmandHandFinal2002.pdf. Accessed March 17, Broeks JG, Lankhorst GJ, Rumping K, Prevo AJ. The long-term outcome of arm function after stroke. Disabil Rehabil 1999;21: Dombovy ML. Rehabilitation and the course of recovery after stroke. In: Whisnant JP, editor. Stroke: populations, cohorts, and clinical trials. Boston: Butterworth-Heinemann; Gowland C. Recovery of motor function following stroke: profile and predictors. Physiother Can 1982;34: McDowell I, Newell C. The theoretical and technical foundations of health measurement. In: McDowell I, Newell C, editors. Measuring health: a guide to rating scales and questionnaires. New York: Univ Pr; p Okkema KA, Culler KH. Functional evaluation of upper extremity use following stroke: a literature review. Top Stroke Rehabil 1998;4: Rudman D, Hannah S. An instrument evaluation framework: description and application to assessments of hand function. J Hand Ther 1998;11: Cronbach LJ. Test validation. In: Thorndike RL, editor. Educational measurement. Washington (DC): American Council on Education; p Lyle RC. A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res 1981;4: Carroll D. A quantitative test of upper extremity function. J Chronic Dis 1965;18: World Health Organization. International classification of impairment, disability and handicap. Geneva: WHO; Barreca S, Gowland C, Stratford P, et al. Development of the Chedoke Arm and Hand Activity Inventory: theoretical constructs, item generation, and selection. Top Stroke Rehabil 2004; 11: Barreca S, Gowland C, Stratford P, et al. Development of the Chedoke-McMaster Arm and Hand Activity Inventory: item selection. Physiother Can 1999;29: Gowland C, Van Hullenaar S, Torresin W, et al. Chedoke- McMaster Stroke Assessment: development, validation and administration manual. Hamilton: Chedoke-McMaster Hospitals and McMaster Univ; Gowland C, Stratford P, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke 1993;24: Stratford PW, Binkley JM, Riddle DL. Health status measures: strategies and analytic methods for assessing change scores. Phys Ther 1996;76: Uniform Data System for Medical Rehabilitation. Functional Independence Measure (the FIM instrument). Guide for the Uniform Data Set for medical rehabilitation (including the FIM instrument). Buffalo: State Univ New York; p III-1 to van der Lee JH, De Groot V, Beckerman H, Wagenaar RC, Lankhorst GJ, Bouter LM. The intra- and interrater reliability of the Action Research Arm test: a practical test of upper extremity function in patients with stroke. Arch Phys Med Rehabil 2001; 82: Brunnstrom S. Movement therapy in hemiplegia: a neurophysiological approach. New York: Harper & Row; Gowland C. Predicting sensorimotor recovery following stroke rehabilitation. Physiother Can 1984;36: Moskowitz E, Lightbody FE, Freitag NS. Long-term follow-up of the poststroke patient. Arch Phys Med Rehabil 1972;53: Sunderland A, Tinson D, Bradley L, Langton-Hewer R. Function after stroke: an evaluation of grip strength as a measure of recov-

7 1622 CHEDOKE ARM AND HAND ACTIVITY INVENTORY, Barreca ery and a prognostic indicator. J Neurol Neurosurg Psychiatry 1989;52: Wade DT, Langton-Hewer R, Wood VA, Skilbeck CE, Ismail HM. The hemiplegic arm after stroke: measurement and recovery. J Neurol Neurosurg Psychiatry 1983;46: Olsen TK. Improvement in function and motor impairment after stroke. J Neurol Rehabil 1989;3: Katrak P, Bowring G, Conroy P, Chilvers M, Poulos R, McNeil D. Predicting upper limb recovery after stroke: the place of early shoulder and hand movement. Arch Phys Med Rehabil 1998;79: Feys H, Hetebrij J, Wilms G, Dom R, de Weerdt W. Predicting arm recovery following stroke: value of site lesion. Acta Neurol Scand 2000;102: Miller P, Stratford P, Gowland C, Van Hullenar S, Torresin W. Comparing two methods to train therapists to use the Chedoke- McMaster Stroke Assessment [abstract]. Presented at: WCPT International Congress; May 25, 1999; Yokohama (Japan). 29. May R, Masson M, Hunter M. Analysis of variance: repeated measures. Application of statistics in behavioral research. New York: Harper & Row; p Streiner DL, Norman GR. Health measurement scales: a practical guide to their development and use. 3rd ed. New York: Oxford Univ Pr; Stratford PW. Getting more from the literature: estimating the standard error for measurement from reliability studies. Physiother Can 2004;56: Stratford PW, Goldsmith CH. Use of the standard error as a reliability index of interest: an applied example using elbow flexor strength data. Phys Ther 1997;77: Beaton DE, Bombardier C, Katz JN, Wright JG. A taxonomy for responsiveness. J Clin Epidemiol 2001;54: Meng X, Rosenthal R, Rubin DB. Comparing correlated correlation coefficients. Psychol Bull 1992;111: Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148: Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979;86: Stratford P, Spadoni G, Kennedy D, Westaway MD, Alcock GK. Seven points to consider when investigating a measure s ability to detect change. Physiother Can 2002;54: Stratford P, Riddle DL, Binkley JM. Assessing for changes in a patient status: a review of current methods and a proposal for a new method of estimating true change. Physiother Can 2001;53: Liang MH, Fossel AH, Larson MG. Comparisons of five health status instruments for orthopedic evaluation. Med Care 1990; 28: Liang MH. Longitudinal construct validity: establishment of clinical meaning in patient evaluative instruments. Med Care 2000; 38(9 Suppl):II84-90.