PERSISTENT LOSS OF upper-extremity (UE) motor function

660 ORIGINAL ARTICLE An Evaluation of the Wolf Motor Function Test in Motor Trials Early After Stroke Dorothy F. Edwards, PhD, Catherine E. Lang, PT, PhD, Joanne M. Wagner, PT, PhD, Rebecca Birkenmeier, OTD, OTR/L, Alexander W. Dromerick, MD ABSTRACT. Edwards DF, Lang CE, Wagner JM, Birkenmeier R, Dromerick AW. An evaluation of the Wolf Motor Function Test in motor trials early after stroke. Arch Phys Med Rehabil 2012;93:660-8. Objective: To examine the internal consistency, validity, responsiveness, and advantages of the Wolf Motor Function Test (WMFT) and compare these results to the Action Research Arm Test (ARAT) in participants with mild to moderate hemiparesis within the first few months after stroke. Design: Data were collected as part of the Very Early Constraint-Induced Therapy for Recovery from Stroke (VEC- TORS) trial, an acute, single-blind randomized controlled trial of constraint-induced movement therapy. Subjects were studied at baseline (day 0), after treatment (day 14), and after 90 days (day 90) poststroke. Setting: Inpatient rehabilitation hospital; follow-up 3 months poststroke. Participants: Hemiparetic subjects (N 51) enrolled in the VECTORS trial. Intervention: None. Main Outcome Measures: At each time point, subjects were tested on (1) the WMFT and ARAT, (2) clinical measures of sensorimotor impairments, (3) reach and grasp movements performed in the kinematics laboratory, and (4) clinical measures of disability. Blinded raters performed all evaluations. Analyses at each time point included calculating effect size as indicators of responsiveness, and correlation analyses to examine relationships between WMFT scores and other measures. Results: The WMFT is internally consistent, valid, and responsive in the early stages of stroke recovery. Sensorimotor and kinematic measures of reach and grasp support the construct validity of the WMFT. Conclusions: In an acute stroke population, the WMFT has acceptable reliability, validity, and responsiveness to change over time. However, when compared with the ARAT, the From the Departments of Kinesiology-Occupational Therapy and Neurology, University of Wisconsin Madison, Madison, WI (Edwards); Programs in Occupational Therapy (Edwards, Lang, Birkenmeier, Dromerick) and Physical Therapy (Lang, Dromerick) and Department of Neurology (Lang), Washington University School of Medicine, St. Louis, MO; Department of Physical Therapy and Athletic Training, Doisy College of Health Sciences, St. Louis, MO (Wagner); Departments of Rehabilitation Medicine and Neurology, Georgetown University, Washington, DC (Dromerick); National Rehabilitation Hospital, Washington, DC (Dromerick); and Washington DC Veterans Affairs Medical Center, Washington, DC (Dromerick). Supported by the National Institutes of Health (grant nos. NS41261, HD047669), the James S. McDonnell Foundation (grant no. 21002032), and the Foundation for Physical Therapy. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Reprint requests to Dorothy F. Edwards, PhD, Program in Occupational Therapy, University of Wisconsin Madison, 2170 Medical Science Center, 1300 University Ave, Madison, WI 53706-1532, e-mail: dfedwards@education.wisc.edu. In-press corrected proof published online on Feb 15, 2012, at www.archives-pmr.org. 0003-9993/12/9304-00810$36.00/0 doi:10.1016/j.apmr.2011.10.005 higher training and testing burdens may not be offset by the relatively small psychometric advantages. Key Words: Outcome assessment (health care); Paresis; Rehabilitation; Treatment outcome. 2012 by the American Congress of Rehabilitation Medicine PERSISTENT LOSS OF upper-extremity (UE) motor function is found in 45% of all stroke survivors, and contributes substantially to stroke-related disability. 1 Although there are many scales developed to assess UE function after stroke, there is no consensus regarding which measure is best suited for routine assessment of patients across the stages of recovery. 2,3 The most common standardized measures used in UE treatment studies are the Fugl-Meyer Assessment (FMA), 4 the Action Research Arm Test (ARAT), 5 and the Wolf Motor Function Test (WMFT). 6 The majority of research examining these measures has been conducted in patients in subacute or chronic stages of stroke recovery. 7-9 Although several studies have examined the reliability and validity of the ARAT and FMA in acute stroke patients, 2,10 the psychometric characteristics of the WMFT in patients in the acute stage of stroke recovery have not been extensively studied. 11 The WMFT is widely used in UE treatment studies of persons in the subacute and chronic stages of stroke recovery. 9,12-14 This scale consists of a series of tasks sequenced according to the joints involved and level of difficulty. 15 Two items measure strength; the remaining 15 items are rated on the speed of performance and quality of motor ability. The WMFT was developed specifically for the assessment of the effects of constraint-induced movement therapy (CIMT). 14 Since that time the WMFT has been used in more than 20 studies. 9 Interrater reliability, internal consistency, content, construct validity, and responsiveness to change have been established for stroke patients in the subacute and chronic stages of recovery. 6,15,16 The psychometric properties of the scale have not been reported in acute populations studied prior to the 90-day time point used to define subacute stroke. The purpose of this study is to examine the responsiveness and validity of the WMFT in a population of hemiparetic subjects within the first weeks and months after stroke. We ARAT CIMT FA FMA NIHSS UE VECTORS WMFT List of Abbreviations Action Research Arm Test constraint-induced movement therapy functional ability Fugl-Meyer Assessment National Institutes of Health Stroke Scale upper extremity Very Early Constraint-Induced Therapy for Recovery from Stroke Wolf Motor Function Test

RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards 661 asked (1) how well do the test items measure the same construct of UE function (internal consistency), (2) how responsive is the WMFT to change in the first weeks after stroke and in the first months after stroke, (3) what is the relationship between the WMFT and the ARAT (concurrent validity), and (4) how do WMFT scores relate to sensorimotor impairment measures typically measured by rehabilitation professionals, objective measures of movement quality obtained via kinematic analyses, and disability scores as measured by the FIM (construct validity). The sensorimotor impairment, kinematic, and disability data collected as part of the Very Early Constraint-Induced Therapy for Recovery from Stroke (VEC- TORS) trial provided the information needed to assess construct validity of the WMFT. We used both clinimetric and psychometric approaches. Although there is considerable overlap between clinimetric and psychometric methods, clinimetric analyses place greater stress on issues such as sensitivity or responsiveness to clinically relevant change, as well as on clinical utility. 17,18 Clinical utility refers to ease and efficiency of use, burden on the patient, and meaningfulness of the information that it provides. METHODS Participants All study participants were enrolled in the VECTORS study. VECTORS is a single-center randomized controlled trial of the early application of CIMT; the purpose of VECTORS is to gather the information necessary to design a definitive multicenter trial of CIMT administered in the immediate poststroke period. Study and treatment procedures are similar to those described elsewhere. 19 VECTORS used a single-blind randomized controlled design. Subjects were randomized to 1 of 3 groups: (1) 2 h/d of conventional treatment, (2) 2 h/d of shaping treatment plus 6 hours of constraint, and (3) 3 h/d of shaping plus constraint 90% of waking hours. The study protocol was approved by the Human Studies Committee at Washington University, and all subjects provided informed consent prior to participation. Persons admitted to the acute neurology stroke service at Barnes-Jewish Hospital in St. Louis, MO, were screened for study eligibility; 1117 individuals were screened for this study. Subjects were included if they had (1) an ischemic or hemorrhagic stroke (with confirmatory neuroimaging) within 28 days of admission to inpatient rehabilitation; (2) persistent hemiparesis, generally as indicated by a score of 1 or 2 on the motor arm item of the National Institutes of Health Stroke Scale (NIHSS); (3) the presence of proximal UE voluntary movement as indicated by a score of 3 or more on the upper arm item of the Motor Assessment Scale (wrist and finger movement were not required); (4) evidence of preserved cognitive function; (5) the ability to follow 2-step commands; and (6) no UE injuries or conditions that limited use prior to the stroke. Subjects were excluded if they (1) could not give informed consent; (2) had clinically significant fluctuations in mental status in the 72 hours prior to enrollment; (3) were not independent prior to the index stroke as measured by scores 95 on the Barthel Index scale or 1 on the Modified Rankin Scale; (3) had hemispatial neglect; and/or (4) were not expected to survive 1 year due to other illnesses (eg, cardiac disease, malignancy). Fifty-one subjects with hemiparesis poststroke are included in this study. Procedures All evaluations were performed by trained study personnel blinded to the treatment type. Prior to collection of baseline data, evaluators were trained to properly administer the WMFT, ARAT, and NIHSS scales. Interrater reliability of more than.95 was established for all measures for all study evaluators. The data in this article include the evaluations performed before randomization (day 0), after treatment (day 14), and at 90 days after stroke onset (day 90). The day 90 time point is the primary study endpoint because most acute stroke intervention trials assess efficacy at 90 days after onset, when the majority of stroke patients are at or near their clinical plateau. 20 Clinical Scale Evaluations Wolf Motor Function Test. The WMFT is a laboratorybased evaluation designed for use in UE stroke rehabilitation studies. 6,15 The Extremity Constraint-Induced Therapy Evaluation trial version of the WMFT is a 17-item measure used to quantify UE functional limitations. 9 It is comprised of 2 strength items and 15 timed task performance items. The task performance items begin with the measurement of simple proximal movements and progress to more complex distal and whole limb movements. The WMFT yields 3 scores: a functional ability (FA) score, which quantifies quality of performance; a timed (time) score, quantifying speed of performance in seconds; and a grip strength (strength) score. The FA score uses a 6-point ordinal scale to rate movement quality on 15 items, where 0 indicates no attempt to use the more affected UE and 5 indicates that movement of the affected UE appears to be normal. The time score is the mean time to complete the same 15 items; scores were truncated at 120 seconds. The strength score is the strength of the grip in kilograms. The test has published reliability and validity in both chronic and subacute populations of stroke patients. 20,21 In the VECTORS study, the key use task was not collected owing to the challenges of constructing and transporting the testing materials; results reported in this study do not include this item. Action Research Arm Test. The ARAT assesses functional limitations of the UE. It uses a 4-point ordinal scale on 19 items, where 0 indicates no movement and 3 indicates normative movement. 5 Item scores are summed to create 4 subscale scores gross motor (9 point maximum), grasp (18 point maximum), grip (12 point maximum), and pinch (18 point maximum) and a total scale score with a maximum score of 57, indicating normative performance. The ARAT has now been shown to be reliable, valid, and responsive to change across a variety of time points poststroke. 7,8,10,22-24 FIM. The FIM is an 18-item global disability measure incorporating concepts and items of functional performance including activities of daily living, bowel and bladder function, social cognition, functional communication, and functional mobility. 25 The FIM is scored on a 7-point ordinal scale, where a score of 1 indicates dependence and a score of 7 indicates independence. In this study, the FONE-FIM interview was administered to the participant as a measure of self-perceived disability. 26 The 15 items requiring motor function were summed to create a FIM activity of daily living, while a subset of 5 items relating to UE use were summed to create a FIM UE score. National Institutes of Health Stroke Scale. The NIHSS assesses cognitive, sensory, and motor impairments as an indicator of stroke severity. 27 This 13-item test results in scores ranging from 0 (no deficit) to 46 (severe deficit). Scores of 6 to 22 are generally considered to be in the moderate range. The NIHSS scores reported here were collected on the acute hospital service.

662 RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards Sensorimotor Impairment Evaluations Various sensorimotor impairments are evaluated by physical and occupational therapists, and impairment findings are used to guide rehabilitation interventions. Here, we evaluated 4 impairments commonly assessed by rehabilitation team members. Light touch sensation was measured at 4 locations on the UE using the Semmes-Weinstein monofilaments. a Normative sensation was scored as 0 (monofilament size 2.83), diminished light touch was scored as 1 (monofilament size 3.61), diminished protective sensation was scored as 2 (monofilament size 4.31), loss of protective sensation was scored as 3 (monofilament size 6.65), and unable to feel largest monofilament was scored as 4. Monofilament values were converted to an ordinal scale and were averaged across the 4 locations to create a composite light touch score. 28,29 Thus, UE somatosensory impairment scores ranged from 0 to 4, with 0 indicating normal and 4 indicating complete loss. Elbow joint spasticity was measured using the Modified Ashworth Scale. 30 Pain in the affected shoulder was measured using a visual analog scale (0 10). Strength of the flexors and extensors of the UE were measured using a hand-held dynamometer b following the protocol of Andrews et al, 31 except that subjects were seated during testing. Strength values were expressed as a ratio of the affected to nonaffected side. A composite UE strength score was determined by averaging the strength of the antigravity muscles used during a reaching task: shoulder flexors, elbow flexors, and wrist extensors. 28,29 Kinematic Measurement of UE Movements The purpose of the kinematic testing was to obtain objective and sensitive quantification of UE movement. We tested each subject s ability to perform reach and reach-to-grasp movements of the affected UE, as previously described. 10,28,32 For the analyses included here, we focused on the reach and on the grasp component of the reach-to-grasp movement. These movements were chosen because they are important for functional use of the UE. Briefly, 3-dimensional movements were recorded using a 6-camera video system. b Thirteen reflective markers were placed on the trunk (3), upper arm (3), forearm (3), dorsum of the hand (1), thumb (1), index finger (1), and target (1). Movements were recorded simultaneously from each camera at 60Hz and stored on a computer disk for later analyses. Subjects sat in a straight-backed chair with the trunk strapped to the back in order to minimize compensatory trunk movements. The start position was relaxed with the tested hand resting on a pillow on the ipsilateral thigh. The target was positioned at 90% of arm s length directly in front of the affected shoulder. The target was a 40-mm diameter sphere for the reach movement and a 38-mm diameter soft cylinder for the reach-to-grasp movement. The reach-to-grasp target cylinder had a long axis of 110mm that was oriented parallel to the subject s chest. Subjects were instructed to touch (or grasp) the target as fast as possible. Prior to any recorded movements, subjects were allowed practice trials of each type of movement in order to familiarize themselves with the task and the instructions. Three trials of each movement were recorded. Offline, c EVaRT, c and KinTrak softwares c were used to extract position, velocity, and angular data, respectively, for the UE from video images. Data were lowpass filtered at 6Hz. Variables of interest for the reach were (1) speed, quantified as peak wrist velocity; (2) efficiency, quantified as reach path ratio; and (3) accuracy, quantified as endpoint error. 29 Peak wrist velocity was the maximum resultant velocity of the wrist attained during the movement. Reach path ratio, a measure of how directly the hand moves toward the target, was calculated as the ratio of the length of the path actually traveled to an ideal straight line between the start and end positions. A reach path ratio of 1 represents a direct movement to the target (ideal), and a reach path ratio 1 represents an abnormally curved movement. Endpoint error was the total distance from the index finger to the target at the end of movement (touch). Variables of interest for the grasp component of the reach-to-grasp movement were (1) speed, quantified as peak aperture ratio; (2) efficiency, quantified as aperture path ratio; and (3) peak aperture. 28,32 Peak aperture rate was the maximum aperture rate (how fast the thumb and index finger tips opened/closed) attained during the movement. Aperture path ratio, a measure of how directly the thumb and index fingers closed, was calculated as the ratio of the length of the aperture curve actually traveled to an ideal straight line between the first peak of the aperture trace and the aperture at the end of movement. An aperture path ratio of 1 represents a single, smooth closing of the thumb and index fingers (normal), and an aperture path ratio of 1 represents abnormal closing of the fingers, typically seen when subjects made multiple attempts to open and close the fingers on the target. Peak aperture was the greatest 3-dimensional distance between thumb and index finger tips attained during the movement. Because of the acute nature of our subjects impairment, especially at the day 0 visit, our analyses were structured to capture the range of behaviors (and attempts) recorded, such that our variables of interest quantified performance regardless of whether or not the subjects could touch or grasp the target. Statistical Analyses Statistical analyses were conducted using SPSS for Windows Version 13.0. d Distributions of variables were tested for normality using Kolmogorov-Smirnov tests. Four variables from the kinematic testing were not normally distributed and needed to be transformed for further statistical analyses. The type of transformation done on a variable was chosen by examining the raw distribution and then selecting the transformation that best minimized skew. The 4 variables were transformed as follows: reach efficiency (reach path ratio) using percentile ranks, reach accuracy (endpoint error) using the natural log function, grasp speed (peak aperture rate) using the square root function, and grasp efficiency (aperture path ratio) using percentile ranks. All statistical analyses on these 4 variables were done using the transformed data. Internal consistency was determined using Cronbach alpha. Responsiveness of the WMFT FA, time, and strength scores from day 0 to day 14 and from day 0 to day 90 was determined using the single population effect size method. 33 The responsiveness (effect size) within the first weeks of stroke was calculated as the mean change from day 0 to day 14 divided by the SD at day 0, and the responsiveness within the first months of stroke was calculated as the mean change from day 0 to day 90 divided by the SD at day 0. We used the criteria described by Lin et al 11 to examine floor and ceiling effects. The floor effect is the percentage of the sample scoring the minimum possible points. The ceiling effect reflects the percentage of scores at the highest value of the scale. Pearson product moment correlations were used to determine the concurrent validity of the WMFT and the ARAT at all 3 time points. Pearson product moment correlations were used to examine construct validity via the relationships between WMFT FA, time and strength scores, sensorimotor impairments, kinematic, and disability measures. Analyses were computed separately at each of the 3 time points. Fifty-one subjects were available at the day 0 and day 14 time points, and 40 subjects at the day 90 time point (5 subjects did not return for

RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards 663 Table 1: Characteristics of Subjects Enrolled in the VECTORS Trial Variable Values Age (y) 63.7 13.6 Admission NIHSS score 8.55 4.52 Time to day 0 evaluation (d since stroke) 9.5 4.5 Time to day 14 evaluation (d since stroke) 25.9 10.6 Time to day 90 evaluation (d since stroke) 110.8 20.7 Premorbid Barthel Index score 99.6 2.2 Premorbid Modified Rankin Scale score 0.30 0.60 Sex Men 21 (42) Women 29 (58) Race White 21 (42) Black 28 (56) Other 1 (2) Stroke type Ischemic 39 (78) Hemorrhagic 11 (22) Affected side Dominant 21 (42) Nondominant 29 (58) Values are mean SD or n (%). their day 90 evaluation and 6 subjects were not evaluated in the kinematic laboratory at day 90). Based on our sample sizes, correlation coefficients with an absolute value.40 were statistically significant at the P.01 level, and correlation coefficients with an absolute value.30 were statistically significant at the P.05 level. RESULTS Persons admitted to the acute neurology stroke service at Barnes-Jewish Hospital in St. Louis, MO, were screened for study eligibility. Mean time to randomization SD was 9.5 4.5 days poststroke. Table 1 presents the demographic and clinical characteristics of the VECTORS sample. Subjects were enrolled based on the presence of a moderate degree of UE motor dysfunction and the absence of severe sensory or cognitive impairments. Using these criteria, study participants displayed a wide range of overall stroke severity as measured by the total NIHSS. The mean NIHSS score SD was 8.5 4.5 (range, 2 21). Subjects ranged in age from 39 to 94 years. Sex, race, and stroke type (ischemic vs hemorrhagic) are representative of the stroke patient population at Barnes-Jewish Hospital. Hemorrhagic stroke was defined as having a primary hemorrhage; those judged to have hemorrhagic transformation of an ischemic infarct were classified as having ischemic stroke. The majority of subjects (58%) were affected on their nondominant side. Table 2 presents mean scores on the functional, sensorimotor, and kinematic variables at the 3 time points. As expected, scores on all UE outcome measures improved over time. The mean WMFT FA score at day 0 was 2.44 with a range of 0.14 to 4.54. On average, VECTORS participants had some purposeful movement at baseline, but the task performance was Table 2: Scores on Study Measures by Time of Testing Variable Day 0 Day 14 Day 90 WMFT scores Function score 2.44 1.10 3.62 1.10 4.20 1.07 Time score (s) 42.5 39.8 18.2 28.6 8.5 14.0 Grip score (kg) 9.6 10.5 16.9 12.6 18.5 12.0 Scores at floor (%) 11.9 6 4.4 Scores at ceiling (%) 0 8 37 ARAT scores Total score 22.5 15.2 38.1 16.5 43.7 14.9 Grasp score 7.78 5.39 13.50 5.53 14.31 5.10 Grip score 5.02 3.52 8.62 3.47 9.69 3.24 Pinch score 4.75 5.29 9.46 6.66 12.54 6.03 Scores at floor (%) 5.9 2 2.1 Scores at ceiling (%) 3.9 22 33 Sensorimotor scores UE strength 0.27 0.27 0.48 0.36 0.54 0.27 Pain 0.90 1.62 0.75 1.32 1.60 1.85 Light touch 1.26 1.50 0.98 1.33 0.54 0.83 Spasticity 0.58 0.81 0.96 0.94 1.14 1.23 Kinematic scores Reach Speed (mm/s) 620 348 719 354 759 371 Efficiency 1.69 1.16 1.18 0.30 1.10 0.20 Accuracy (mm) 119 86 78 83 61 70 Grasp Speed (mm/s) 116 91 166 122 179 124 Efficiency 2.48 2.62 1.93 2.16 1.46 1.38 Peak aperture 94 33 114 26 120 25 (mm) FIM motor score 57.79 11.08 75.04 11.48 80.67 12.40 FIM UE score 23.08 4.95 29.29 4.99 31.81 4.67 NOTE. Values are mean SD or as otherwise indicated.

664 RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards typically quite effortful. Table 2 also presents the floor and ceiling effects for the WMFT and ARAT at each time of evaluation. At baseline, 12% of the participants had scores of 1 on the WMFT and none had a score of 59, which is the maximum score, suggesting that the scale does not have significant floor or ceiling effects for assessment of patients with mild to moderate stroke in the acute stage of recovery. By day 14, the mean FA score increased by 1.18 points. Eight percent of the sample achieved mean scores of 5. The FA score change between day 14 and day 90 was.58, suggesting that the greatest improvement was observed on this scale between day 0 and day 14. Thirty-seven percent of the sample had FA scores of 5 (normative movement) at day 90. Time and strength scores are also reported in table 2. The pattern of results for these variables is similar to that of the FA scores. The greatest improvement occurred between day 0 and day 14, although there were still significant improvements (P.001) in both time and strength scores between day 14 and day 90. Internal Consistency Internal consistency of the WMFT was assessed at all time points. Cronbach alpha coefficients for the FA scale were.96,.97, and.98 at day 0, day 14, and day 90, respectively. Similar coefficients were observed for the time scale. Coefficients of.90 or greater are indicative of high levels of internal consistency. 34 Responsiveness The responsiveness of the WMFT was also examined. The single population effect size method 33 was used to compute effect size coefficients as indicators of scale responsiveness. Based on Cohen criteria, coefficients of.80 or greater are considered large,.50 to.80 moderate, and.50 small. 35 Using these criteria the responsiveness of the WMFT scores ranged from moderate to large (table 3). These coefficients varied by both test and time interval. Overall, the FA scores were more responsive than the time and strength scores across the 3 times of measurement. The responsiveness of the WMFT FA score increased from 1.09 (day 0 day 14) to 1.63 (day 0 day 90), suggesting that the FA score was the most sensitive indicator of clinical change in the acute stage of stroke recovery. Although the responsiveness of the time and strength scores also improved with time, the effects were not as robust. Coefficients for time and strength ranged from moderate (.61 and.69, respectively) for the interval between day 0 to day 14 to good (.85 and.84) between day 0 and day 90. Concurrent Validity Concurrent validity of the WMFT in this population was examined through comparison of the 3 WMFT scores with the total ARAT score at each time point (table 4). The highest Table 3: Responsiveness of WMFT and ARAT Scores as Measured by the Single Population, Effect Size Method Variable Day 0 to Day 14 Day 0 to Day 90 WMFT function score 1.093 1.630 WMFT time score 0.612 0.854 WMFT grip score 0.694 0.848 ARAT total score 1.018 1.390 ARAT gross motor score 0.729 0.984 ARAT grasp score 1.042 1.224 ARAT grip score 1.017 1.324 ARAT pinch score 0.854 1.494 Table 4: Correlations Between the WMFT FA Scores and the ARAT Total Score at the 3 Time Points Variable Day 0 Day 14 Day 90 WMFT FA function score 0.745 0.827 0.863 WMFT FA time score 0.641 0.825 0.772 WMFT FA grip score 0.702 0.631 0.553 correlations were observed between the FA and the ARAT total scores. The correlations increased from.75 at day 0 to.87 at day 90. Time and ARAT total score coefficients ranged from.64 at day 0 to.77 at day 90. The correlations between the ARAT and strength scores decreased from.70 at day 0 to.55 at day 90. These results support the concurrent validity of the WMFT. Construct Validity Construct validity was examined through comparison of WMFT scores to measures of sensorimotor status, kinematic assessments of reach and grasp, and disability as measured by the FIM. Relationships between the FA, time, and strength scores and 4 sensorimotor impairments frequently measured by occupational and physical therapists are shown in figure 1. UE strength was moderately to well correlated with FA, time, and strength across all 3 time points, such that the greater the strength, the higher the FA and strength scores and the faster the time score. Spasticity was inversely related to FA, time, and strength scores, such that greater spasticity was associated with lower FA and strength scores and slower time scores. Light touch sensation was essentially unrelated to all 3 scores over time. Pain appeared to have greater effects on performance at day 14 particularly for strength, but in general, pain had minimal influence on WMFT scores in this sample. Relationships between WMFT scores and kinematic measures of movement are shown in figure 2. Kinematic measures of reach (see fig 2, left column) were related to FA scores such that faster, more efficient, and more accurate reaching performance was associated with better FA scores over time. The time scores were similarly associated with the kinematic measures of reach, such that faster, more efficient, and more accurate reaching performance was associated with faster scores. The correlations between the strength scores and the 3 reach measures were consistent across the 3 time points with coefficients in the moderate range. Examination of the WMFT scores and kinematic measures of grasp (see fig 2, right column) indicate a similar pattern of relationships between the FA scores and grasp. Faster, more efficient grasping performance and a greater ability to open the fingers were associated with higher FA scores. Time and strength correlation coefficients remained fairly consistent across the 3 time points for all grasp variables. As expected, we observed moderate correlations between the kinematic measures and WMFT scores. This was expected because the WMFT assesses the broad construct of UE function, whereas the kinematic measures capture performance on only 2 movements (reach, grasp) essential for UE function. Relationships between WMFT scores and FIM scores are shown in figure 3. Across the 3 WMFT scores, correlation coefficients ranged from moderate (.40) to high (.70) and increased in magnitude from day 0 to day 90. In general, the correlations among WMFT scores and the FIM motor and UE scores were higher for FA and strength scores than for time scores across the 3 assessment intervals.

RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards 665 Fig 1. Pearson correlations between WMFT FA scores (A), time scores (B), and grip strength scores (C) with sensorimotor impairment measures at the 3 time points. Note that the sign of the correlations are reversed for the time scores because better time scores are indicated by smaller numbers. Abbreviations: d0, day 0; d14, day 14; d90, day 90. DISCUSSION Previous articles have demonstrated the reliability of the WMFT in chronic and subacute populations. 9,15,36 The present study expands on earlier findings by examining the internal consistency, responsiveness, and validity of the WMFT in the acute phase of stroke recovery. Our results indicate that the WMFT has a high degree of internal consistency, is responsive to change, and has high concurrent validity with the ARAT in the first few weeks and months after stroke. Our data suggest that the WMFT can be used immediately after stroke onset for patients with mild to moderate stroke. Only 12% of our participants were unable to perform scale items at baseline. We also found lower rates of floor effects and higher rates of ceiling effects for both the WMFT and ARAT at days 14 and 90 than have been previously reported. 11 These differences may reflect the effects of study inclusion criteria requiring proximal arm movement at the time of enrollment. Thus, our findings support the use of the WMFT in acute assessment of individuals with mild to moderate UE hemiparesis early after stroke onset. Furthermore, by examining the relationship of WMFT scores to kinematic measures of grasp and reach and sensorimotor function, our results extend the evidence for the construct validity of the WMFT beyond the findings presented in other studies. We have previously reported convergent validity of the WMFT using 2 disability measures (the FIM and the Motor Activity Log) obtained from VECTORS participants 90 days after stroke. 37 In the present study, UE strength and spasticity were more highly correlated with the FA and time scores than with light touch or pain. As expected, UE strength was also highly correlated with the WMFT strength scores, and low to moderate correlations were observed between strength and light touch and pain. The moderate correlations reflect differences in the domains of measurement associated with each approach. Kinematic analysis allows for highly quantitative measures of specific aspects of movement beyond that which can be done with the more global scoring used by the FA scale. The moderate relationships between the WMFT scores and the kinematic measures of efficiency, speed, and accuracy during reaching and grasping indicate that despite differences in the scaling of the measures, the WMFT and kinematic scores are assessing related characteristics of UE function. 28,32 The timing of the assessments and the relatively small sample assessed in this study also affect the magnitude of the relationships among the kinematic and WMFT scores. Relationships between the WMFT and other measures may be more variable in this early time period after stroke than in the chronic stage of stroke recovery. Jørgenson et al 20 demonstrated that improvements in function tend to lag behind improvements in neurologic impairments by 1 to 2 weeks; the small fluctuations in correlation coefficients observed here may reflect a slight shifting in the magnitude of the relationships over the first few weeks and months poststroke. Another possible explanation for the small fluctuations is that they represent slightly different estimations of the true value of the relationship in the population. A much larger sample size in which we could calculate narrow confidence intervals for the correlation coefficients would be required to distinguish between these 2 possibilities. The WMFT compares well with the ARAT in this population. We published an assessment of the sensitivity of the ARAT in the same participants. 10 In the same sample, the FA scale was the most responsive, followed by the ARAT total score, and then the WMFT time and strength scores. The responsiveness from day 0 to day 90 for both the WMFT FA scale and the ARAT exceeded 1.0, indicating that either test would be an appropriate measure for use in clinical trials.

666 RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards Fig 2. Pearson correlations between WMFT FA scores (A), time scores (B), and grip strength scores (C) with kinematic measures of reach (left) and grasp (right) at the 3 time points. Note that the sign of the correlations are reversed for the time scores because better time scores are indicated by smaller numbers. Abbreviations: d0, day 0; d14, day 14; d90, day 90; Grasp aperture, peak aperture during grasp; Grasp efficiency, aperture path ratio; Grasp speed, peak aperture rate; Reach accuracy, endpoint error; Reach efficiency, reach path ratio; Reach speed, peak wrist velocity. Furthermore, the high correlation between the 2 at day 90 suggests that they have a similar ability to measure the underlying construct of UE function. Lin et al 11 compared the responsiveness of the WMFT and ARAT and reported that the ARAT was more responsive than the WMFT (days 14 90:.70 vs.58) in a sample of patients admitted to a stroke rehabilitation service who were evaluated 14 days and 90 days after stroke. In our sample the responsiveness of both measures was much higher, and contrary to the Lin findings the WMFT was more responsive (days 0 90: 1.39 vs 1.60). Several important factors may explain these differences. First, the patients in the Lin study were more impaired at baseline and at each subsequent time of testing than the VECTORS participants, and the intervals used for comparison (Lin day 14 90, VECTORS day 0 90) may influence the results. It should also be noted that the VECTORS participants represent a more select sample, in that they met the inclusion criteria for a phase II randomized controlled trial and also agreed to participate in the trial. Which measure is superior in acute stroke research, the WMFT or the ARAT? From a traditional psychometric perspective, the WMFT has modest advantages in responsiveness and will therefore detect smaller degrees of change. For the trialist, the responsiveness of a scale is an important factor in estimating power and sample size for clinical trials; where subjects are difficult to recruit or per-subject study costs are high, the WMFT might be the better choice. In other contexts, such as most rehabilitation studies or routine clinical care, the small differences in responsiveness may be outweighed by the WMFT s equipment needs, time of administration, training requirements, and potential subject burden. The ARAT has some clinimetric advantages without sacrificing reliability and validity. 10 Clinimetric strategies call for more parsimonious scales with as few items as possible without compromising sensitivity to change. 38 Testing burden is challenging in early stroke recovery when fatigue and frustration can affect test performance. Shorter scales are less likely to be affected by pain, fatigue, and limited motivation on the part of the patients. For the clinical trialist, this means quicker evaluations and less testing fatigue. Study Limitations Despite the findings in support of the measurement properties of the WMFT in patients assessed early in the recovery

RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards 667 generalizability of the results, these findings should be confirmed in a larger multisite sample with a broader range of acute, stroke-related UE impairments. CONCLUSIONS We found that the internal consistency, validity, and responsiveness of the WMFT in persons with mild or moderate stroke are similar to our previously published values for the ARAT in the acute stage of stroke recovery. Comparisons with the FIM, sensorimotor, and kinematic measures of reach and grasp provide further evidence for the construct validity of the WMFT. The choice between the WMFT and ARAT should be driven by the resources available and the goals of the evaluation. Fig 3. Pearson correlation coefficients between WMFT FA scores (A), time scores (B), and grip strength scores (C) with disability measures at all 3 time points. Note that the sign of the correlations are reversed for the time scores because better time scores are indicated by smaller numbers. Abbreviations: d0, day 0; d14, day 14; d90, day 90. process, this study has limitations. In particular, the findings of this single-site phase II trial are based on a small sample of persons with mild to moderate stroke. Thus, to fully ensure the References 1. Gresham GE, Duncan PW, Stason WB. Post-stroke rehabilitation clinical practice guideline. Washington, DC: US Department of Health and Human Services; 1995; No. 16. 2. Rabadi MH, Rabadi FM. Comparison of the Action Research Arm Test and the Fugl-Meyer Assessment as measures of upperextremity motor weakness after stroke. Arch Phys Med Rehabil 2006;87:962-6. 3. Sivan M, O Connor RJ, Makower S, Levesley M, Bhakta B. Systematic review of outcome measures used in the evaluation of robot-assisted upper limb exercise in stroke. J Rehabil Med 2011; 43:181-9. 4. Fugl-Meyer AR, Jääskö L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient. 1. A method for evaluation of physical performance. Scand J Rehabil Med 1975;7:13-31. 5. Lyle R. A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res 1981;4:483-92. 6. Wolf SL, Catlin PA, Ellis M, Archer AL, Morgan B, Piacentino A. Assessing the Wolf Motor Function Test as outcome measure for research in patients after stroke. Stroke 2001;32:1635-9. 7. 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;8: 14-9. 8. DeWeert W, Harrison MA. Measuring recovery of arm-hand function in stroke patients: a comparison of the Brunnstrom-Fugl- Meyer Test and the Action Research Arm Test. Physiother Can 1985;37:65-70. 9. Wolf SL, Thompson PA, Morris DM, et al. The EXCITE trial: attributes of the Wolf Motor Function Test in patients with subacute stroke. Neurorehabil Neural Repair 2005;19:194-205. 10. Lang CE, Wagner JM, Dromerick AW, Edwards DF. Measurement of upper-extremity function early after stroke: properties of the action research arm test. Arch Phys Med Rehabil 2006;87: 1605-10. 11. Lin JH, Hsu MJ, Sheu CF. Psychometric comparisons of 4 measures for assessing upper-extremity function in people with stroke. Phys Ther 2009;89:840-50. 12. Miltner WH, Bauder H, Sommer M, Dettmers C, Taub E. Effects of constraint-induced movement therapy on patients with chronic motor deficits after stroke: a replication. Stroke 1999;30:586-92. 13. Taub E, Miller NE, Novack TA, et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil 1993; 74:347-54. 14. Wolf SL, Lecraw DE, Barton LA, Jann BB. Forced use of hemiplegic upper extremities to reverse the effect of learned nonuse among chronic stroke and head-injured patients. Exp Neurol 1989; 104:125-32. 15. Morris DM, Uswatte G, Crago JE, Cook EW 3rd, Taub E. The reliability of the Wolf Motor Function Test for assessing upper

668 RELIABILITY AND VALIDITY OF WOLF MOTOR FUNCTION TEST, Edwards extremity function after stroke. Arch Phys Med Rehabil 2001;82:750-5. 16. Wolf SL, Winstein CJ, Miller JP, et al. Effect of constraintinduced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA 2006;296:2095-104. 17. Law M, Hurley P, Russell D, King G, Hanna S. Selecting outcomes measures in children s rehabilitation: a comparison of methods. Arch Phys Med Rehabil 2003;84:496-9. 18. Law M, Letts L. A critical review of scales of activities of daily living. Am J Occup Ther 1989;43:522-8. 19. Dromerick AW, Lang CE, Birkenmeier RL, et al. Very Early Constraint-Induced Movement during Stroke Rehabilitation (VECTORS): a single-center RCT. Neurology 2009;73:195-201. 20. Jørgensen HS, Nakayama H, Raaschou HO, Vive-Larsen J, Støier M, Olsen TS. Outcome and time course of recovery in stroke. Part II: Time course of recovery. The Copenhagen stroke study. Arch Phys Med Rehabil 1995;76:406-12. 21. Dromerick AW, Edwards DF, Hahn M. Does the application of constraint-induced movement therapy during acute rehabilitation reduce arm impairment after ischemic stroke? Stroke 2000;31: 2984-8. 22. Hsieh CL, Hsueh IP, Chiang FM, Lin PH. Inter-rater reliability and validity of the Action Research Arm Test in stroke patients. Age Ageing 1998;27:107-13. 23. Platz T, Pinkowski C, van Wijck F, Kim IH, di Bella P, Johnson G. Reliability and validity of arm function assessment with standardized guidelines for the Fugl-Meyer Test, Action Research Arm Test and Box and Block Test: a multicentre study. Clin Rehabil 2005;19:404-11. 24. van der Lee JH, Roorda LD, Beckerman H, Lankhorst GJ, Bouter LM. Improving the Action Research Arm Test: a unidimensional hierarchical scale. Clin Rehabil 2002;16:646-53. 25. Granger CV, Hamilton BB. The Uniform Data System for Medical Rehabilitation report of first admissions for 1991. Am J Phys Med Rehabil 1993;72:33-8. 26. Chang WC, Slaughter S, Cartwright D, Chan C. Evaluating the FONE FIM: Part I. Construct validity. J Outcome Meas 1997;1: 192-218. 27. Brott T, Adams HP Jr, Olinger CP, et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke 1989;20: 864-70. 28. Wagner JM, Lang CE, Sahrmann SA, et al. Relationships between sensorimotor impairments and reaching deficits in acute hemiparesis. Neurorehabil Neural Repair 2006;20:406-16. 29. Zackowski KM, Dromerick AW, Sahrmann SA, Thach WT, Bastian AJ. How do strength, sensation, spasticity and joint individuation relate to the reaching deficits of people with chronic hemiparesis? Brain 2004;127:1035-46. 30. Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth Scale of muscle spasticity. Phys Ther 1987;67:206-7. 31. Andrews AW, Thomas MW, Bohannon RW. Normative values for isometric muscle force measurements obtained with hand-held dynamometers. Phys Ther 1996;76:248-59. 32. Lang CE, Wagner JM, Bastian AJ, et al. Deficits in grasp versus reach during acute hemiparesis. Exp Brain Res 2005;166:126-36. 33. Kazis LE, Anderson JJ, Meenan RF. Effect sizes for interpreting changes in health status. Med Care 1989;27:S178-89. 34. Portney LG. Watkins MP. Foundations of clinical research: applications to practice. 2nd ed. Upper Saddle River: Prentice Hall; 1999. 35. Sivan M. Interpreting effect size to estimate responsiveness of outcome measures. Stroke 2009;40:709. 36. Winstein CJ, Miller JP, Blanton S, et al. Methods for a multisite randomized trial to investigate the effect of constraint-induced movement therapy in improving upper extremity function among adults recovering from a cerebrovascular stroke. Neurorehabil Neural Repair 2003;17:137-52. 37. Dromerick AW, Lang CE, Birkenmeier R, Hahn MG, Sahrmann SA, Edwards DF. Relationships between upper-limb functional limitation and self-reported disability 3 months after stroke. J Rehabil Res Dev 2006;43:401-8. 38. Wright JG, Feinstein AR. A comparative contrast of clinimetric and psychometric methods for constructing indexes and rating scales. J Clin Epidemiol 1992;45:1201-18. Suppliers a. North Coast Medical, 1765 E Main Ave, Morgan Hill, CA 95037-3213. b. MICROFET2; Hoggan Health Industries, 12411 S 265 W, Draper, UT 84020-5412. c. Motion Analysis Corp, 3617 Westwind Blvd, Santa Rosa, CA 95403-1067. d. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.