Psychometric Properties of the Wolf Motor Function Test (WMFT) and Its Modified Versions: A Systematic Review With Meta-Analysis

Abstract

Background

The Wolf Motor Function Test (WMFT) and its modified versions are widely used to assess upper limb (UL) function in stroke survivors. However, comprehensive evaluations of its psychometric properties are lacking.

Objective

To perform a systematic review with meta-analysis on the psychometric properties (following the COnsensus-based Standards for the selection of health Measurement INstruments [COSMIN] taxonomy) of the WMFT and modified versions in stroke survivors.

Methods

Six databases were searched until May 2024 for studies examining at least one WMFT measurement property in stroke patients. Two independent reviewers conducted study selection, data extraction, and quality assessment using the COSMIN Risk of Bias checklist and quality of evidence (QoE) with the Grading of Recommendations Assessment, Development, and Evaluation approach. Meta-analyses synthesized psychometric properties reported in at least two studies.

Results

Twenty-five studies (N = 2044) were included. Regarding the WMFT Functional Ability Scale (FAS) and TIME scales, internal consistency (alpha ≥ .88), intra-rater (intraclass correlation coefficient [ICC] ≥ .97) and inter-rater (ICC ≥ .92) reliability, measurement error for TIME, construct validity (strong correlations [r ≥| .64|] with Fugl-Meyer Assessment and Action Research Arm Test), and responsiveness (ES ≥ 0.48) were rated sufficiently with QoE from very low to high. Measurement error for FAS was assessed as inconsistent with moderate QoE, and cross-cultural validity was rated as indeterminate with very low QoE. Content validity was not assessed. Few studies investigated the psychometric properties of the modified versions.

Conclusions

WMFT demonstrates robust psychometric properties in assessing UL function in stroke survivors. While the WMFT-modified versions showed promising properties, further research is needed to use them. Future studies should focus on WMFT measurement error, content, and cross-cultural validity.

Trial Review Registration:

PROSPERO: CRD42021237425.

Keywords

psychometrics outcome assessment health care reproducibility of results stroke neurorehabilitation

Introduction

Stroke is the second leading cause of death and the third cause of disability worldwide.¹ The majority of stroke survivors suffer from functional impairments affecting their quality of life and burden on society.² Upper limb (UL) sensorimotor deficits persist in ~80% of survivors, with only a small percentage showing full functional recovery within 6 months post-stroke.³ Sensorimotor deficits that may occur following a stroke are characterized by weakness or paralysis, sensory loss, spasticity, and abnormal motor synergies⁴ that impact the ability to perform daily activities.⁵

Stroke rehabilitation guidelines and the Stroke Recovery and Rehabilitation Roundtable recommend utilizing psychometrically sound outcome measures.^6,7 Standardized outcomes are crucial to accurately quantify post-stroke sensorimotor impairments and to support the development and monitoring of personalized treatment plans, enhancing communication among clinicians.^7,8

The Wolf Motor Function Test (WMFT)⁹ is widely-used to assess post-stroke UL function in clinical and research settings.^7,10 According to the International Classification of Functioning, Disability, and Health (ICF) terminology,¹¹ the WMFT assesses both the UL function and activity domains. It was originally developed to evaluate the clinical effect of constraint-induced movement therapy in stroke and traumatic brain injury survivors with mild-to-moderate UL motor impairments.⁹ The WMFT assesses 2 domains of sensorimotor function (i.e., movement quality and motor performance).¹² The original 21-item WMFT⁹ was modified to a 17-item scale^13,14 (15 function- and 2 strength-based) with tasks arranged by increasing difficulty involving proximal-to-distal joint movements.¹⁵ The 15 functional tasks are scored on a 6-point Likert scale from 0 (does not attempt with involved UL) to 5 (involved UL does attempt; movement appears normal), and time to complete each task is recorded (max 120 seconds per task; when exceeded, a score of 120 is recorded). This yields 2 scores: Functional Ability Score (FAS) rates the quality of task execution out of 75 pt, and TIME reports the total time to complete all tasks. Lower FAS scores indicate greater UL activity limitation, while higher TIME scores reflect worse performance.

Other WMFT versions have been proposed. The 13-item Graded WMFT (gWMFT) was developed to assess moderate-to-severe UL impairments accurately.¹⁶ To improve administration efficiency, a shortened 6-item Streamlined WMFT (sWMFT) version was proposed, targeting individuals at different recovery stages (subacute and chronic).¹⁷ For a description of the WMFT versions, please refer to Supplemental Material 1.

Several studies have assessed the measurement characteristics of WMFT and its modified versions, highlighting the soundness of its psychometric properties.^13,14,17 However, to our knowledge, no systematic reviews have yet been conducted to evaluate the extent, methodological quality, and findings of studies investigating WMFT measurement properties. A systematic review with meta-analysis offers the opportunity to synthesize and critically assess the existing evidence, to identify gaps, such as which properties are lacking or insufficient, to resolve inconsistencies, such as different study designs or the use of modified versions, and to provide a comprehensive overview of WMFT psychometric properties.

This review aimed to evaluate the psychometric properties (i.e., reliability, validity, and responsiveness) of WMFT versions in stroke survivors and to provide recommendations for clinical and research use.

Methods

Protocol and Registration

This study followed Cochrane Handbook guidance,¹⁸ adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement¹⁹ and COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) guidelines²⁰ and was prospectively registered on PROSPERO (No: CRD42021237425).

Search Strategy

A systematic search was performed on MEDLINE, CINAHL, EMBASE, PsycINFO, CENTRAL, and Web of Science Core Collection from their inception to May 2024, using COSMIN-based search recommendations²¹ (details in Supplemental Material 2). Additional articles were identified through manual searches of reference lists of included and related studies. Google Scholar was also used to search for combinations of tool names and measurement properties. The eligibility of the first 300 titles retrieved was assessed.²² Before performing the searches, cookies were cleared.²³ No additional limits (e.g., publication date and language) were used.

Inclusion Criteria

To be eligible, studies were required to meet the following criteria based on the PICO framework. Population (P): Post-stroke survivors; Interventions (I): Studies investigating at least one WMFT measurement property, defined according to COSMIN taxonomy, that is, reliability (internal consistency, intra- or inter-rater reliability, and measurement error), validity (content validity, structural validity, construct validity/hypothesis testing, cross-cultural validity/measurement invariance) and responsiveness²⁴ as defined in Table 1. Comparison (C): No restrictions were applied regarding the presence/type of comparator; studies without a comparator were also eligible. Outcome (O): assessment of one or more WMFT measurement properties. Systematic reviews, non-peer-reviewed articles, letters, and commentaries were excluded.

Table 1.

Measurement Property Definitions.

Domain	Definition	Clinically relevant question
Reliability	The extent to which the measurement is free from measurement error	Can the measurement consistently differentiate between true changes in a patient’s condition and variability caused by measurement error, ensuring stable scores when the patient’s condition has not changed?
Intra-rater reliability	The degree to which scores remain consistent (i.e., the same) for patients who have not changed when the same individuals (i.e., raters or responders) administer repeated measurements on different occasions (i.e., at 2 or more time points)	Does the measurement provide consistent results when administered by the same clinician, researcher, or responder at different time points?
Inter-rater reliability	The degree to which scores remain consistent (i.e., the same) for patients who have not changed when different individuals (i.e., raters) administer repeated measurements on the same occasion	Does the measurement yield the same results when administered by different clinicians or researchers on the same occasion?
Internal consistency	The degree of interrelatedness among the items, that is, how strongly the items in a measurement instrument are related to each other	Do the items within the measurement assess the same construct?
Measurement error	The systematic and random errors in a patient’s score are not due to true changes in the construct being measured.	How much of the observed change in the patient’s score reflects the true change rather than being influenced by systematic or random errors?
Standard error of measurement (SEM)	The SEM reflects the precision of an individual-level measurement. It indicates how much an observed score might deviate from the true score due to measurement error.	How much might an observed score deviate from the true score due to error?
Minimal detectable change (MDC)	The MDC represents the margin of error (amount of random measurement error) for detecting a true change in the construct being measured.	What is the smallest (minimal) change that can be confidently interpreted as a true change beyond measurement error?
Validity	The degree to which an outcome measures the construct(s) it is intended to measure	Does the outcome measure capture the construct it is intended to evaluate in the clinical population?
Content validity	The extent to which the content of a measurement tool is an adequate reflection of the construct to be measured	Does the measurement tool comprehensively represent all relevant aspects of the construct being measured?
Construct validity/hypothesis testing	The degree to which the scores of a measurement tool are consistent with hypotheses (e.g., relationships to scores from other instruments) assuming the instrument measures the construct to be measured	Do the scores of the measurement tool align with expected relationships or hypotheses that support its measurement of the intended construct?
Structural validity	The extent to which the scores of an instrument reflect the underlying dimension of the construct being assessed	Do the scores from the instrument reflect the underlying structure or dimensions of the construct being evaluated?
Cross-cultural validity/measurement invariance	The extent to which the items in a translated or culturally adapted outcome measure reflect the performance of the original version’s items	Does the translated or adapted version of the outcome measure perform equivalently to the original version across diverse cultural or linguistic environments?
Responsiveness	The ability of a measurement tool to detect change over time in the construct to be measured	Can the measurement tool detect changes in a patient’s condition over time following an intervention?

Note. All the psychometric (measurement) properties are defined according to the COSMIN taxonomy.²⁴

Study Selection

Two independent reviewers (FN and AU) performed the literature search and study selection. After removing duplicates, titles and abstracts were initially screened using Rayyan software.²⁵ Then, full texts were retrieved and reviewed for possible final inclusion. Reasons for exclusion were given for each excluded study. Disagreements were resolved through discussion with a third reviewer (LP).

Data Extraction

Two reviewers (FN and LSP) independently extracted the following data: publication year, sample characteristics (i.e., sample size, sex, age, type, and phase of stroke), study characteristics (i.e., country, setting, and inclusion criteria), measurement properties examined, and related results. Disagreements were resolved through consensus with a third reviewer (LP).

Risk of Bias(RoB) Assessment

The RoB assessment, reflecting the methodological quality of each included study investigating the measurement property, was conducted independently by two reviewers (AC and DP) using the COSMIN RoB checklist.^26,27 This checklist consists of 10 boxes evaluating the risk of bias across three measurement domains (i.e., reliability, validity, and responsiveness). Each box contains items evaluating various aspects of design requirements and statistical methods. Each item is rated using a 4-point Likert scale, that is, very good, adequate, doubtful, and inadequate quality. The total score was calculated based on the lowest rating for all items within a box (i.e., worst score counts). If multiple measurement properties were assessed, the RoB was evaluated separately for each property. Discrepancies were resolved through a third reviewer (LP). Only boxes referring to the measurement properties of the included studies were used, that is, Box#3 (Structural Validity), Box#4 (Internal Consistency), Box#5 (Cross-Cultural Validity/Measurement Invariance), Box#6 (Reliability), Box#7 (Measurement Error), Box#9 (Hypotheses Testing for Construct Validity), and Box#10 (Responsiveness).

Data Synthesis and Analysis

Study characteristics were summarized using descriptive statistics, and a qualitative synthesis (i.e., a synthesis enabling conclusions about evidence quality) was conducted for each measurement property. Meta-analysis pooled internal consistency, intra- and inter-rater reliability indexes, construct validity correlation, and responsiveness findings. We included all studies in the meta-analysis regardless of their quality. A priori, we required at least two studies investigating the same psychometric property to proceed with the meta-analysis. We performed preliminary data processing before calculating the summary correlation across studies; specifically, multiple correlations within one study (e.g., age group) were averaged (weighted mean), or if a study assessed the same measurement property across more than one group, a weighted mean was computed, thus ensuring that a single study did not disproportionately contribute to the meta-analysis.^23,28 To summarize the data, we considered Cronbach’s alpha for internal consistency and intraclass correlation coefficient (ICC), Cohen’s Kappa, weighted linear Kappa, or weighted quadratic Kappa for intra- and inter-rater reliability. Measurement error was assessed by considering the standard error of the measurement (SEM) and minimal detectable change (MDC). For construct validity, Pearson’s or Spearman’s correlation coefficients were examined. For responsiveness, we analyzed the effect sizes (ES) and standardized response means (SRM). The 95% confidence interval (CI) was computed for each statistical index.

Meta-analyses were performed using the “meta” package in R.²⁹ We estimated correlations using Fisher’s z transformation and applied inverse variance weighting for pooling. Metrics including Cronbach’s alpha, ICC, correlation coefficients, ES, and SRM were pooled using the “metacor” function, while parameters such as SEM and MDC were pooled using the “metamean” function. The variance was estimated using the standard error with a P-value equal to .05.³⁰ The random effects model, which gives more weight to studies with smaller sample sizes, was utilized to combine the pooled results. The Restricted Maximum Likelihood Estimation was used for τ² estimation, and Q-profile for estimating confidence intervals of τ² and τ.³¹ Statistical heterogeneity was assessed using I² statistics, which categorized the heterogeneity as not important (I² = 0%-40%), substantial (I² = 30%-60%), or considerable (I² = 75%-100%).¹⁸ The Egger test and graphical methods were used to investigate the publication bias if ≥10 studies were pooled in each meta-analysis.³²

Measurement Properties Rating

Rating of measurement properties facilitates the interpretation of the results. Measurement properties were evaluated against the Criteria for Good Measurement Properties³³ and rated as being sufficient (+), insufficient (−), or indeterminate (?) (see Supplemental Material 3 for details). To evaluate construct validity, reviewers established a set of a priori hypotheses based on anticipated correlations between constructs evaluated by WMFT and its modified versions and those assessed through comparator measures. WMFT construct validity and the comparator outcome measures were defined according to the ICF. We anticipated that correlations would be ≥.60 for outcomes assessing similar constructs and <.60 for those evaluating unrelated constructs. For responsiveness, a moderate ES (≥0.3) was expected to achieve a sufficient (+) rating. This threshold acknowledges the influence of multiple factors affecting responsiveness (e.g., study design, time points, interventions, and populations). However, it is still meaningful enough to suggest changes in scores.³⁴ If a study provided multiple results for a single measurement property, an overall rating was assigned based on the following criteria: if ≥75% of the results agreed on a specific rating, that rating was considered as the overall score (either sufficient [+] or insufficient [−]), otherwise, an indeterminate (?) rating was given.²⁷

Quality of the Evidence (QoE)

The QoE reflects the level of confidence that the pooled or summarized results of each measurement property estimate are trustworthy. For each measurement property, the QoE was independently determined by two reviewers (AC and DP) using the modified Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach.²⁷ Evidence was evaluated as high, moderate, low, or very low considering the RoB (i.e., COSMIN methodological quality), inconsistency, imprecision (i.e., total study sample size), and indirectness (i.e., evidence derived from populations different from the one of interest). Discrepancies were resolved by discussion with a third author (LP). Table 2 shows the definitions of the quality levels. Supplemental Materials 4 reports full details.

Table 2.

Definitions of the Quality of the Evidence.

Recommendation	Description
High	There is a high level of confidence that the true measurement property lies close to the estimate.
Moderate	There is a moderate level of confidence in the estimate. The true measurement property is likely close but could be significantly different.
Low	Confidence in the measurement property estimate is limited. The true measurement property may differ significantly.
Very Low	There is very little confidence in the estimate. The true measurement property is likely substantially different.

Note. Details are presented in Supplemental Material 4.

Results

Study Selection

Out of 4127 titles after removing duplicates, twenty-five studies were included in the qualitative synthesis^14,15,35-57 and twenty-three in the meta-analysis (Figure 1). Two studies^39,40 investigating the sWMFT structural validity could not be pooled due to differences in methodological approaches (i.e., Rasch model, which focuses on the Item Response Theory and assesses the fit of data to the model’s requirements). The list of excluded studies and reasons for exclusion are reported in Supplemental Material 5.

Figure 1.

The selection process of the study inclusion.

Study Characteristics

The studies were published between 2001^14,15 and 2020.⁵⁴ About eleven (44.0%; 95% CI [26.3, 63.4]) studies were conducted in Asia,^{35,36,39,40,44-46,49-51,57} seven (28.0%; 95% CI [13.9, 48.2]) in North America,^{14,41-43,47,55,56} five (20.0%; 95% CI [8.6, 39.9]) in Europe,^{15,37,38,52,54} and two (8.0%; 95% CI [2.0, 26.9]) in South America^48,53 (Table 1).

The sample size of the included studies was 8 to 360 (median = 48; sum = 2044). Participants were 53 to 71 years old (median = 60 years) and mostly men (64.0%: 95% CI [62.0, 67.0]). The prevalent setting was outpatient (52.0%; 95% CI [33.0, 71.0], Table 3).

Table 3.

Characteristics of the Study Included in the Systematic Review.

Study	Year	Country	Setting^a	Inclusion criteria of stroke participants	N (F)	Age (y)^b	Clinical characteristics	Type of stroke	Chronicity^c
Wolf motor function test
Adhikari et al³⁵	2016	Nepal	Inpatient	Infarction or hemorrhagic stroke; Unilateral with stroke onset < 1 mo; FMA-UL ≥ 1 in mass flexion and extension of hand function; MAS ≤ 3 grade for muscle tone; mild or no cognitive impairments (MMSE ≥ 19); normal or corrected to normal vision and hearing; age: 18-79 y	8 (2)	61.9 ± 11.7	MMSE: 24.5 ± 2.9; FMA-UL motor: 34.5 ± 8.2; FMA-UL total: 64.3 ± 18.8; MAS: 1.9 ± 0.8	I (6) H (2)	6.4 ± 2.5 mo (LS/Ch)
Ang and Man³⁶	2006	China	Outpatient	Medically stable; ability to follow simple instructions; no psychiatric history or other neurological lesions	45 (29)	53.5 ± 14.4	Side of hemiplegia: 24 right, 21 left	NR	NR
Berardi et al³⁷	2018	Italy	Inpatient	Chronic hemiplegia or hemiparesis after 1-y post stroke; MMSE ≥ 21; ability to extend the wrist, move the thumb, and 2 phalanges	24 (10)	69.9 ± 13.4	Dominant right hand: 16; affected right hand: 10; affected dominant hand: 6	NR	NR
Bürge et al³⁸	2013	Switzerland	Inpatient	Age > 18; French-speaking; medically stable without cognitive deficits that that could interfere with their participation in the tests; ability to activate the flexors and abductors of the shoulder, extend the wrist, the thumb, and 2 other fingers; ability to sit independently without support for 5 min	44 (13)	61.7 ± 13.6	Side of cerebral lesion: 27 right, 17 left; FMA-UL total <45: 6; FMA-UL total 45-59: 27 FMA-UL total 60-66): 11	I (33) H (9) Other (2)	38 ± 14 d: n = 25 (ES); 87 ± 17 d: n = 15 (ES); 1217 ± 1177 d n = 4 (Ch)
Duff et al⁴¹	2015	US	Inpatient	Adults early post-stroke	360 (NR)	NR	NR	NR	NR
Edwards et al⁴²	2012	US	Inpatient	Ischemic or hemorrhagic stroke within 28 d of admission to inpatient rehabilitation; persistent hemiparesis (motor arm item of NIHSS = 1-2); proximal UE voluntary movement (upper arm item of the Motor Assessment Scale > 3); preserved cognitive function; ability to follow 2-step commands; no UL injuries or conditions that limited use prior to the stroke	51 (29)	63.7 ± 13.6	Side of hemiplegia: 21 dominant, 29 non-dominant; Admission NIHSS score: 8.55 ± 4.52	I (39) H (11)	First evaluation: 9.5 ± 4.5 d (A/ES); Second evaluation: 25.9 ± 10.6 d (ES); Third evaluation: 110.8 ± 20.7 d (ES/LS)
Fritz et al⁴³	2009	US	Mixed	Subacute stroke, 3-9 mo post-stroke	96 (36)	62.3 (range = 19-90)	Side of hemiplegia: 43 right, 53 left
Hsieh et al⁴⁵	2009	Taiwan	Outpatient	First-ever stroke onset at least 6 mo previously; Brunnstrom stage III for the proximal UL or above; No excessive spasticity of the affected UL (MAS ≤ 2.5)	57 (18)	54.6 ± 11.5	Side of stroke: 31 left, 26 right;MMSE: 28.54 ± 1.56;FMA-UL at pre-treatment: 49.72 ± 9.30;ARAT at pre-treatment: 42.72 ± 12.11;Brunnstrom stage of the proximal UL, median: 5;Brunnstrom stage of the distal UL, median: 5;FIM-total at pre-treatment: 117.88 ± 11.02;FIM-motor at pre-treatment: 84.79 ± 8.97.	NR	12.98 ± 7.62 mo (LS/Ch)
Hui-Chan⁴⁶	2006	China	Not reported	12 mo post-stroke	22 (7)	59.9 ± 7.7	NR	I (17)H (5)	42.5 ± 25.11 mo (Ch)
Lang et al⁴⁷	2008	US	Inpatient	Ischemic or hemorrhagic stroke within 28 d of admission to inpatient rehabilitation; persistent hemiparesis, score 1-3 at motor arm item NIHSS; some UL voluntary activity; preserved cognitive function, consciousness items of NIHSS = 0-1, SBMOC test > 19; ability to follow 2-step commands, item 1c NIHSS = 0; no UL injury or conditions	52 (31)	64.0 ± 14.0	Affected side: 23 dominant, 29 nondominant;Premorbid BI: 99.6 ± 2.2;Premorbid Modified Rankin Index: 0.3 ± 0.6	I (41)H (11)	First Evaluation: 9.5 ± 4.5 d (A/ES);Second Evaluation:25.9 ± 10.6: d (ES)
Lin et al⁴⁹	2009	Taiwan	Inpatient	Diagnosis of cerebral hemorrhage or cerebral infarction; first onset of stroke; Stroke onset within 2 wk before hospital admission; ability to follow instructions	53 (22)	64.3 ± 11.7	Side of stroke: 29 right, 24 left;FMA-UL: 32.0 ± 25.9;ARAT: 18.6 ± 22.7;Admission NIHSS: 5.3 ± 1.8; BI: 9.1 ± 5.6;UE-STREAM: 10.0 ± 7.8	I (34)H (19)	Evaluation at 14 d post-stroke (ES)
					42 (16)	63.9 ± 11.2	Side of stroke: 21 right, 21 left;FMA-UL: 43.9 ± 25.3;ARAT: 31.4 ± 24.2;BI: 13.8 ± 5.8;UE-STREAM: 13.1 ± 7.7	26 (I)16 (H)	Evaluation at 30 d post-stroke (ES)
					36 (16)	64.0 ± 11.6	Side of stroke: 17 right, 19 left;FMA-UL: 48.0 ± 23.7;ARAT: 37.7 ± 23.4;BI: 16.8 ± 4.7;UE-STREAM: 14.7 ± 7.1	I (24)H (12)	Evaluation at 90 d post-stroke (LS)
					35 (14)	64 ± 12.5	Side of stroke: 14 right, 21 left;FMA-UL: 46.0 ± 24.0;ARAT: 36.3 ± 23.6;BI: 16.6 ± 5.0;UE-STREAM: 14.5 ± 7.0	I (23)H (12)	Evaluation at 180 d post-stroke (LS/Ch)
Lin et al⁵⁰	2009	Taiwan	Outpatient	6 mo post first-ever stroke; able to reach Brunnstrom stage III for upper extremity; no excessive spasticity in the shoulder and elbow joints of the affected upper extremity (MAS ≤ 2.5)	57 (18)	54.6 ± 11.5	Side of stroke: 31 left, 26 right;MMSE: 28.54 ± 1.56;Brunnstrom stage of the proximal UE: 5	NR	12.98 ± 7.62 mo (Ch)
Morris et al¹⁵	2001	UK	Outpatient	Ability to extend the more affected wrist, thumb, and at least 2 additional digits 10° at each joint; ability to understand and follow verbal directions	24 (9)	61(range: 14-86)	Side of hemiplegia: 68% right;64% dominant UL;Pre-morbidly right-handed	NR	6 y; range: 2-17 y (Ch)
Ng et al⁵¹	2008	China	Outpatient	Time since onset post-stroke > 6 mo; ability to maintain a sitting position for at least 30 min; no language problems; ability to follow a 1-2 step command	12 (4)	57.3 ± 6.9	Side of hemiplegia: 6 left, 6 right;ARAT lower level: 12.2 ± 14.9, n = 7ARAT upper level: 51.8 ± 7.4, n = 5	NR	69.3 ± 54.7 mo (Ch)
Nijland et al⁵²	2010	Netherlands	Outpatient	Hemiparesis of the UE, with at least some voluntary muscle contraction (MRC ≥ 1); no severe deficits in communication, memory, and understanding (MMSE > 22); absence of orthopedic UL limitations.	40 (17)	60.0 ± 16.6	Side of hemiplegia: 23 left, 17;MRC (0-5), median (IQR): 4.0 (4-5);ARAT, median (IQR): 38 (22-46)	NR	median (IQR): 0.41 (0.25-0.77) y (ES/LS/Ch)
Pereira et al⁵³	2011	Brazil	Outpatient	Age > 18 y; 6 mo post-stroke; no cognitive impairment	15 (5)	57.9 ± 11.1	Side of hemiplegia: 8 left; 7 right;FMA-UL: 41.6 ± 18.3;MMSE: 26.7 ± 2.7	NR	68.5 ± 53.5 mo (Ch)
Whitall et al⁵⁵	2006	US	Outpatient	Unilateral ischemic stroke; age > 25 y; 6 mo post-stroke, completion of all inpatient and outpatient therapy; residual hemiparetic motor dysfunction	66 (33)	58.0 ± 14.0	NIHSS, median (range): 0 (0-3)	NR	59.6 ± 74.6 mo (Ch)
Wolf et al¹⁴	2001	US	NR	Medical stability; sitting and standing balance control; intact cognition	19 (NR)	61.4 ± 9.5	NR	NR	4.9 ± 6.4 y (Ch)
Woodbury et al^{56 d}	2010	US	Outpatient	Ischemic or hemorrhagic stroke within 3 to 9 mo prior to enrollment. Hemiparesis of UL	189 (NR)	61 ± 13.5	NR	NR	NR
Graded Wolf motor function test
Pereira et al⁴⁸	2015	Brazil	Outpatient	Severe chronic hemiparesis; >6 mo post-stroke; mean lesion time of 6.9 ± 6.5 y; preserved cognitive function	10 (4)	53.2 ± 11.4	Side of hemiplegia: 8 right, 2 left;MMSE 25.2 ± 3.42;FMA-UL: 24.2 ± 10.23	NR	6.9 ± 6.5 y (Ch)
Turtle et al⁵⁴	2020	UK	Inpatient	Age > 18 y; stroke diagnosis within 3 mo with upper limb motor loss and upper limb rehabilitation; ability to understand and follow verbal and written commands	28 (10)	71.3 ± 9.6	Side of hemiplegia: 18 left; 10 right;	NR	14.73 ± 8.36 d
Turtle et al⁵⁴	2020	UK	Inpatient		21 (5)	70.5 ± 8.7	Side of hemiplegia: 15 left, 6 right	NR	3 mo after 14.73 ± 8.36 d (A/ES)
Streamline Wolf motor function test
Chen et al³⁹	2012	Taiwan	Outpatient	First-ever stroke with onset between 3 and 9 mo or more than 12 mo; ability to understand and respond to questions (MMSE ≥ 22); Brunnstrom ≥ stage III or higher for the proximal part of the affected arm; no excessive spasticity at any joint of the arm (MAS ≤ 2)	97 (33)	Median (range):56.6 (31.3-86.33)	Side of stroke: 50 right, 47 left	NR	median (range): 20 (12-89) mo (Ch)
Chen et al³⁹	2012	Taiwan	Outpatient		75 (17)	Median (range): 55.6 (29.6-77)	Side of stroke: 40 right, 35 left	NR	median (range): 7 (3-9) mo (LS/Ch)
Chen et al⁴⁰	2014	Taiwan	Outpatient	Brunnstrom ≥ stage II for the affected upper limb; MMSE ≥ 18; first-ever stroke; No major medical problems	94 (21)	53.2 ± 12.8	Side of hemiplegia: 50 left, 44 right;MMSE: 27.4 ± 2.62	NR	6.69 ± 1.23 mo (LS/Ch)
					58 (16)	53.7 ± 11.5	Side of hemiplegia: 42 left,16 right;MMSE: 27.58 ± 2.92	NR	10.38 ± 1.07 mo (Ch)
					199 (65)	54.8 ± 11.2	Side of hemiplegia: 134 left, 65 right;MMSE: 27.90 ± 2.10	NR	25.92 ± 15.96 mo (Ch)
Fu et al⁴⁴	2012	Taiwan	Inpatient	Subacute stroke patients, 3-9 mo post-stroke	51 (13)	55.0 ± 11.7	Side of stroke: 22 right, 29 left;MMSE: 27.08 ± 3.39;sFMA-UL: 8.63 ± 2.69;Brunnstrom stage, median (range): 4.5 (3-6);FIM: 109.96 ± 16.36	NR	5.87 ± 1.31 mo (LS/Ch)
Wu et al⁵⁷	2011	Taiwan	Inpatient	First-ever stroke; onset of more than 3 mo and less than 10 mo; Brunnstrom stage ≥III for the proximal part of the affected upper limb; no serious cognitive deficits (MMSE ≥ 24); no excessive spasticity at any joint of the upper limb (MAS ≤ 2)	64 (14)	53.0 ± 13.5	MMSE: 7.72 ± 2.31;FMA-UL: 44.84 ± 12.77;Brunnstrom stage of the proximal part of UE, median (range): 4 (3-6)	NR	7.88 ± 1.69 mo (Ch)

Abbreviations: ARAT: Action Research Arm Test; BI: Barthel Index; F: females; FIM: Functional Independence Measure; FMA-UL: Fugl-Meyer Assessment; MMSE: Mini-Mental State Examination; MRC: Medical Research Council; N: number; NIHSS: National Institutes of Health Stroke Scale; NR: not reported; SBMOC: Short Blessed Memory Orientation and Concentration; sFMA-UL: shortened Fugl-Meyer Assessment; UE-STREAM: Upper-Extremity subscale of the Stroke Rehabilitation Assessment of Movement; UL: upper limb; y: years; H, hemorrhagic stroke; I, ischemic stroke.

Outpatients included stroke participants who accessed services without being hospitalized (e.g., private clinics, general medicine clinics, outpatient clinics, or at home), while inpatients referred to patients who were hospitalized (e.g., hospitals, rehabilitation units, neurology departments, or intensive care units).

Age is reported as mean ± standard deviation unless otherwise specified.

Chronicity is reported as mean ± standard deviation in days, months, or years unless otherwise specified. Note that chronicity was classified according to the definitions provided by Bernhardt et al,⁵⁸ that is, ES: Early Subacute: 7 d mo; LS: Late Subacute: 3 to 6 mo; Ch: Chronic: >6 mo.

Demographics of the sample are reported in the primary study.⁵⁹

Nineteen studies (76.0%; 95% CI [55.8, 88.8]) evaluated the full version of the WMFT,^{14,15,35-38,41-43,45-47,49-53,55,56} teo (8.0%; 95% CI [2.0, 26.9]) the gWMFT,^48,54 and four (16.0%; 95% CI [6.4, 35.7]) the sWMFT.^39,40,44,57

Studies assessed internal consistency (N = 9),^{14,15,37-39,42,52,54,56} intra-rater (N = 9)^{15,35,37,46,49,52-55} and inter-rater reliability (N = 11),^{15,35,37,38,41,43,48,49,52-54} measurement error (N = 6),^{35,38,43,49,50,54} structural validity (N = 3),^39,40,56 cross-cultural validity (N = 6),^{35,37-39,48,53} construct validity (N = 14),^{35-38,42,44-46,49-52,55,57} responsiveness (N = 7),^{42,44,45,47,49,50,57} and interpretability (by the minimal clinically important difference [MCID]; N = 3).^47,49,50 No studies investigated content validity.

Methodological Quality

RoB assessment was conducted for each study investigating at least one psychometric property. If a study reported multiple psychometric properties, each property was evaluated. The quality of internal consistency was very good for all studies (N = 25).^{14,15,37-39,42,52,54,56} Studies analyzing intra- and inter-rater reliability showed very good (N = 7),^{9,38,41,43,49,52,54} adequate (N = 5),^{15,37,48,53,55} or doubtful (N = 2)^35,46 quality. Studies investigating measurement error were very good (N = 4),^38,43,49,54 doubtful (N = 1),³⁵ or inadequate (N = 1).⁵⁰ Studies on structural validity were rated as very good (N = 1),⁴⁰ adequate (N = 1),⁵⁶ and doubtful (N = 1)³⁹ quality. Cross-cultural validity studies were of inadequate quality (N = 6).^{35,37-39,48,53} Studies on construct validity had very good (N = 12)^{35-37,42,44,45,49-52,55,57} or adequate (N = 2)^38,46 quality. Studies examining responsiveness had very good (N = 6)^{42,44,45,47,50,57} or inadequate (N = 1)⁴⁹ quality (Table 4).

Table 4.

Quality of the studies and rating of the psychometric proprieties of the studies included in the systematic review.

Study	Subscale	Reliability			Validity			Responsiveness
Study	Subscale	Internal consistency	Reliability	Measurement error	Structural validity	Cross-cultural validity/measurement invariance	Construct validity	Responsiveness
Wolf Motor function test
Adhikari et al³⁵	FAS					Inadequate^a(?)	Very good (+)
Adhikari et al³⁵	TIME		Doubtful (+)	Doubtful (+)		Inadequate^a(?)	Very good (+)
Ang et al³⁶	FAS						Very good (+)
Ang et al³⁶	TIME						Very good (+)
Berardi et al³⁷	FAS	Very good (+)	Adequate (+)			Inadequate^a (?)	Very good (+)
Berardi et al³⁷	TIME	Very good (+)	Adequate (+)			Inadequate^a (?)	Very good (+)
Bürge et al³⁸	FAS	Very good (+)	Very good (+)	Very good (−)		Inadequate^a (?)	Adequate (+)
Bürge et al³⁸	TIME	Very good (+)	Very good (+)	Very good (−)		Inadequate^a (?)	Adequate (+)
Duff et al⁴¹	FAS		Very good (+)
Edwards et al⁴²	FAS	Very good (+)					Very good (+)	Very good (+)
Edwards et al⁴²	TIME						Very good (+)	Very good (+)
Fritz et al⁴³	FAS		Very good (+)	Very good (+)
Fritz et al⁴³	TIME		Very good (+)	Very good (+)
Hsieh et al⁴⁵	FAS						Very good (+)	Very good (+)
Hsieh et al⁴⁵	TIME						Very good (+)	Very good (+)
Hui-Chan⁴⁶	FAS		Doubtful (+)				Adequate (+)
Hui-Chan⁴⁶	TIME		Doubtful (+)				Adequate (+)
Lang et al⁴⁷	FAS							Very good (+)
Lang et al⁴⁷	TIME							Very good (+)
Lin et al⁴⁹	FAS		Very good (+)	Very good (−)			Very good (+)	Inadequate (+)
Lin et al⁵⁰	FAS			Inadequate (+)			Very good (+)	Very good (+)
Lin et al⁵⁰	TIME			Inadequate (+)			Very good (+)	Very good (+)
Morris et al¹⁵	FAS	Very good (+)	Adequate (+)
Morris et al¹⁵	TIME	Very good (+)	Adequate (+)
Ng et al⁵¹	FAS						Very good (+)
Nijland et al⁵²	TIME	Very good (+)	Very good (+)				Very good (+)
Nijland et al⁵²	FAS	Very good (+)	Very good (+)				Very good (+)
Pereira et al⁵³	FAS		Adequate (+)			Inadequate (?)
Pereira et al⁵³	TIME		Adequate (+)			Inadequate (?)
Whitall et al⁵⁵	FAS		Adequate (+)				Very good (+)
Whitall et al⁵⁵	TIME		Adequate (+)				Very good (+)
Wolf et al¹⁴	TIME	Very good (+)	Very Good (+)
Woodbury et al⁵⁶	FAS	Very good (+)			Adequate (+)
Graded Wolf Motor function test
Pereira et al⁴⁸	FAS		Adequate (+)			Inadequate (?)
Pereira et al⁴⁸	TIME		Adequate (+)			Inadequate (?)
Turtle et al⁵⁴	FAS	Very Good (+)	Very Good (+)	Very Good (?)
Turtle et al⁵⁴	TIME	Very Good (+)	Very Good (+)	Very Good (?)
Streamline Wolf Motor function test
Chen et al³⁹	FASc	Very Good (+)			Doubtful (+)	Inadequate(+)
Chen et al³⁹	FASs	Very Good (+)			Doubtful (+)	Inadequate(+)
Chen et al⁴⁰	FAS				Very Good (+)
Fu et al⁴⁴	TIME						Very Good (−)	Very Good (+)
Wu et al⁵⁷	TIME						Very Good (±)	Very Good (+)

Abbreviations: FAS, Functional Ability Scale; FASc: Functional Ability Scale for chronic subjects; FASs: Functional Ability Scale for subacute subject.

+: sufficient rating; −: insufficient rating; ?: indeterminate rating; ±: non-consistent.

Cross-cultural transition, according to Beaton et al.⁶⁰

Psychometric Properties

Table 5 reports the individual psychometric properties for each WMFT version. Table 6 summarizes the meta-analysis summary results. Supplemental Material 6 to 17 shows the forest plots. Publication bias was not assessed as no meta-analysis included ≥10 studies. Table 7 reports the QoE based on GRADE.

Table 5.

Results of the Studies of the Psychometric Properties of the Wolf Motor Function Test and Its Modified Version Included in the Systematic Review.

	Scale	Internal consistency				Intra-rater reliability				Inter-rater reliability				Measurement error				Construct validity					Responsiveness
Study	Scale	N	C	Value	R	N	C	Value	R	N	C	Value	R	N	C	Value	R	N	Correlation with	C	Value	R	N	C	Value	R
Wolf Motor function test
Adhikari et al³⁵	FAS																	8	FMA-UL	ρ	0.91	+
	TIME					8	ICC_2.1	1.00	+	8	ICC_3.1	1.00	+	8	SEM	0.00^a	+	8	FMA-UL	ρ	−0.95	+
	TIME													8	MDC	0.00^a	+
Ang et al³⁶	FAS																	45	FMA-UL	ρ	0.90	+
																		45	Brunnstrom’s stage	ρ	0.94	+
																		45	Brunnstrom’s scale	ρ	0.93	+
																		45	mBI	ρ	0.53	—
	TIME																	44	FMA-UL	ρ	−0.94	+
																		44	Brunnstrom’s stage	ρ	−0.90	+
																		44	Brunnstrom’s scale	ρ	−0.89	+
																		44	mBI	ρ	−0.60	+
Berardi et al³⁷	FAS	24	α	0.98	+
Berardi et al³⁷	TIME	24	α	0.91	+
Bürge et al³⁸	FAS	44	α	0.93^b	+					44	ICC	0.92^b	+	44	SEM	1.94	—	44	FMA-UL	ρ	0.75^b	+
	FAS													44	MDC	5.38	—
	TIME	44	α	0.84^b	+					44	ICC	0.98^b	+	44	SEM	21.61	—	44	FMA-UL	ρ	−0.60^b	+
	TIME													44	MDC	59.91	—
Duff et al⁴¹	FAS									233	WK	0.84	+
Edwards et al⁴²	FAS	51	α	0.97^b	+													51	ARAT	r	0.81^b	+	51	ES	1.09^c	+
	FAS																						40	ES	1.63^d	+
	TIME																	51	ARAT	r	−0.75 ^b	+	51	ES	0.61^c	+
	TIME																						40	ES	0.85^d	+
Fritz et al⁴³	FAS									96	ICC	0.96	+	96	SEM	0.10	+
	FAS													96	MDC	0.10	+
	TIME									96	ICC	0.98	+	96	SEM	0.20	+
	TIME													96	MDC	0.70	+
Hsieh et al⁴⁵	FAS																	57	ARAT	ρ	0.72^b	+	57	SRM	1.30^d	+
																		57	FMA-UL	ρ	0.61^b	+
																		57	FIM	ρ	0.29	—
	TIME																	57	ARAT	ρ	0.60^b	+	57	SRM	0.38^d	+
																		57	FMA-UL	ρ	0.73^b	+
																		57	FIM	ρ	0.40	—
Hui-Chan⁴⁶	FAS					22	ICC	0.99	+									22	FMA-UL	r	0.84	+
Hui-Chan⁴⁶	TIME					22	ICC	0.99	+									22	FMA-UL	r	−0.73	+
Lang et al⁴⁷	FAS																						23	ES	0.91^c	+
	FAS																						29	ES	1.00^c	+
	TIME																						23	ES	0.48^c	+
Lin et al⁴⁹	FAS					30	ICC	0.97	+	30	ICC	0.92	+	30	SEM	3.56^b	—	42	ARAT	ρ	0.91^b	+	42	ES	0.44 ^c	+
														30	MDC	12.00^b	—	42	FMA-UL	ρ	0.94^b	+	42	ES	0.58 ^d	+
																							42	ES	0.64^e	+
Lin et al⁵⁰	FAS													57	SEM	0.16	+						57	ES	0.41^c,a	+
	FAS													57	MDC	0.44^a	+
	TIME													57	SEM	1.87	+						57	ES	0.36^c,a	+
	TIME													57	MDC	5.17^a	+
Morris et al¹⁵	FAS	15	α	0.92^b	+	20	r	0.95	+	22	ICC_2.1	0.88^b	+
Morris et al¹⁵	TIME	16	α	0.89^b	+	20	r	0.90	+	22	ICC_2.1^a	0.98^b	+
Ng et al⁵¹	FAS																	12	ARAT	ρ	0.96	+
																		12	FTHUE	ρ	0.92	+
																		12	QoM of the MAL	ρ	0.86	+
																		12	AoU of the MAL	ρ	0.92	+
Nijland et al⁵²	FAS	40	α	0.98	+	40	ICC	0.95	+	40	ICC	0.94	+					40	ARAT	ρ	0.86	+
Nijland et al⁵²	TIME																	40	ARAT	ρ	−0.89	+
Pereira et al⁵³	FAS					15	LWK	0.93	+	15	ICC	0.94	+
Pereira et al⁵³	TIME					15	ICC	0.99	+	15	ICC	0.99	+
Whitall et al⁵⁵	FAS					66	ICC_2.1	0.97	+									66	FMA-UL	ρ	0.88	+
Whitall et al⁵⁵	TIME					66	ICC_2.1	0.99	+									66	FMA-UL	ρ	−0.88	+
Wolf et al¹⁴	TIME	19	α	0.92	+													19	FMA-UL	ρ	−0.62^b	+
Woodbury et al⁵⁶	FAS	189	α	0.98	+
Graded Wolf motor function test
Pereira et al⁴⁸	FAS									10	ICC	0.98	+
Pereira et al⁴⁸	TIME									10	ICC	0.99	+
Turtle et al⁵⁴	FAS	24	α	0.99^b	+		ICC_3.1	0.99	+	28	ICC_2.1	0.98	+	21	SEM	0.19^a	?
	FAS													21	MDC	0.53^a	?
	TIME	24	α	0.98^b	+		ICC_3.1	0.99	+	28	ICC_2.1	0.99	+	21	SEM	3.64^a	?
	TIME													21	MDC	10.09 ^a	?
Streamline Wolf Motor function test
Chen et al⁴⁰	FAS^f	172	α	0.91	+
Chen et al⁴⁰	FAS^g	172	α	0.91	+
Fu et al⁴⁴	TIME																	51	S-FMA	r	0.49^b	—	51	ES	0.47^c,b	+
Fu et al⁴⁴	TIME																						51	SRM	0.33^e	+
Wu et al⁵⁷	TIME																	64	FMA-UL	ρ	0.64	+	64	ES	0.52^e	+
																		64	SIS	ρ	0.53	—	64	SRM	0.37^e	+

Abbreviations: AoU of the MAL, Amount of the Use of the Motor Activity Log; ARAT: Action Research Arm Test; BI: Barthel Index; C, coefficient; ES: effect size; FAS: Functional Ability Scale; FTHUE: Functional Test for the Hemiplegic Upper Extremity; FIM: Functional Independence Measure; FMA-UL: Fugl-Meyer Assessment Upper Limb; ICC: intraclass correlation coefficient; LWK: linear weighted Kappa; mBI: modified Barthel Index; MDC: minimal detectable change; N: number of subjects; QoM of the MAL, quality of the movement of the Motor Activity Log.

G; ρ: Spearman’s rank correlation coefficient; r: Pearson’s correlation coefficient; R, rating; SEM: standard error of the measurement; sFMA-UL: shortened Fugl-Meyer Assessment; SIS: Stroke Impact Scale; SRM: standardized response mean; α:Cronbach’s alpha.

Rating rules were according to Supplemental Material 3.

Value calculated using the data reported in the article.

Value expressed as the mean of multiple values.

Time interval of 2/3 wk.

Time interval of 90 d.

Time interval of 6 mo.

For sub-acute patients.

For chronic patients.

Table 6.

Summary of the Meta-analyses for Each Scale and Each Psychometric Property.

Scale	Psychometric properties and statistical index	Subscale	N	Sample size	Fixed effects	95% CI fixed effects	Random effects	95% CI random effects	Q	Tau²	I ²
WMFT	Internal consistency (Cronbach’s alpha)	FAS	6	363	0.97	0.97, 0.98	0.97	0.95, 0.98	20.58	0.07	76
	Internal consistency (Cronbach’s alpha)	TIME	4	102	0.88	0.83, 0.92	0.88	0.83, 0.92	2.31	0.00	0
	Intra-rater reliability (ICC)	FAS	6	193	0.97	0.95, 0.97	0.97	0.95, 0.98	9.36	0.03	47
	Intra-rater reliability (ICC)	TIME	5	131	0.99	0.98, 0.99	0.99	0.96, 1.00	28.36	0.51	86
	Inter-rater reliability (ICC)	FAS	7	480	0.90	0.88, 0.92	0.92	0.88, 0.95	41.29	0.06	85
	Inter-rater reliability (ICC)	TIME	5	185	0.98	0.98, 0.99	0.99	0.97, 0.99	11.62	0.12	66
	Measurement error (SEM)	FAS	4	227	0.11	0.10, 0.12	1.42	−0.18, 3.02	301.88	2.64	99
	Measurement error (SEM)	TIME	3	205	0.21	0.19, 0.23	7.81	−5.58, 21.20	356.26	139.20	99
	Measurement error (MDC)	FAS	4	227	0.11	0.10, 0.12	4.41	−0.98, 9.79	426.60	29.88	99
	Measurement error (MDC)	TIME	3	205	0.75	0.68, 0.82	21.69	−15.39, 58.77	343.25	1067.01	99
	Construct validity with ARAT (rho)	FAS	5	202	0.84	0.79, 0.88	0.86	0.76, 0.92	13.70	0.07	71
	Construct validity with ARAT (rho)	TIME	3	148	−0.75	−0.82, −0.67	−0.77	−0.89, −0.55	11.44	0.11	83
	Construct validity with FMA-UL (rho)	FAS	7	260	−0.82	−0.85, −0.77	−0.81	−0.90, −0.67	33.06	0.14	82
	Construct validity with FMA-UL (rho)	TIME	8	291	−0.64	−0.71, −0.56	−0.72	−0.91, −0,29	225.71	0.71	97
	Responsiveness (ES)	FAS	4	202^a	0.51	0.48, 0.55	0.76	0.48, 1.05	124.16	0.10	97
		FAS	2	82^b	0.69	0.61, 0.78	1.10	0.07, 2.13	60.48	0.54	98
		TIME	3	131^a	0.43	0.39, 0.47	0.48	0.34, 0.63	27.31	0.01	93
Graded WMFT	Inter-rater reliability (ICC)	FAS	2	38	0.98	0.96, 0.99	0.98	0.96, 0.99	0.00	0.00	0
Graded WMFT	Inter-rater reliability (ICC)	TIME	2	38	0.99	0.97, 0.99	0.99	0.97, 0.99	0.16	0.00	0
Streamline WMFT	Responsiveness (ES)	TIME	2	115^a	0.50	0.45, 0.54	0.50	0.45, 0.54	1.17	0.00	15
Streamline WMFT	Responsiveness (SRM)	TIME	2	115^a	0.35	0.32, 0.38	0.35	0.31, 0.39	1.50	0.00	33

Abbreviations: ARAT: Action Research Arm Test; CI: confidence interval; ES: effect size; FAS: Functional Ability Scale; FMA-UL: Fugl-Meyer Assessment Upper Limb; I²: I² statistics; ICC, intraclass correlation coefficient; MDC: minimal detectable change; n: number of studies included in the meta-analysis; N: number of studies included; Q: Q-statistics; rho, correlation coefficient; SEM: standard error of the measurement; SRM: standardized response mean; WMFT: Wolf Motor Function Test.

Time interval of 2/3 wk.

Time interval of 90 days.

Table 7.

Quality of the Evidence According to Modified GRADE and Overall Rating of the Psychometric Properties.

Measurement properties	Subscale	Risk of bias	Inconsistency	Imprecision^a	Indirectness	Quality of evidence	Overall rating
Wolf motor function test
Reliability
Internal consistency	FAS	0	0	0	0	High	+
Internal consistency	TIME	0	0	0	0	High	+
Reliability	FAS	0	0	0	0	High	+
Reliability	TIME	0	0	0	0	High	+
Measurement error	FAS	0	−1	0	0	Moderate	±
Measurement error	TIME	−1	0	0	0	Moderate	+
Validity
Structural validity	FAS	−1	0	0	0	Moderate	+
Structural validity	TIME	—	—	—	—	N/A	N/A
Cross-cultural validity/measurement invariance	FAS	−3	0	0	0	Very Low	?
Cross-cultural validity/measurement invariance	TIME	−3	0	0	0	Very Low	?
Construct validity	FAS	0	0	0	0	High	+
Construct validity	TIME	0	0	0	0	High	+
Responsiveness	FAS	0	0	0	0	High	+
Responsiveness	TIME	0	0	0	0	High	+
Graded Wolf motor function test^b
Reliability
Internal consistency	FAS	0	0	−2	0	Low	+
Internal consistency	TIME	0	0	−2	0	Low	+
Reliability	FAS	0	0	−2	0	Low	+
Reliability	TIME	0	0	−2	0	Low	+
Measurement error	FAS	0	0	−2	0	Low	?
Measurement error	TIME	0	0	−2	0	Low	?
Validity
Cross-cultural validity/measurement invariance	FAS	−3	0	0	0	Very Low	?
Cross-cultural validity/measurement invariance	TIME	−3	0	0	0	Very Low	?
Streamline Wolf Motor function test^c
Reliability
Internal consistency	FAS	0	0	0	0	High	+
Internal consistency	TIME	—	—	—	—	N/A	N/A
Validity
Structural validity	FAS	−1	0	0	0	Moderate	+
Structural validity	TIME	—	—	—	—	N/A	N/A
Cross-cultural validity/measurement invariance	FAS	−3	0	0	0	Very Low	+
Cross-cultural validity/measurement invariance	TIME	—	—	—	—	N/A	N/A
Construct validity	FAS	—	—	—	—	N/A	N/A
Construct validity	TIME	0	−1	0	0	Moderate	±
Responsiveness	FAS	—	—	—	—	N/A	N/A
Responsiveness	TIME	0	0	0	0	High	+

Abbreviations: N/A, not available; FAS, Functional Ability Scale.

Not considered for content validity, structural validity, and cross-cultural validity.

No studies were available for structural validity, construct validity, and responsiveness.

No studies were available for reliability, measurement error.

Scores are given based on the modified GRADE criteria defined in Supplemental Material 4. +: sufficient rating; −: insufficient rating; ?: indeterminate rating; ±: inconsistent rating.

Wolf Motor Function Test

Internal Consistency

Six studies^{15,37,38,42,52,56} investigated WMFT-FAS internal consistency (N = 363) with a pooled Cronbach’s alpha of .97 (95% CI [0.95, 0.98]; I² = 76%, Table 6, Supplemental Material 6a).

Four studies^14,15,37,38 examined the internal consistency of WMFT-TIME (N = 102), with a pooled Cronbach’s alpha of .88 (95% CI [0.83, 0.92]; I² = 0%, Table 6, Supplemental Material 6b).

The overall rating of the internal consistency was sufficient for FAS and TIME with a high QoE. (Table 7).

Reliability

Intra-rater reliability of WMFT-FAS was investigated in six studies^{15,46,49,52,53,55} (N = 193). The pooled ICC was 0.97 (95% CI [0.95, 0.98]; I² = 47%, Table 6, Supplemental Material 7a).

Five studies^{15,35,46,53,55} explored intra-rater reliability of WMFT-TIME (N = 131), with a pooled ICC of 0.99 (95% CI [0.96, 1.00]; I² = 86%, Table 6, Supplemental Material 7b).

Inter-rater reliability of WMFT-FAS was investigated in seven studies^{15,38,41,43,49,52,53} (N = 480), with a pooled ICC of 0.92 (95% CI [0.88, 0.95]; I² = 85%, Table 6, Supplemental Material 8a).

Five articles^{15,35,38,43,53} (N = 185) studied the inter-rater reliability of WMFT-TIME. The pooled ICC was 0.99 (95% CI [0.97, 0.99]; I² = 66%, Table 6, Supplemental Material 8b).

The overall rating for intra- and inter-rater reliability was sufficient for FAS and TIME with a high QoE (Table 7).

Measurement Error

Four studies explored the SEM and MDC of WMFT-FAS^38,43,49,50 (N = 227). The pooled SEM was 1.42 pt (95% CI [−0.18, 3.02]; I² = 99%, Table 6, Supplemental Material 9a), and the pooled MDC was 4.41 pt (95% CI [−0.98, 9.79]; I² = 99%, Table 6, Supplemental Material 10a).

Four studies^35,38,43,50 explored SEM and MDC of the WMFT-TIME (N = 205). The pooled SEM was 7.81 seconds (95% CI [−5.58, 21.20]; I² = 99%, Table 6, Supplemental Material 9b), and the pooled MDC was 21.69 seconds (95% CI [−15.39, 58.77]; I² = 99%, Table 6, Supplemental Material 10b).

The overall rating for measurement error was inconsistent for FAS and sufficient for TIME, both with moderate QoE (Table 7).

Structural Validity

Woodbury et al⁵⁶ assessed FAS structural validity using confirmatory factor analysis and Rasch analysis, supporting a single-factor structure and found to be sufficient with a moderate QoE (Table 7). The TIME structural validity was not assessed.

Cross-Cultural Validity

Following international recommendations, translation and back-translation processes did not present any major difficulties in adapting to Italian,³⁷ French,³⁸ Brazilian-Portuguese,^48,53 or Nepali.³⁵

Overall, cross-cultural validity ratings for WMFT-FAS and TIME were indeterminate, with a very low QoE (Table 7).

Construct Validity

The correlation between the WMFT-FAS and the Action Research Arm Test (ARAT) scores was investigated in five studies^{42,45,49,51,52} N = 202). The pooled rank correlation coefficient was 0.86 (95% CI [0.76, 0.92]; I² = 71%, Table 6, Supplemental Material 11a).

Three studies^42,45,52 (N = 148) compared WMFT-TIME and ARAT scores. The pooled rank correlation coefficient was −0.77 (95% CI [−0.98, −0.55]; I² = 83%), Table 6, Supplemental Material 11b).

The correlation between the WMFT-FAS and Fugl-Meyer Assessment-Upper Limb (FMA-UL) scores was assessed in seven studies^{35,36,38,45,46,49,55} (N = 260), with a pooled rank correlation coefficient of −0.81 (95% CI [−0.90, −0.67]; I² = 82%, Table 6, Supplemental Material 12a).

Seven studies^{14,35,36,38,45,46,55} compared WMFT-TIME and FMA-UL scores (N = 260), showing a pooled rank correlation coefficient of −0.81 (95% CI [−0.90, −0.67]; I² = 97%; Table 6, Supplemental Material 12b).

Ang and Man³⁶ compared WMFT-FAS and TIME scores with Brunnstrom’s stage and Brunnstrom’s scale and modified Barthel Index scores. Results showed correlations of 0.94 (95% CI [0.89; 0.97]), 0.93 (95% CI [0.88, 0.96]), and 0.53 (95% CI [0.28, 0.71]), respectively, for WMFT-FAS, and −0.90 (95% CI [−0.94, −0.82]), −0.89 (95% CI [−0.94, −0.81]), and −0.60 (95% CI [0.76, −0.37]), respectively, for WMFT-TIME. Hsieh et al⁴⁵ reported correlations between WMFT and the Functional Independence Measure scores equal to 0.29 (95% CI [0.03, 0.52]) for FAS and 0.40 (95% CI [0.16, 0,60]) for TIME. Construct validity was also examined by comparing WMFT-FAS and Functional Test for the Hemiplegic Upper Extremity scores, revealing a correlation coefficient of 0.92 (95% CI [0.73, 0.98]).⁵¹ Additionally, a relationship was found between WMFT-FAS and Quality of Movement and Amount of Use Motor Activity Log scores (coefficients: 0.86 (95% CI [0.56, 0.96] and 0.92 (95% CI [0.73, 0.98], respectively).⁵¹

The overall rating of construct validity for the FAS and TIME was sufficient, with a high QoE (Table 7).

Responsiveness

The meta-analysis was stratified based on the time-interval between assessments. The interventions were heterogeneous among the four included studies.^42,47,49,50 In Edwards et al⁴² and Lang et al,⁴⁷ participants underwent either early high-dose (90% of waking hours) or moderate-dose (6 hours) Constraint-Induced Therapy or conventional treatment (e.g., VECTORS trial). Lin et al⁴⁹ reported distributed Constraint-Induced Therapy, bilateral arm training, or traditional rehabilitation. Lin et al⁵⁰ did not describe the intervention.

All four studies^42,47,49,50 reported responsiveness of WMFT-FAS (time interval between assessments, 14-21 days; N = 202) with a pooled ES of 0.76 (95% CI [0.48, 1.05]; I² = 97%, Table 6, Supplemental Material 13a).

Three studies investigated the responsiveness of WMFT-TIME (time interval between assessments, 14-21 days)^42,47,50 (N = 131), with a pooled ES of 0.48 (95% CI [0.34, 0.63]; I² = 93%, Table 6, Supplemental Material 13b).

Two studies^42,49 (N = 82) analyzed the responsiveness of WMFT-FAS (time interval between assessments, 90 days) and had a pooled ES of 1.10 (95% CI [0.07, 2.13]; I² = 98%, Table 6, Supplemental Material 14a).

The overall responsiveness rating for the FAS and TIME was sufficient, with a high QoE (Table 7).

Feasibility

Two studies^42,52 found no evidence of floor or ceiling effects for WMFT-FAS and TIME, whereas one study⁴⁹ reported a floor effect in one of four evaluations and a ceiling effect in three of four assessments.

Two studies investigated the MCID of WMFT-FAS and TIME. Lang et al⁴⁷ in acute stroke (within 14 days post-stroke) reported FAS MCID raw values of 1.0 point for the dominant UL and 1.2 points for the nondominant UL. They also observed a TIME MCID of −19.0 seconds for the dominant UL. Lin et al⁵⁰ in chronic stroke survivors (12.98 ± 87.62 months) found an MCID of 0.33 points and −1.64 seconds for the WMFT-FAS and TIME, respectively.

Graded WMFT

Inter-Rater Reliability

The inter-rater reliability of gWMFT-FAS was studied in two articles^48,54 (N = 38). The meta-analysis showed an ICC of 0.98 (95% CI [0.96, 0.99]; I² = 0%, Table 6, Supplemental Material 15a).

Two studies^48,54 reported the inter-rater reliability of gWMFT-TIME (N = 38) with a pooled ICC of 0.99 (95% CI [0.97, 0.99]; I² = 0%, Table 6, Supplemental Material 15b).

The overall inter-reliability rating was sufficient for FAS and TIME, with a low QoE (Table 4).

Measurement Error

The SEM and MDC of gWMFT-FAS and TIME were reported in one study⁵⁴ (N = 21); the SEM was 0.19 pt (95% CI [0.14, 0.23]) and 3.64 pt (95% CI [2.87, 4.41]) for gWMFT-FAS and TIME, respectively. Moreover, the MDC was 0.53 seconds (95% CI [0.42, 0.64]) and 10.09 seconds (95% CI [7.95, 12.23]) for gWMFT-FAS and TIME, respectively.

The overall rating of the measurement error for gWMFT-FAS and TIME with a low QoE was indeterminate (Table 7).

Cross-Cultural Validity

One study⁴⁸ translated the gWMFT into Brazilian-Portuguese without cross-cultural adaptation and differential item functioning (DIF) investigation.

The overall rating of cross-cultural validity for gWMFT-FAS and TIME was indeterminate, with a very low QoE (Table 7).

Other Psychometric Properties

Other gWMFT psychometric properties were not assessed.

Feasibility

Two studies^48,54 investigated the presence of the ceiling/floor effects. Pereira et al⁴⁸ did not report ceiling/floor effects for the gWMFT-FAS. Turtle et al⁵⁴ found a floor effect for TIME and FAS at two different assessments.

Streamlined WMFT

Internal Consistency

One study³⁹ assessed the internal consistency of the sWMFT (N = 172), with a Cronbach’s alpha of .91 (95% CI [0.88, 0.93])

The overall rating of the internal consistency was sufficient for FAS, with a high QoE (Table 7).

Structural Validity

The chronic and subacute versions of sWMFT (6 items, FAS Component)³⁹ and the combined version (8 items)⁴⁰ have sufficient structural validity, demonstrating a one-factor structure through Rasch analysis.

The overall structural validity rating was sufficient for FAS, with a moderate QoE (Table 6).

Cross-Cultural Validity

Measurement invariance had a sufficient rating for DIF in sex, age, or stroke hemispheric location.³⁹

The overall rating of the cross-cultural validity was sufficient for FAS, with a very low QoE (Table 7).

Construct Validity

Two studies investigated the TIME component.^44,57

The overall rating of construct validity was inconsistent for TIME with a moderate QoE (Table 7).

Responsiveness

Two studies^44,57 assessed the responsiveness of sWMFT-TIME (N = 115). Fu et al⁴⁴ did not report the intervention, while in Wu et al,⁵⁷ participants received constraint-induced therapy (CIT), bilateral arm training, or conventional rehabilitation. The combined ES was 0.50 (95% CI [0.45-0.54]; I² = 15%, Table 6, Supplemental Material 16a), whereas the pooled SRM was 0.35 (95% CI [0.31, 0.39]; I² = 33%), Table 6, Supplemental Material 17a).

The overall rating of responsiveness was sufficient for TIME, with a high QoE (Table 7).

Other Psychometric Properties

Other sWMFT psychometric properties were not assessed.

Discussion

This study determined the quality of the evidence supporting the use of WMFT for assessing UL function in post-stroke individuals following the COSMIN methodology.^20,24,26,27 The review included twenty-five studies involving 2009 participants in clinical and research settings, while the meta-analysis provided quantitative benchmarks (i.e., MDC and ES) for interpreting change scores. Most studies focused on WMFT and found sufficient psychometric properties with very low-to-high QoE. Fewer studies investigated the gWMFT and sWMFT, showing some evidence of their psychometric properties (i.e., sufficient reliability, validity, and responsiveness) with very low-to-high QoE. For WMFT, the FAS and TIME subscales demonstrated high internal consistency and inter- and intra-rater reliability with a high overall QoE. However, the measurement error for FAS and TIME was inconsistent and sufficient, respectively, supported by a moderate QoE. FAS structural validity was rated as moderate, while cross-cultural validity for both subscales was undetermined with very low QoE. Construct validity and responsiveness had high QoE for both subscales. We found sufficient internal consistency and reliability for gWMFT, supported by low QoE, undetermined measurement error and cross-cultural validity with low and very low QoE, respectively. The sWMFT showed high QoE and sufficient internal consistency (FAS) and responsiveness (TIME). Validity ratings were mixed, with structural and construct validity rated as moderate and sufficient, while QoE for cross-cultural validity was very low. Notably, no studies investigated content validity according to COSMIN standards for any version. However, the development studies^9,16,17 provided an adequate theoretical framework for the constructs.

Most studies, which had very good methodological quality, supported excellent internal consistency, intra-rater reliability, and inter-rater reliability of WFMT. These findings underscore WMFT reliability, indicating that it produces stable results between assessments in the same patient when evaluated by the same or different evaluators.

FAS and TIME measurement error (i.e., SEM and MDC), however, should be further investigated, as the QoE was moderate for both, and the overall rating was inconsistent for the latter. SEM represents absolute reliability,⁶¹ while MDC defines the margin of error for detecting true changes in the measured construct.^61,62 To be rated as sufficient, SEM and MDC should demonstrate values lower than the MCID³³ (Supplemental Material 3). The rating was mostly influenced by studies of very good methodological quality, which reported higher SEM and MDC (in both FAS and TIME components³⁸; one study⁵⁰ investigating FAS) in comparison to MCID values retrieved by Lang et al⁴⁷ It should be noted that measurement error is computed through distribution-based approaches and is directly influenced by sample intra-subject variability (i.e., MDC depends on SEM, which is a function of the standard deviation of the measured outcome).^62,63 Therefore, even in the presence of excellent inter-rater reliability, SEM and MDC values may still be substantial.⁶¹ Overall, pooled results indicated that when administering WMFT-FAS and TIME, a minimum change of 4.41 points and 21.69 seconds, respectively, should be considered as true change beyond systematic and random error.⁶¹

WMFT-FAS and TIME have been rigorously evaluated for construct validity with very good or adequate methodological quality. The a priori hypothesis tests were met for all studies, with most examining the relationship between FMA-UL and ARAT. As expected, the pooled coefficient showed a strong correlation with FMA-UL and ARAT, indicating that WMFT is closely related to both impairment (function)-level constructs (as measured by FMA-UL) and activity-level constructs (as measured by ARAT).¹²

WMFT has been translated and adapted into several languages.^{35,37,38,48,53} Three studies^35,37,38 were performed according to guideline recommendations,⁶⁰ and one study⁵³ used a back-to-forward translation.⁶⁴ While the reported psychometric properties were rated as sufficient, no study conducted a cross-cultural DIF analysis,⁶⁵ with no evidence that the measure is invariant across cultural contexts.^61,66 Therefore, caution should be taken when comparing or pooling WMFT scores derived from other countries in secondary studies, as DIF may lead to systematic bias in scoring,⁶⁵ especially on self-reported competence (i.e., FAS). No studies investigated DIF regarding demographic or clinical characteristics. Further research is needed to determine whether dominance of the affected UL may influence scale performance, as impaired UL dominance affects functional recovery.^67,68 Structural validity was assessed only for FAS, with one methodologically adequate study reporting a sufficient unidimensional construct through Rasch analysis.⁵⁶

WMFT showed sufficient ability to detect change over time, making it suitable for tracking motor recovery following interventions. Notably, the pooled results highlighted that ESs were higher in the acute compared to the acute-subacute phase of stroke. Such findings could be explained by the ES = 1.63⁴¹ for WMFT-FAS scores derived from the VECTORS trial,⁶⁹ which involved intensive training. Responsiveness is influenced by demographic factors, clinical characteristics, intervention type, and the implementation environment.^61,70 We stratified studies by the assessment time window rather than intervention type due to intervention heterogeneity. This approach could enhance the results generalizability, as the interventions ranged from conventional rehabilitation to intensive care. The pooled ES may assist researchers in estimating sample sizes and power for clinical trials and provide benchmarks for interpreting change scores while acknowledging the variability and characteristics of the original studies.

Considering feasibility, all but one study reported no floor or ceiling effects, indicating that the measurement range may be suitable for stroke survivors. The WMFT MCID was investigated in 2 studies^47,50 with different stroke chronicity populations, that is, acute⁴⁷ and chronic.⁵⁰ We suggest considering MCID WMFT values provided by Lang et al,⁴⁷ as they were determined through an anchor-based approach with a global rating of perceived changes^61,71 following interventions. MCID values are intimately linked to the sample from which they are derived.^71,72 Therefore, caution should be used when using these thresholds, as MCID values were derived from an acute stroke population (14 days post-stroke). Individuals’ meaningful improvement expectations may vary across recovery stages.^61,73

gWMFT was designed to assess individuals with moderate-to-severe UL impairment.¹⁶ However, limited studies have evaluated its psychometric properties. gWMFT (FAS and TIME) showed sufficient internal consistency and intra- and inter-rater reliability. However, measurement error was indeterminate, and all the remaining properties have not been studied. As a result, while some psychometric properties of the modified versions have been investigated, caution is advised when administering and interpreting gWMFT scores until further evidence is available.

The sWMFT was developed in the EXCITE trial to reduce administration time.¹⁷ Notably, intra- and inter-rater reliability and measurement error were not reported. The unidimensional factor for FAS was successfully studied, with no significant DIF reported for relevant demographic and clinical characteristics. The responsiveness for TIME was rated as sufficient. However, the MCID was not reported. Overall, the sWMFT warrants further investigation, and users should exercise caution as the SEM and MDC are unknown.

Limitations

Our study has some limitations. Publication bias could occur as we did not search study registers and excluded non-peer-reviewed studies. Additionally, we were unable to evaluate it as there were <10 studies in the meta-analysis. However, we searched 6 databases using a sensitive search string (without language or time restrictions) as recommended by COSMIN²¹ and prospectively registered the review on PROSPERO. While no librarian was involved in the search process, we retrieved and adopted the search strategy specifically developed by COSMIN for identifying the measurement properties of health outcome measures.²¹ Finally, studies where WMFT versions were not primary outcome measures were excluded following COSMIN recommendations.²⁷

Future Directions

Given the comprehensive evidence supporting its psychometric properties, WMFT is a reliable and valid tool for assessing UL function in stroke survivors. The pooled responsiveness scores may help researchers achieve more precise sample estimations. At the same time, SEM, MDC, and MCID can aid both clinicians and researchers in detecting statistical and clinical changes following interventions, considering the sample from which they were derived. However, a formal study to investigate WMFT content validity is needed,⁷⁴ involving evaluation of relevance, comprehensiveness, and comprehensibility, and including all stakeholders, and also to support TIME structural validity.

Future research should investigate complementary methods to assess motor function alongside WMFT by integrating technology that enables more granular assessment. For example, incorporating kinetic and kinematic analysis into specific WMFT items could provide valuable insights into recovery trajectories. The shift toward more detailed evaluations aligns with the growing emphasis on personalized rehabilitation.

Finally, gWMFT and sWMFT provide preliminary evidence of sufficient psychometric properties. However, the lack of more extensive evidence indicates that they should not be preferred over the full version. Future research should focus on filling these gaps in the use of WMFT scales in research and clinical practice.

Conclusions

Our findings showed that the WMFT demonstrated sufficient internal consistency and intra-/inter-rater reliability for both the FAS and TIME components, with a high QoE. Measurement error was inconsistent for WMFT-FAS but sufficient for WMFT-TIME, and both were rated as moderate QoE. Construct validity for FAS and TIME was sufficient, with high QoE. Structural validity was supported for FAS, but no evidence was available for TIME. Cross-cultural validity was rated as indeterminate with very low QoE. Content validity was not formally investigated. Responsiveness was sufficient, with a high QoE.

The gWMFT demonstrated sufficient intra-/inter-rater reliability but indeterminate measurement error and cross-cultural validity, both rated with low QoE. The sWMFT showed sufficient internal consistency, structural validity, and responsiveness for TIME, with moderate-to-high QoE, but lacked evidence on intra-/inter-rater reliability and measurement error. Further studies are needed to strengthen confidence in the full measurement properties of the sWMFT and gWMFT for clinical and research use.

Supplemental Material

sj-docx-1-nnr-10.1177_15459683251327568 – Supplemental material for Psychometric Properties of the Wolf Motor Function Test (WMFT) and Its Modified Versions: A Systematic Review With Meta-Analysis

Supplemental material, sj-docx-1-nnr-10.1177_15459683251327568 for Psychometric Properties of the Wolf Motor Function Test (WMFT) and Its Modified Versions: A Systematic Review With Meta-Analysis by Lorena Sabrina Pometti, Daniele Piscitelli, Alessandro Ugolini, Francesco Ferrarello, Francesco Notturni, Andrea Coppari, Serena Caselli, Fabio La Porta, Mindy F. Levin and Leonardo Pellicciari in Neurorehabilitation and Neural Repair

Footnotes

Author Contributions

Lorena Sabrina Pometti: Data curation; Investigation; Methodology; Writing - original draft; Writing - review & editing. Daniele Piscitelli: Conceptualization; Data curation; Methodology; Writing - original draft; Writing - review & editing. Alessandro Ugolini: Formal analysis; Software; Visualization; Writing - original draft. Francesco Ferrarello: Formal analysis; Investigation; Methodology; Writing - original draft; Writing - review & editing. Francesco Notturni: Investigation; Writing - original draft. Andrea Coppari: Formal analysis; Investigation; Writing - original draft. Serena Caselli: Methodology; Writing - review & editing. Fabio La Porta: Formal analysis; Writing - review & editing. Mindy F. Levin: Methodology; Writing - review & editing. Leonardo Pellicciari: Conceptualization; Methodology; Supervision; Writing - original draft; Writing - review & editing.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Daniele Piscitelli

Francesco Ferrarello

Mindy F. Levin

Leonardo Pellicciari

Supplementary material for this article is available on the Neurorehabilitation & Neural Repair website along with the online version of this article.

References

Feigin

Brainin

Norrving

, et al. World Stroke Organization (WSO): global stroke fact sheet 2022. Int J Stroke. 2022;17(1):18-29. doi:10.1177/17474930211065917

Tsao

Aday

Almarzooq

, et al. Heart disease and stroke statistics-2023 update: a report from the American Heart Association. Circulation. 2023;147(8):e93-e621. doi:10.1161/CIR.0000000000001123

Kwakkel

Kollen

van der Grond

Prevo

AJ.

Probability of regaining dexterity in the flaccid upper limb: impact of severity of paresis and time since onset in acute stroke. Stroke. 2003;34(9):2181-2186. doi:10.1161/01.STR.0000087172.16305.CD

Raghavan

Upper limb motor impairment after stroke. Phys Med Rehabil Clin N Am. 2015;26(4):599-610. doi:10.1016/j.pmr.2015.06.008

Nichols-Larsen

Clark

Zeringue

Greenspan

Blanton

Factors influencing stroke survivors’ quality of life during subacute recovery. Stroke. 2005;36(7):1480-1484. doi:10.1161/01.STR.0000170706.13595.4f

Bernhardt

Borschmann

Boyd

, et al. Moving rehabilitation research forward: developing consensus statements for rehabilitation and recovery research. Int J Stroke. 2016;11(4):454-458. doi:10.1177/1747493016643851

Santisteban

Teremetz

Bleton

Baron

Maier

Lindberg

PG.

Upper limb outcome measures used in stroke rehabilitation studies: a systematic literature review. PLoS One. 2016;11(5):e0154792. doi:10.1371/journal.pone.0154792

Sullivan

Crowner

Kluding

, et al. Outcome measures for individuals with stroke: process and recommendations from the American Physical Therapy Association neurology section task force. Phys Ther. 2013;93(10):1383-1396. doi:10.2522/ptj.20120492

Wolf

Lecraw

Barton

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(2):125-132.

10.

Alt Murphy

Resteghini

Feys

Lamers

. An overview of systematic reviews on upper extremity outcome measures after stroke. BMC Neurol. 2015;15:29. doi:10.1186/s12883-015-0292-6

11.

World Health Organization. International Classification of Functioning, Disability and Health (ICF). World Health Organization; 2001.

12.

Demers

Levin

MF.

Do activity level outcome measures commonly used in neurological practice assess upper-limb movement quality?

Neurorehabil Neural Repair. 2017;31(7):623-637. doi:10.1177/1545968317714576

13.

Taub

Miller

Novack

, et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil. 1993;74(4):347-354.

14.

Wolf

Catlin

Ellis

Archer

Morgan

Piacentino

Assessing Wolf Motor Function Test as outcome measure for research in patients after stroke. Stroke. 2001;32(7):1635-1639. doi:10.1161/01.str.32.7.1635

15.

Morris

Uswatte

Crago

Cook

III Taub

The reliability of the Wolf Motor Function Test for assessing upper extremity function after stroke. Arch Phys Med Rehabil. 2001;82(6):750-755. doi:10.1053/apmr.2001.23183

16.

Constraint Induced Movement Therapy Research Group. Manual: Graded Wolf Motor Function Test. University of Alabama and Birmingham Veteran’s Administration Centre; 2002.

17.

Bogard

Wolf

Zhang

Thompson

Morris

Nichols-Larsen

Can the wolf motor function test be streamlined?

Neurorehabil Neural Repair. 2009;23(5):422-428. doi:10.1177/1545968308331141

18.

Higgins

JPT

Thomas

Chandler

, et al. Cochrane Handbook for Systematic Reviews of Interventions version 6.4 Cochrane. 2023. Accessed July 4, 2024. https://www.training.cochrane.org/handbook.

19.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.

20.

Gagnier

Lai

Mokkink

Terwee

CB.

COSMIN reporting guideline for studies on measurement properties of patient-reported outcome measures. Qual Life Res. 2021;30(8):2197-2218. doi:10.1007/s11136-021-02822-4

21.

Terwee

Jansma

Riphagen

de Vet

HC.

Development of a methodological PubMed search filter for finding studies on measurement properties of measurement instruments. Qual Life Res. 2009;18(8):1115-1123. doi:10.1007/s11136-009-9528-5

22.

Haddaway

Collins

Coughlin

Kirk

The role of Google Scholar in evidence reviews and its applicability to grey literature searching. PLoS One. 2015;10(9):e0138237. doi:10.1371/journal.pone.0138237

23.

Piscitelli

Ferrarello

Ugolini

Verola

Pellicciari

Measurement properties of the gross motor function classification system, gross motor function classification system-expanded & revised, manual ability classification system, and communication function classification system in cerebral palsy: a systematic review with meta-analysis. Dev Med Child Neurol. 2021;63(11):1251-1261. doi:10.1111/dmcn.14910

24.

Mokkink

Terwee

Patrick

, et al. The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health-related patient-reported outcomes. J Clin Epidemiol. 2010;63(7):737-745. doi:10.1016/j.jclinepi.2010.02.006

25.

Ouzzani

Hammady

Fedorowicz

Elmagarmid

Rayyan-a web and mobile app for systematic reviews. Syst Rev. 2016;5(1):210. doi:10.1186/s13643-016-0384-4

26.

Mokkink

Boers

van der Vleuten

CPM

, et al. COSMIN risk of bias tool to assess the quality of studies on reliability or measurement error of outcome measurement instruments: a Delphi study. BMC Med Res Methodol. 2020;20(1):293. doi:10.1186/s12874-020-01179-5

27.

Mokkink

de Vet

HCW

Prinsen

CAC

, et al. COSMIN Risk of Bias checklist for systematic reviews of Patient-Reported Outcome Measures. Qual Life Res. 2018;27(5):1171-1179. doi:10.1007/s11136-017-1765-4

28.

Senn

SJ.

Overstating the evidence: double counting in meta-analysis and related problems. BMC Med Res Methodol. 2009;9(1):10. doi:10.1186/1471-2288-9-10

29.

Schwarzer

Carpenter

Rücker

, SpringerLink. Meta-Analysis with R. 1st 2015. ed. Use R! Springer International Publishing; 2015.

30.

Parmar

Torri

Stewart

Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med. 1998;17(24):2815-2834. doi:10.1002/(sici)1097-0258(19981230)17:24<2815::aid-sim110>3.0.co;2-8

31.

Viechtbauer

Bias and efficiency of meta-analytic variance estimators in the random-effects model. J Educ Behav Stat. 2005;30(3):261-293.

32.

Sterne

Egger

Funnel plots for detecting bias in meta-analysis: guidelines on choice of axis. J Clin Epidemiol. 2001;54(10):1046-1055. doi:10.1016/s0895-4356(01)00377-8

33.

Prinsen

CAC

Mokkink

Bouter

, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Qual Life Res. 2018;27(5):1147-1157. doi:10.1007/s11136-018-1798-3

34.

Cohen

Statistical Power Analysis for the Behavioral Sciences. 2nd ed. L. Erlbaum Associates; 1988:xxi, 567 p.

35.

Adhikari

Tretriluxana

Chaiyawat

Reliability and validity of the Nepali Wolf Motor Function Test following Cross-cultural Adaptation. Kathmandu Univ Med J (KUMJ). 2016;14(53):3-8.

36.

Ang

Man

DW.

The discriminative power of the Wolf Motor Function Test in assessing upper extremity functions in persons with stroke. Int J Rehabil Res. 2006;29(4):357-361. doi:10.1097/MRR.0b013e3280118e65

37.

Berardi

Dhrami

Tofani

Valente

Sansoni

Galeoto

Cross-cultural adaptation and validation in the Italian population of the wolf motor function test in patients with stroke. Funct Neurol. 2018;33(4):229-253.

38.

Bürge

Kupper

Badan Ba

Leemann

Berchtold

Qualities of a French version of the Wolf Motor Function Test: a multicenter study. Ann Phys Rehabil Med. 2013;56(4):288-299. doi:10.1016/j.rehab.2013.03.003

39.

Chen

Lin

Chen

CK.

Rasch validation of the streamlined Wolf Motor Function Test in people with chronic stroke and subacute stroke. Phys Ther. 2012;92(8):1017-1026. doi:10.2522/ptj.20110175

40.

Chen

Lin

, et al. Measurement properties of streamlined wolf motor function test in patients at subacute to chronic stages after stroke. Neurorehabil Neural Repair. 2014;28(9):839-846. doi:10.1177/1545968314526643

41.

Duff

Nelsen

, et al. Interrater reliability of the Wolf Motor Function Test-Functional Ability Scale: why it matters. Neurorehabil Neural Repair. 2015;29(5):436-443. doi:10.1177/1545968314553030

42.

Edwards

Lang

Wagner

Birkenmeier

Dromerick

AW.

An evaluation of the Wolf Motor Function Test in motor trials early after stroke. Arch Phys Med Rehabil. 2012;93(4):660-668. doi:10.1016/j.apmr.2011.10.005

43.

Fritz

Blanton

Uswatte

Taub

Wolf

SL.

Minimal detectable change scores for the Wolf Motor Function Test. Neurorehabil Neural Repair. 2009;23(7):662-667. doi:10.1177/1545968309335975

44.

Lin

, et al. Psychometric comparison of the shortened Fugl-Meyer Assessment and the streamlined Wolf Motor Function Test in stroke rehabilitation. Clin Rehabil. 2012;26(11):1043-1047. doi:10.1177/0269215511431474

45.

Hsieh

Lin

Chang

Chen

Liu

JS.

Responsiveness and validity of three outcome measures of motor function after stroke rehabilitation. Stroke. 2009;40(4):1386-1391. doi:10.1161/STROKEAHA.108.530584

46.

Hui-Chan

A study on intrarater reliability and criterion validity of Wolf Motor Function Test. Chin J Rehabil Med. 2006;21(12):1084-1086.

47.

Lang

Edwards

Birkenmeier

Dromerick

AW.

Estimating minimal clinically important differences of upper-extremity measures early after stroke. Arch Phys Med Rehabil. 2008;89(9):1693-1700. doi:10.1016/j.apmr.2008.02.022

48.

Pereira

Vieira

Pompeu

Menezes

IDS

Anjos

SMD

Ovando

AC.

Translation, cultural adaptation and reliability of the brazilian version of the Graded Wolf Motor Function Test in adults with severe hemiparesis. Fisioterapia em Movimento. 2015;28(4):667-676. doi:10.1590/0103-5150.028.004.Ao03

49.

Lin

Hsu

Sheu

, et al. Psychometric comparisons of 4 measures for assessing upper-extremity function in people with stroke. Phys Ther. 2009;89(8):840-850. doi:10.2522/ptj.20080285

50.

Lin

Hsieh

Chen

Jang

Liu

JS.

Minimal detectable change and clinically important difference of the Wolf Motor Function Test in stroke patients. Neurorehabil Neural Repair. 2009;23(5):429-434. doi:10.1177/1545968308331144

51.

AKY

Leung

DPK

Fong

KNK

. Clinical utility of the action research arm test, the wolf motor function test and the motor activity log for hemiparetic upper extremity functions after stroke: a pilot study. Hong Kong J Occ Ther. 2008;18(1):20-27. doi:10.1016/s1569-1861(08)70009-3

52.

Nijland

van Wegen

Verbunt

van Wijk

van Kordelaar

Kwakkel

A comparison of two validated tests for upper limb function after stroke: the Wolf Motor Function Test and the Action Research Arm Test. J Rehabil Med. 2010;42(7):694-696. doi:10.2340/16501977-0560

53.

Pereira

Michaelsen

Menezes

Ovando

Lima

RCM

Teixeira-Salmela

LF.

Confiabilidade da versão brasileira do Wolf Motor Function Test em adultos com hemiparesia. Brazilian J Phys Ther. 2011;15(3):257-265. doi:10.1590/s1413-35552011000300013

54.

Turtle

Porter-Armstrong

Stinson

The reliability of the graded Wolf Motor Function Test for stroke. British J Occ Ther. 2020;83(9):585-594. doi:10.1177/0308022620902697

55.

Whitall

Savin

Jr. Harris-Love

Waller

SM.

Psychometric properties of a modified Wolf Motor Function test for people with mild and moderate upper-extremity hemiparesis. Arch Phys Med Rehabil. 2006;87(5):656-660. doi:10.1016/j.apmr.2006.02.004

56.

Woodbury

Velozo

Thompson

, et al. Measurement structure of the Wolf Motor Function Test: implications for motor control theory. Neurorehabil Neural Repair. 2010;24(9):791-801. doi:10.1177/1545968310370749

57.

Lin

, et al. Assessing the streamlined Wolf Motor Function Test as an outcome measure for stroke rehabilitation. Neurorehabil Neural Repair. 2011;25(2):194-199. doi:10.1177/1545968310381249

58.

Bernhardt

Hayward

Kwakkel

, et al. Agreed definitions and a shared vision for new standards in stroke recovery research: the Stroke Recovery and Rehabilitation Roundtable taskforce. Int J Stroke. 2017;12(5):444-450. doi:10.1177/1747493017711816

59.

Wolf

Winstein

Miller

, et al. Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA. 2006;296(17):2095-2104. doi:10.1001/jama.296.17.2095

60.

Beaton

Bombardier

Guillemin

Ferraz

MB.

Guidelines for the process of cross-cultural adaptation of self-report measures. Spine (Phila Pa 1976). 2000;25(24):3186-3191. doi:10.1097/00007632-200012150-00014

61.

De Vet

Terwee

Mokkink

Knol

DL.

Measurement in Medicine: A Practical Guide. Cambridge University Press; 2011.

62.

Piscitelli

Banina

Lam

Chen

Levin

MF.

Psychometric properties of a new measure of upper limb performance in post-stroke individuals: trunk-based index of performance. Neurorehabil Neural Repair. 2023;37(1):66-75. doi:10.1177/15459683221143462

63.

Crosby

Kolotkin

Williams

GR.

Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol. 2003;56(5):395-407. doi:10.1016/s0895-4356(03)00044-1

64.

Acquadro

Conway

Hareendran

Aaronson

, European Regulatory I, Quality of Life Assessment G. Literature review of methods to translate health-related quality of life questionnaires for use in multinational clinical trials. Value Health. 2008;11(3):509-521. doi:10.1111/j.1524-4733.2007.00292.x

65.

Krogsgaard

Brodersen

Christensen

, et al. How to translate and locally adapt a PROM. Assessment of cross-cultural differential item functioning. Scand J Med Sci Sports. 2021;31(5):999-1008. doi:10.1111/sms.13854

66.

Holland

Wainer

Differential Item Functioning. Routledge; 2012.

67.

Hmaied Assadi

Feige Gross-Nevo

Dudkiewicz

Barel

Rand

. Improvement of the upper extremity at the subacute stage poststroke: does hand dominance play a role? Neurorehabil Neural Repair. 2020;34(11):1030-1037. doi:10.1177/1545968320962502

68.

Harris

Eng

JJ.

Individuals with the dominant hand affected following stroke demonstrate less impairment than those with the nondominant hand affected. Neurorehabil Neural Repair. 2006;20(3):380-389. doi:10.1177/1545968305284528

69.

Dromerick

Lang

Birkenmeier

, et al. Very early constraint-induced movement during stroke rehabilitation (VECTORS): a single-center RCT. Neurology. 2009;73(3):195-201. doi:10.1212/WNL.0b013e3181ab2b27

70.

Terwee

Roorda

Knol

De Boer

De Vet

HC.

Linking measurement error to minimal important change of patient-reported outcomes. J Clin Epidemiol. 2009;62(10):1062-1067. doi:10.1016/j.jclinepi.2008.10.011

71.

Podurgiel

Piscitelli

Denegar

Challenges in applying minimal clinically important difference: a critical review. Int J Rehabil Res. 2024;47(1):10-19. doi:10.1097/MRR.0000000000000613

72.

Wang

Hart

Stratford

Mioduski

JE.

Baseline dependency of minimal clinically important improvement. Phys Ther. 2011;91(5):675-688. doi:10.2522/ptj.20100229

73.

Copay

Subach

Glassman

Polly

Jr. Schuler

TC.

Understanding the minimum clinically important difference: a review of concepts and methods. Spine J. 2007;7(5):541-546. doi:10.1016/j.spinee.2007.01.008

74.

Terwee

Prinsen

CAC

Chiarotto

, et al. COSMIN methodology for evaluating the content validity of patient-reported outcome measures: a Delphi study. Qual Life Res. 2018;27(5):1159-1170. doi:10.1007/s11136-018-1829-0

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.77 MB