Journal of Neurologic Physical Therapy:
A Systematic Review of Voluntary Arm Recovery in Hemiparetic Stroke: Critical Predictors for Meaningful Outcomes Using the International Classification of Functioning, Disability, and Health
Chen, Shu-Ya PT, MS; Winstein, Carolee J. PT, PhD, FAPTA
Division of Biokinesiology and Physical Therapy, School of Dentistry, University of Southern California, Los Angeles, California.
Address correspondence to: Carolee Winstein, E-mail: email@example.com
Background and Purpose: Stroke is the leading cause of long term disability in the United States. Most poststroke individuals experience difficulty in moving the contralesional limbs to perform daily activities. Studies were systematically reviewed to identify the best predictors of arm-specific motor recovery and further analyzed using the International Classification of Functioning, Disability, and Health to distinguish outcome measures that best reflected participation in life situations.
Methods: We used PubMed, CINAHL, the Cochrane Library, and EMBASE with the keywords stroke, upper extremity, recovery of function, and predictor. Inclusion and exclusion criteria were developed and applied to further refine the search. Finally, methodological quality was assessed.
Results: Fifty-six studies published between 1979 and 2008 met the criteria. There was a 317% increase in the frequency of articles on predictors of arm recovery over a nearly 30-year period. Thirty-six percent were of high methodological quality (score, ≥10 of 15). Early neurophysiologic and sensorimotor measures were the best predictors at follow-up of arm-specific outcomes. There was no outcome measure classified at the International Classification of Functioning, Disability, and Health participation level. Only one study provided information about an outcome-related minimal clinically important difference.
Discussion and Conclusions: Initial measures of the integrity of neural connectivity and voluntary motor behavior were the best predictors of arm-specific functional recovery. The paucity of valid and reliable instruments to capture the more distal outcomes associated with activities and participation has likely limited the breadth of available evidence in this field. This suggests an urgent need for the development of direct measures of arm use in real-life environments.
Approximately 780,000 people have a new or recurrent stroke every year, leading to a heavy burden of medical expense in the United States.1 Many survivors have impairments as a result of stroke, making it a leading cause of long-term disability in the United States.1 After discharge from acute hospital care, most poststroke individuals experience difficulty in moving the contralesional limbs, resulting in an inability to accomplish daily activities.2 A recent large cohort study determined that 6 months after stroke, 65% of persons (n = 434) with an average Barthel Index (BI) score of 90 of 100 could not incorporate the paretic arm into life activities.3 This finding suggests that the BI may be insensitive to meaningful recovery of the paretic arm. Exploration of the true recovery of paretic arm use after hemiparetic stroke is becoming more important in light of these inconsistencies in what constitutes full recovery.
Many measures have been used to capture functional motor recovery of the upper extremity. Each may reflect motor capability within a spectrum of disablement from impairment (eg, muscle strength) to disability (eg, role in home and community). The World Health Organization's International Classification of Functioning, Disability, and Health (ICF) provides a multidimensional framework for health and disability.4 In the ICF, meaningful recovery of arm function is defined as the ability to incorporate the paretic arm in home and community activities and therefore to enhance participation. One example is the ability to independently carry a tray of food across a crowded room during a community outing. In fact, a recent questionnaire survey (n = 220) of stroke survivors found that the most important outcome was the ability to use the paretic arm in meaningful ways.5,6 One person said “It is a big deal to be able to use your arm again. I think most of the doctors think it is not. It is a big deal to be able to use your arm again psychologically as well physically.”5 This underscores the importance of the clinical meaningfulness of outcomes from the person's perspective.
To better identify poststroke upper extremity motor recovery, an objective, valid, and reliable measure of arm use in real-life situations is sorely needed. Unfortunately, there are very few such tools. Subcomponents embedded in the activities of daily living measurement, such as the BI, are commonly used in clinical practice and research.7 To the authors' knowledge, the accelerometer,8 the Motor Activity Log (MAL),9 and the Actual Amount of Use Test10 are the only instruments developed and used for this purpose. The MAL is a semistructured questionnaire interview that relies on individual recall and the Actual Amount of Use Test is a laboratory based covert observation of spontaneous arm use in structured scenarios. Although the MAL has been widely used in stroke intervention studies, its reliability11 and longitudinal construct validity12 has been questioned.
Previous investigations into predictors of functional motor recovery have been limited to global outcomes,13–15 with little specificity provided for predictors of upper limb motor recovery. For example, a systematic review identified predictors of poststroke functional recovery in activities of daily living.16 These predictors included age, previous stroke, urinary continence, consciousness at onset, disorientation in time and place, severity of paralysis, sitting balance, admission activities of daily living score, level of social support, and metabolic rate of glucose outside the infarct area in hypertensive individuals. However, the review did not distinguish predictors of upper limb motor function from the more global functional recovery outcomes. The current review focuses on the predictors of arm-specific recovery.
Previous studies investigating upper extremity motor function recovery have used a wide range of outcome measures that span from basic motor capabilities to global motor function.17–22 For this review, we use the ICF to categorize predictors and outcome measures to allow further evaluation and comparison across studies. Therefore, the purpose of this focused review is to identify potential predictors of poststroke upper extremity motor recovery. The questions we posed to guide the review were (1) which predictors have the highest sensitivity for capturing arm-specific functional motor recovery after stroke and (2) which of the currently used outcome measures best reflects arm use in real-life situations?
Inclusion and Exclusion Criteria
Table 1 summarizes the seven inclusion and five exclusion criteria for study selection. A diagnosis of unilateral stroke had to be confirmed by either clinical neurologic examination or brain imaging data. We excluded strokes that result in bilateral involvement because the etiology, pathogenesis, and recovery patterns are considerably different from the unilateral hemiparetic type.23,24 Only studies whose participants were aged 18 years or older were included due to the potential influence of developmental processes on motor recovery.25 Studies whose participants had cerebellar lesions were not included because their consequent motor deficits are different from that of cerebral lesions.26 Studies selected for inclusion had to have included at least 10 research participants because determination of clinical prediction rules for outcome research requires a ratio of 10 per predictor variable investigated.27
Outcomes of special interest are measures of arm-specific voluntary motor recovery after stroke. Therefore, studies that used other outcome measures were excluded. Studies that examined voluntary motor function using global outcome measures without distinguishing a subscale of upper extremity motor function were excluded because a global score neither specifies upper extremity motor function nor necessarily considers interaction between subscales.28 The Functional Independence Measure29 and the BI30 are two commonly used global outcome measures that contain a subscale of upper extremity motor function, but these subscales were not used separately in most of the studies we identified.31–34
Literature Search and Screening Strategy
Articles published before April 2008 were searched in PubMed, CINAHL, the Cochrane Library, and EMBASE with the keywords stroke, upper extremity, recovery of function, and predictor, and their synonyms and plurals. Nine hundred thirty-five articles were found for further consideration, of which 473 were in English and study participants were older than 18 years of age. Three hundred twenty-three of those were excluded according to criteria listed in Table 1, leaving 150 articles. Of those, 63 articles included a measurement category specific to upper extremity function. Seven articles with a sample size fewer than 10 were then excluded. Our search for predictors of poststroke functional upper extremity motor recovery yielded a total of 56 articles (6% = 56 of 935) that met the review criteria.
Evidence Strength Assessment
The 56 studies were subjected to a methodological quality evaluation that included internal validity, statistical validity, and external validity. Grading criteria were summarized in Table 2. We modeled our criteria after previous works by Hier and Edelstein27 and Kwakkel et al.16 Each item was scored using a bimodal distribution: yes (1) or no (0). The total possible score of methodological quality was 15, with 5 points given for each validity domain. Studies with 10 points or more were subjected to further analysis.
Internal Validity Criteria
Adequate Definitions of Outcome Measures and Predictor Measures
Outcome measures were used for the follow-up phase of stroke and predictor measures were used to capture the initial capability of the contralesional arm use. Both should be precisely defined to reach a common level of understanding and to ensure an accurate interpretation by readers.27
Reliable or Valid Measurements
Measures were considered reliable and valid if the study reported the reliability and validity or provided reference from the literature for the psychometric properties of each measure.35
Results were obtained by testers who were unaware of the study design to reduce observation bias from influencing conclusions.36
Appropriate Time Point to Capture Predictors
The time point for obtaining predictor measures was considered appropriate if the measures were obtained within the first three months after stroke onset.
Control of Dropout
An acceptable dropout rate was defined as a sample size at follow-up large enough to ensure statistical power in the prediction model.36 A score of 1 was given when the author provided a paragraph of text description about participant dropout or when a table indicated participant numbers at each of the different evaluation time points.
Statistical Validity Criteria
Control for Statistical Significance
Control for statistical significance was considered sufficient if the rationale for testing the cause-and-effect relationship between predictor and outcome measures was specified, and the variance explained was tested for statistical significance.36
Adequate Sample Size
A sample size was considered adequate in a prognostic study if there were at least 10 patient cases for each predictor measure examined.27
Control for Multicollinearity
To determine the best predictors for functional upper extremity motor recovery, the interaction between two or more predictors had to be tested to avoid redundant predictors in the prediction model.36 NA was given instead of a 0 for this criterion when studies did not establish a regression model for outcome prediction. In addition, the studies scored as NA would be dropped out when calculating the overall score for evidence strength. Evidence strength where multicollinearity was not tested was considered weaker than in the case where a test for multicollinearity was included.
External Validity Criteria
Identification of Stroke Pathology
The results had to be stratified by stroke pathology, (ie, infarction or hemorrhage) because these two stroke pathology types may exhibit a different time course of recovery.27
Specification of Inclusion and Exclusion Criteria
Relevant characteristics for participants in a stroke group (eg, age, stroke stage, lesion localization) had to be specified.37
Description of Additional Treatment Effects During the Period of Observation
Information about medical or paramedical interventions had to be specified because additional treatment effects may influence the identification of predictors for upper extremity function recovery.
Cross-Validation of the Prediction Model
Prediction models had to be validated in a second independent group of poststroke individuals to ensure that results could be generalized across samples.36
Description of Clinical Meaningfulness
Minimal clinically important differences (MCIDs) that are essential to patients and clinicians in predictor and outcome measures had to be specified38 because a clinically meaningful change in a score on any particular outcome measure is important.
Fifty-six articles met both the inclusion and exclusion criteria. Cumulatively, these 56 articles involved 2965 persons with unilateral stroke in various brain regions other than the cerebellum (Table 3). In total, there were approximately 85 predictors and 30 outcome measures described in the subgroup of 56 studies. Twenty-one studies were published during the 20-year period between 1979 and 1999, and 35 studies were published between 2000 and April 2008. Annually, 1.05 studies were published before 2000 compared with 4.38 studies published after 2000. This shows a 317% [(4.38–1.05)/1.05] yearly rate of increase for these kinds of studies.
Table 3 summarizes the number of persons classified by stroke stage (acute/subacute or chronic) and stroke pathology (infarction or hemorrhage). Approximately 90% (2659 of 2965) of the stroke participants in the selected studies were examined in the acute/subacute stage. With regard to stroke pathology, 37% (1106 of 2965) were classified with infarction, 2.5% (75 of 2965) were classified with hemorrhage, 30.5% (907 of 2965) were not classified by type of stroke pathology, and 30% (877 of 2965) did not report stroke pathology.
The ICF descriptions of body functions/structures, activities, and participation were used to classify predictors and outcome measures. These are presented in Tables 4–7. Other than predictors in the region of demography and medical history (eg, age, gender, time since stroke onset, number of comorbid conditions), predictors within each domain in the ICF were further categorized into neurophysiologic factors (Table 4), motor behavioral factors (Table 5), and sociocognitive factors (Table 6). We found that most of the predictors and outcome measures were from the body functions/structures domain of the ICF. Relatively few identified predictors and outcome measures were from the activities domain. None of the measures were from the participation domain.
Among the neurophysiologic factors, transcranial magnetic stimulation (TMS)–related measures were found to be the most commonly used to examine upper extremity motor function after stroke (Table 4). They were used in 27% (15 of 56) of the studies reviewed. The predictors of physical factors included visual functions, speech functions, upper extremity sensorimotor functions, and lower extremity motor functions (Table 5). Not surprisingly, the most popular predictor measures were tests of motor impairment, sensory impairment, and arm and hand function. However, only three studies53,55,65 examined hand dexterity, which is very important for upper extremity function in daily life. Incorporating the contralesional arm into life activities is a realistic and important goal for poststroke persons.6 All sociocognitive factors fell into the body functions/structures category. The defining feature of each, including perceptions, expectations, and emotion-related measures, is that they reside in the body/mind-body of the individual; most could be sequelae of the neurologic insult or be premorbidly present but impairments all the same.86 Predictors that capture sociocognitive factors such as self-efficacy may also provide important insight into the recovery of voluntary arm use in daily life.87–89 However, we did not find any studies using self-efficacy as a predictor measure in their models. Cognition, memory, emotion, and motivation were the sociocognitive factors that were observed and used to predict upper extremity outcomes at follow-up (Table 6).
At follow-up, most outcome measures were found to be in the domain of body functions/structure. Motor capability of the contralesional limb in the domain of body functions/structure was described with measures of range of motion, muscle strength, arm movement-related functions, and kinematic analysis (Table 7). Muscle strength was found to be the most popular outcome measure in this review. Twenty-two studies (39%) considered hand or arm muscle strength using the manual muscle test,39,58 hand-held dynamometry,67,70,90 the Medical Research Council guidelines,40,43–45,52,59,80,81 the Motricity Index,22,41,46,47,66,76,85 the Canadian Neurological Scale (distal arm),48 or the NIH Stroke Scale (arm).57 Although the kinematic measures are not as convenient for clinical assessments, they provide specific information about qualitative features of upper extremity movements, such as accuracy, efficiency, and speed79 and may be sensitive to motor capacity changes.
Few outcome measures were found to be in the domains of activities and participation. Unlike the muscle strength measures, the Fugl-Meyer Assessment (FMA) upper extremity motor score includes both body functions/structures and activities domains. The FMA was widely used in the studies (n = 14) reviewed here.19,42,49–51,54,60,61,67,68,75,76,83,84 However, none of the measures were considered representative of the participation domain. In the context of activities, the MAL is the only measure useful to quantify how much and how well the contralesional arm is used voluntarily in daily life activities. However, the three studies that used the MAL were ranked weak in methodological quality evaluation.67,71,82
Across the 56 studies, the time points of capturing outcomes varied from a few weeks45,82 to more than a year (Table 8).49,81 A six-month observation period was considered sufficient for determination of functional outcomes.91 Approximately 41% (23 of 56) of the selected studies had chosen a six-month period for capturing outcomes.17,19,22,48,50,51,53,56,60,62–64,67–69,70,71,73,76–78,85,92 The quality evaluation of internal, statistical, and external validity is summarized in Table 8.
In the 56 included studies, more than 95% used reliable and valid instruments for predictors and outcome measures (54 and 56 studies, respectively). Approximately 35% (19 of 56) used raters who were unaware of individuals' pathologic information and treatment program arrangement. Use of blinded raters is intended to decrease potential observational bias in data collection.36 In 98% (55 of 56) of the sample, participant dropout information (eg, withdrawal because of death or migration) was provided.
In 45 of the 56 studies, the time point of first obtaining predictor measures was within three months of stroke onset. The timing of this initial evaluation varied from several days (17 studies)7,39,40,43–45,49,50–53,57,58,64,66,68,76 to several months or years (7 studies).17,62,67,69,71,72,74,82,92 Of the studies in which predictor measures were obtained during the first few days after stroke onset, 53% (9 of 17) used TMS-related measures,39,40,43–45,49–52 24% (4 of 17) used imaging-related measures,53,57,58,68 and 24% (4 of 17) employed functional examinations.7,64,66,76 In studies in which the first time predictor measures were obtained more than three months after stroke onset, the study population was likely in the subacute/chronic stage of recovery, and the primary study purpose was to measure responsiveness to various interventions.17,62,67,71,82,92
Eighty-four percent (47 of 56) of the studies provided rationales for the statistical methods of detecting the relationship strength between predictor and outcome measures. The remaining 16% limited analysis to descriptive plots of predictor-outcome relationships. Thirty-four percent (19 of 56) of the studies did not have a sample size sufficient to provide a valid test of the prediction model. Forty-eight percent (27 of 56) of the studies used regression analysis. A control for multicollinearity was observed in only 9 studies of the subgroup of the studies that established the regression models (33%, 9 of 27 studies).44,56,60,62,69,71,74,79,83
Twenty-eight studies provided information about stroke pathology. The inclusion and exclusion criteria were specified in most of the studies (47 of 56). Sixty-eight percent (38 of 56) reported additional medical or paramedical interventions during the period of study. Only one study used cross-validation to establish a prediction model in a second group of poststroke participants.72 Similarly, only one study provided information about the clinical meaningfulness of the predictor and outcome measures.69
Using the best evidence criteria, 20 studies (36%) were retained for further analysis of significant predictors.21,44–47,50,51,53,54,56,60,62,63,69,74,75,79–81,83 The best predictors of arm-specific outcome measures were the initial neurophysiologic factors44–47,50,51,53,54,80,81 and initial motor capability.44,53,56,60,62,63,69,74,75,79,83 Among the neurophysiologic factors, the presence of a motor evoked potential (MEP), MEP amplitude, and MEP latency were the most frequently used variables. All were TMS-related measures.44–47,50,51,84 Lesion location in the brain and neuronal activation in the motor cortex were the predictor measures captured by imaging techniques.53,54,80,81 The most commonly used predictor measures of initial motor capability were deep sensation,60 muscle tone,60,79 active range of motion,62,79 muscle strength,44,56,79 and performance-based measures (eg, the FMA).53,56,60,63,69,74,75,83
Fifty-six of 935 studies met criteria for inclusion in this systematic review. Between 1979 and April 2008, there was a 317% increase in the frequency of published studies designed to determine the critical predictors of voluntary arm recovery after stroke. The frequency of such studies rose from 1.05 per year before 2000 to nearly 4.38 per year after 2000. There has been a profound increase in the prevalence of this kind of clinical research during the last eight years. This increase may be due in part to recent developments of valid and reliable predictor measures of poststroke upper extremity motor function that should prove useful to inform appropriate therapeutic programs. However, it should be noted that for the most part, this review was based primarily on prognostic studies rather than clinical intervention studies. In addition the included research studies were selected from a limited set of databases that did not include dissertations, theses, or conference proceedings, and further, the citations from selected articles were not checked for inclusion. Although clinical practice in neurologic physical therapy is becoming more evidence-based, it is far from routine use of prognostic tests and indicators to prescribe therapeutic programs.
To optimize the predictive value of arm-specific outcomes, 90% of the stroke participants from the selected studies were investigated in the acute/subacute stage. This is likely the case because assessments conducted in later stages of recovery may not provide sufficient predictive value for functional outcomes.56 Most of the studies of stroke rehabilitation recruit persons with either ischemic or hemorrhagic stroke because the outcome response to intervention is somewhat different between these two groups.93 For the purpose of this review, we included both types of stroke as long as the primary lesion was isolated to one hemisphere and resulted in a hemiparetic syndrome. This approach would have been appropriate to test the hypothesis that predictors of upper extremity motor function recovery are different between these two stroke etiologies. However, it is not possible to separate the results by stroke pathology because 54% (30 of 56) of the studies did not report stroke pathology. Only three studies reported inclusion of people with hemorrhagic stroke. Although we did not obtain complete information about the prevalence of each stroke type, it is reasonable to estimate that those with an infarction are much more numerous than those with hemorrhage (approximately 70% and 30%, respectively). According to statistics published by the American Stroke Association in 2008, approximately 10% of ischemic strokes and 40% of hemorrhagic strokes result in death within one month after stroke onset among persons aged 45 to 64 years,94 indicating that a greater number of individuals with infarct strokes than those with hemorrhagic strokes survive the acute phase. Because the clinical symptoms and recovery time course between infarct and hemorrhagic-type strokes are different,27 this information is most helpful here and for future prognostic studies.
This review focused on measurement used to capture voluntary arm use. To better understand arm-specific functional recovery, the ICF was used to categorize both predictors and outcome measures because it has been used widely to characterize health and disability in clinical practice and research through a framework for delivering goal-oriented rehabilitation.4 Recently, Salter et al95–97 selected 20 popular outcome measures with acceptable reliability and validity and further classified them using the ICF. Based on whether these outcomes captured the context of body functions/structures, activities, or participation, they were categorized into only one ICF domain. However, a number of stroke-related measures are designed to capture basic motor capacity and/or general functional recovery. Therefore, to classify a given measure into a single ICF category may be insufficient for fully characterizing the measure. In addition, there are other limitations to this kind of single-domain approach. First, the boundaries between ICF levels are not clear-cut.28 Second, the testing items of an instrument may be used to measure motor capabilities at different ICF levels. For example, the FMA includes items both at the levels of body functions/structures (eg, reflex) and activities (eg, grasp). Therefore, for this review, we allowed a measure to be classified in more than one ICF level. In this way, it provides a more comprehensive analysis of the various measurement instruments used for prognostication of functional upper extremity motor recovery after stroke.
Together, neurophysiologic measures43–47,50,51,53,54,80,81 and initial sensorimotor abilities44,53,56,60,62,63,74,75,79,83 have been shown to be the best predictors of arm-specific outcomes. The neurophysiologic measures provide a useful metric of the integrity of motor pathways, which is a viable index of functional motor performance.42 Additionally, neurophysiologic measures can be gathered in the early poststroke stage and require minimal participant cooperation.98 Measures of initial sensorimotor function provide a direct evaluation of motor behavior and reveal fundamental motor capability consequent to the stroke. These sensorimotor measures may be more sensitive to the effects of specific interventions than the neurophysiologic measures.60 However, it is also the case that a majority of the voluntary sensorimotor function measures are more vulnerable to associated neurologic impairments such as aphasia, apraxia, and neglect than direct neurophysiologic measures. Although the majority of the predictor models tested included a number of predictor variables, only five of 26 included neuroimaging and clinical variables together in their models.42,44,53,61,68 In all cases, the combined predictor variables resulted in a higher percentage of the outcome variance explained.
Unfortunately, there is still considerable controversy pertaining to the magnitude of a clinically meaningful change in the majority of upper extremity motor ability outcome measures. This highlights the importance of MCID determination for studies related to stroke interventions, particularly those pertaining to upper extremity recovery. It is of special note that only one study reviewed here provided information about MCID.69
The general definition of participation in the ICF is “the involvement of an individual in a life situation.”96 More specifically, the participation level of functional upper extremity motor recovery is defined as the involvement of the arms in life situations. In this review, very few predictor or outcome measures were classified at the participation level. Most of the measures pertained to the body functions/structures level. Using measures at the body functions/structures level may gather sufficient information to examine predictors, but does not provide significant information to examine outcome measures. Because voluntary arm use in daily activities is the primary objective for poststroke individuals,5,6 outcome measures classified at the participation level would be expected to provide information about how individuals incorporate their paretic arms in life situations at home or in the community.
Other than outcome measures that capture motor capability level of arm-specific function in daily life,7,20,21,75 the MAL is a useful measure to quantify how much and how well the contralesional arm is used in daily life activities. The MAL has been shown to have high internal consistency (Cronbach α ≥ 0.88) and moderate construct validity (Spearman ρ = 0.63) in persons with chronic stroke.12 In addition, the MAL has good interrater reliability (interclass correlation coefficient = 0.90–0.94).12 This review identified three studies that used either raw or change scores for the MAL posttreatment as an outcome measure to investigate the predictors of functional upper extremity motor recovery after Constraint-induced movement therapy in individuals more than six months poststroke.67,71,82 The potential predictors investigated in these three studies were individual/stroke characteristics (eg, side of stroke location, time since stroke, hand dominance, age, sex, and ambulatory status);71 persons' cognitive functions (measured by the Mini-Mental State Examination, the short-form Token Test, the Sustained Attention to Response Task, Logical Memory and Visual Reproduction subtests from the Wechsler Memory Scale, and the Trail Making Test Form B);82 and upper extremity motor severity (measured by the FMA).67
Interestingly, measures of individual/stroke characteristics and cognitive functions did not reach significance when examining the prediction model of actual arm use measured by the MAL.71,82 However, motor severity classified by the FMA was correlated with the changed MAL score after a community-based upper extremity group exercise among individuals with chronic stroke (mean poststroke duration 5.1 years).67 Persons in the mildly impaired group gained 1 point on the MAL, whereas persons in the moderately/severely impaired groups gained 0.2 to 0.5 points on the MAL.67 It seems that motor severity classified by the FMA is a better discriminator than the MAL in predicting which person might benefit the most from the community-based group exercise. Yet it is not clear whether motor severity captured in the acute stroke stage has good predictive value in relation to outcome measures of voluntary arm use. More studies using measures at the activities and participation level are needed to establish cause-and-effect relationships between predictors and outcome measures of voluntary arm use in real-life situations.
This is the first systematic review focused on critical predictors of arm-specific motor recovery that incorporated the ICF for meaningful outcomes of voluntary arm recovery in hemiparetic stroke. Until recently, clinical research has paid little attention to the more distal outcomes represented by the activities and participation categories in the ICF. Not surprisingly, initial measures that capture the integrity of neural connections (eg, in the corticospinal tract) and voluntary motor behavior were the best predictors of functional arm recovery at follow-up. This finding supports the usefulness of a top-down approach in which task-oriented training programs are developed and aimed at reducing the functional limitations apparent in contralesional limb use.99 The fact that few outcome measures for voluntary arm use exist at the participation level of the ICF underscores the need to develop reliable and valid measures of arm use in real-life environments.
The authors thank Pamela Corley for her help in developing the keywords/databases search and Dr. Sharon Myers, Jill Stewart, and Hsiu-Chen Lin for their suggestions or comments on the preparation of the manuscript.
1. Rosamond W, Flegal K, Furie K, et al. Heart disease and stroke statistics–2008 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2008;117:e25–e146.
2. Mayo NE, Wood-Dauphinee S, Ahmed S, et al. Disablement following stroke. Disabil Rehabil. 1999;21:258–268.
3. Mayo NE, Wood-Dauphinee S, Cote R, et al. Activity, participation, and quality of life 6 months poststroke. Arch Phys Med Rehabil. 2002;83:1035–1042.
4. World Health Organization. ICF: International Classification of Functioning, Disability and Health. Geneva: World Health Organization; 2001.
5. Barker RN, Brauer SG. Upper limb recovery after stroke: the stroke survivors' perspective. Disabil Rehabil. 2005;27:1213–1223.
6. Barker RN, Gill TJ, Brauer SG. Factors contributing to upper limb recovery after stroke: a survey of stroke survivors in Queensland Australia. Disabil Rehabil. 2007;29:981–989.
7. Nakayama H, Jorgensen HS, Raaschou HO, et al. Recovery of upper extremity function in stroke patients: the Copenhagen Stroke Study. Arch Phys Med Rehabil. 1994;75:394–398.
8. Uswatte G, Giuliani C, Winstein C, et al. Validity of accelerometry for monitoring real-world arm activity in patients with subacute stroke: evidence from the extremity constraint-induced therapy evaluation trial. Arch Phys Med Rehabil. 2006;87:1340–1345.
9. Taub E, Miller NE, Novack TA, et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil. 1993;74:347–354.
10. Taub E, Crago JE, Uswatte G. Constraint induced movement therapy: a new approach to treatment in physical rehabilitation. Rehabil Psvchol. 1998;43:152–170.
11. Uswatte G, Taub E, Morris D, et al. The motor activity log-28: assessing daily use of the hemiparetic arm after stroke. Neurology. 2006;67:1189–1194.
12. van der Lee JH, Beckerman H, Knol DL, et al. Clinimetric properties of the motor activity log for the assessment of arm use in hemiparetic patients. Stroke. 2004;35:1410–1414.
13. Carod-Artal FJ, Gonzalez-Gutierrez JL, Herrero JA, et al. Functional recovery and instrumental activities of daily living: follow-up 1-year after treatment in a stroke unit. Brain Inj. 2002;16:207–216.
14. Frankel MR, Morgenstern LB, Kwiatkowski T, et al. Predicting prognosis after stroke: a placebo group analysis from the National Institute of Neurological Disorders and Stroke rt-PA Stroke Trial. Neurology. 2000;55:952–959.
15. Fullerton KJ, Mackenzie G, Stout RW. Prognostic indices in stroke. Q J Med. 1988;66:147–162.
16. Kwakkel G, Wagenaar RC, Kollen BJ, et al. Predicting disability in stroke–a critical review of the literature. Age Ageing. 1996;25:479–489.
17. Wolf SL, Baker MP, Kelly JL. EMG biofeedback in stroke: effect of patient characteristics. Arch Phys Med Rehabil. 1979;60:96–102.
18. De Souza LH, Hewer RL, Miller S. Assessment of recovery of arm control in hemiplegic stroke patients, Part 1: Arm function tests. Int Rehabil Med. 1980;2:3–9.
19. Kusoffsky A, Wadell I, Nilsson BY. The relationship between sensory impairment and motor recovery in patients with hemiplegia. Scand J Rehabil Med. 1982;14:27–32.
20. La Joie WJ, Reddy NM, Melvin JL. Somatosensory evoked potentials: their predictive value in right hemiplegia. Arch Phys Med Rehabil. 1982;63:223–226.
21. Dudgeon BJ, DeLisa JA, Miller RM. Optokinetic nystagmus and upper extremity dressing independence after stroke. Arch Phys Med Rehabil. 1985;66:164–167.
22. Parker VM, Wade DT, Langton HR. Loss of arm function after stroke: measurement, frequency, and recovery. Int Rehabil Med. 1986;8:69–73.
23. Kruger E, Teasell R, Salter K, et al. The rehabilitation of patients recovering from brainstem strokes: case studies and clinical considerations. Top Stroke Rehabil. 2007;14:56–64.
24. Teasell R, Foley N, Doherty T, et al. Clinical characteristics of patients with brainstem strokes admitted to a rehabilitation unit. Arch Phys Med Rehabil. 2002;83:1013–1016.
25. Mosch SC, Max JE, Tranel D. A matched lesion analysis of childhood versus adult-onset brain injury due to unilateral stroke: another perspective on neural plasticity and recovery of social functioning. Cogn Behav Neurol. 2005;18:5–17.
26. Umphred D, ed. Neurological Rehabilitation. 5th ed. St. Louis: Mosby; 2007.
27. Hier DB, Edelstein G. Deriving clinical prediction rules from stroke outcome research. Stroke. 1991;22:1431–1436.
28. Wade DT. Measurement in Neurological Rehabilitation. New York: Oxford University Press; 1992.
29. Keith RA, Granger CV, Hamilton BB, et al. The functional independence measure: a new tool for rehabilitation. Adv Clin Rehabil. 1987;1:6–18.
30. Collin C, Wade DT, Davies S, et al. The Barthel ADL Index: a reliability study. Int Disabil Stud. 1988;10:61–63.
31. Rose L, Bakal DA, Fung TS, et al. Tactile extinction and functional status after stroke. A preliminary investigation. Stroke. 1994;25:1973–1976.
32. Woldag H, Gerhold LL, de Groot M, et al. Early prediction of functional outcome after stroke. Brain Inj. 2006;20:1047–1052.
33. Arac N, Sagduyu A, Binai S, et al. Prognostic value of transcranial magnetic stimulation in acute stroke. Stroke. 1994;25:2183–2186.
34. Chira-Adisai W, Yan K, Shahani BT. Changes in serial correlation coefficients and fractional parameters during functional recovery in stroke patients. Electromyogr Clin Neurophysiol. 2001;41:79–86.
35. Lyden PD, Lau GT. A critical appraisal of stroke evaluation and rating scales. Stroke. 1991;22:1345–1352.
36. Portney LG, Watkins MP. Foundations of Clinical Research: Applications to Practice. 3rd ed. Upper Saddle River, NJ: Pearson Education; 2009.
37. Symposium recommendations for methodology in stroke outcome research. Task Force on Stroke Impairment, Task Force on Stroke Disability, and Task Force on Stroke Handicap. Stroke. 1990;21(9 Suppl):II68–II73.
38. Hsieh YW, Wang CH, Wu SC, et al. Establishing the minimal clinically important difference of the Barthel Index in stroke patients. Neurorehabil Neural Repair. 2007;21:233–238.
39. Nagao S, Kawai N. Prediction of motor function by magnetic brain stimulation in patients with intracerebral hematoma. Neurol Med Chir (Tokyo). 1992;32:268–274.
40. Pennisi G, Rapisarda G, Bella R, et al. Absence of response to early transcranial magnetic stimulation in ischemic stroke patients: prognostic value for hand motor recovery. Stroke. 1999;30:2666–2670.
41. Keren O, Ring H, Solzi P, et al. Upper limb somatosensory evoked potentials as a predictor of rehabilitation progress in dominant hemisphere stroke patients. Stroke. 1993;24:1789–1793.
42. Stinear CM, Barber PA, Smale PR, et al. Functional potential in chronic stroke patients depends on corticospinal tract integrity. Brain. 2007;130(Part 1):170–180.
43. Delvaux V, Alagona G, Gerard P, et al. Post-stroke reorganization of hand motor area: a 1-year prospective follow-up with focal transcranial magnetic stimulation. Clin Neurophysiol. 2003;114:1217–1225.
44. Nascimbeni A, Gaffuri A, Imazio P. Motor evoked potentials: prognostic value in motor recovery after stroke. Funct Neurol. 2006;21:199–203.
45. Rapisarda G, Bastings E, de Noordhout AM, et al. Can motor recovery in stroke patients be predicted by early transcranial magnetic stimulation? Stroke. 1996;27:2191–2196.
46. Dachy B, Biltiau E, Bouillot E, et al. Facilitation of motor evoked potentials in ischemic stroke patients: prognostic value and neurophysiologic correlations. Clin Neurophysiol. 2003;114:2370–2375.
47. Dachy B, Denis L, Deltenre P. Usefulness of transcranial magnetic stimulation to predict the development of reflex sympathetic dystrophy poststroke: a pilot study. Arch Phys Med Rehabil. 2002;83:730–732.
48. Cruz Martinez A, Tejada J, Diez Tejedor E. Motor hand recovery after stroke. Prognostic yield of early transcranial magnetic stimulation. Electromyogr Clin Neurophysiol. 1999;39:405–410.
49. Hendricks HT, Hageman G, van Limbeek J. Prediction of recovery from upper extremity paralysis after stroke by measuring evoked potentials. Scand J Rehabil Med. 1997;29:155–159.
50. Hendricks HT, Pasman JW, Merx JL, et al. Analysis of recovery processes after stroke by means of transcranial magnetic stimulation. J Clin Neurophysiol. 2003;20:188–195.
51. Hendricks HT, Pasman JW, van Limbeek J, et al. Motor evoked potentials in predicting recovery from upper extremity paralysis after acute stroke. Cerebrovasc Dis. 2003;16:265–271.
52. Tzvetanov P, Rousseff RT, Atanassova P. Prognostic value of median and tibial somatosensory evoked potentials in acute stroke. Neurosci Lett. 2005;380:99–104.
53. Binkofski F, Seitz RJ, Hacklander T, et al. Recovery of motor functions following hemiparetic stroke: a clinical and magnetic resonance-morphometric study. Cerebrovasc Dis. 2001;11:273–281.
54. Schiemanck SK, Kwakkel G, Post MWM, et al. Impact of internal capsule lesions on outcome of motor hand function at one year post-stroke. J Rehabil Med. 2008;40:96–101.
55. Higgins J, Mayo NE, Desrosiers J, et al. Upper-limb function and recovery in the acute phase poststroke. J Rehabil Res Dev. 2005;42:65–76.
56. Kwakkel G, Kollen BJ, van der Grond J, et al. Probability of regaining dexterity in the flaccid upper limb: impact of severity of paresis and time since onset in acute stroke. Stroke. 2003;34:2181–2186.
57. Karibe H, Shimizu H, Tominaga T, et al. Diffusion-weighted magnetic resonance imaging in the early evaluation of corticospinal tract injury to predict functional motor outcome in patients with deep intracerebral hemorrhage. J Neurosurg. 2000;92:58–63.
58. Yoshioka H, Horikoshi T, Aoki S, et al. Diffusion tensor tractography predicts motor functional outcome in patients with spontaneous intracerebral hemorrhage. Neurosurgery. 2008;62:97–103, discussion 103.
59. Matteis M, Vernieri F, Troisi E, et al. Early cerebral hemodynamic changes during passive movements and motor recovery after stroke. J Neurol. 2003;250:810–817.
60. Feys H, De Weerdt W, Nuyens G, et al. Predicting motor recovery of the upper limb after stroke rehabilitation: value of a clinical examination. Physiother Res Int. 2000;5:1–18.
61. Feys H, Hetebrij J, Wilms G, et al. Predicting arm recovery following stroke: value of site of lesion. Acta Neurol Scand. 2000;102:371–377.
62. Fritz SL, Light KE, Patterson TS, et al. Active finger extension predicts outcomes after constraint-induced movement therapy for individuals with hemiparesis after stroke. Stroke. 2005;36:1172–1177.
63. Kwakkel G, van Dijk GM, Wagenaar RC. Accuracy of physical and occupational therapists' early predictions of recovery after severe middle cerebral artery stroke. Clin Rehabil. 2000;14:28–41.
64. Meldrum D, Pittock SJ, Hardiman O, et al. Recovery of the upper limb post ischaemic stroke and the predictive value of the Orpington Prognostic Score. Clin Rehabil. 2004;18:694–702.
65. Katrak P, Bowring G, Conroy P, et al. Predicting upper limb recovery after stroke: the place of early shoulder and hand movement. Arch Phys Med Rehabil. 1998;79:758–761.
66. Roy CW, Sands MR, Hill LD, et al. The effect of shoulder pain on outcome of acute hemiplegia. Clin Rehabil. 1995;9:21–27.
67. Pang MY, Harris JE, Eng JJ. A community-based upper-extremity group exercise program improves motor function and performance of functional activities in chronic stroke: a randomized controlled trial. Arch Phys Med Rehabil. 2006;87:1–9.
68. Prabhakaran S, Zarahn E, Riley C, et al. Inter-individual variability in the capacity for motor recovery after ischemic stroke. Neurorehabil Neural Repair. 2008;22:64–71.
69. Fritz SL, George SZ, Wolf SL, et al. Participant perception of recovery as criterion to establish importance of improvement for constraint-induced movement therapy outcome measures: a preliminary study. Phys Ther. 2007;87:170–178.
70. Pandyan AD, Cameron M, Powell J, et al. Contractures in the post-stroke wrist: a pilot study of its time course of development and its association with upper limb recovery. Clin Rehabil. 2003;17:88–95.
71. Fritz SL, Light KE, Clifford SN, et al. Descriptive characteristics as potential predictors of outcomes following constraint-induced movement therapy for people after stroke. Phys Ther. 2006;86:825–832.
72. Barreca SR, Finlayson MA, Gowland CA, et al. Use of the Halstead Category Test as a cognitive predictor of functional recovery in the hemiplegic upper limb: a cross-validation study. Clin Neuropsychol. 1999;13:171–181.
73. Canning CG, Ada L, Adams R, et al. Loss of strength contributes more to physical disability after stroke than loss of dexterity. Clin Rehabil. 2004;18:300–308.
74. Loewen SC, Anderson BA. Predictors of stroke outcome using objective measurement scales. Stroke. 1990;21:78–81.
75. Hatakenaka M, Miyai I, Sakoda S, et al. Proximal paresis of the upper extremity in patients with stroke. Neurology. 2007;69:348–355.
76. Smania N, Paolucci S, Tinazzi M, et al. Active finger extension: a simple movement predicting recovery of arm function in patients with acute stroke. Stroke. 2007;38:1088–1090.
77. Katrak PH. Shoulder shrug—a prognostic sign for recovery of hand movement after stroke. Med J Aust. 1990;152:297–301.
78. Sunderland A, Tinson D, Bradley L, et al. Arm function after stroke. An evaluation of grip strength as a measure of recovery and a prognostic indicator. J Neurol Neurosurg Psychiatry. 1989;52:1267–1272.
79. Wagner JM, Lang CE, Sahrmann SA, et al. Sensorimotor impairments and reaching performance in subjects with poststroke hemiparesis during the first few months of recovery. Phys Ther. 2007;87:751–765.
80. Jang SH, Cho SH, Kim YH, et al. Diffusion anisotrophy in the early stages of stroke can predict motor outcome. Restor Neurol Neurosci. 2005;23:11–17.
81. Wenzelburger R, Kopper F, Frenzel A, et al. Hand coordination following capsular stroke. Brain. 2005;128(Part 1):64–74.
82. Mark VW, Woods AJ, Mennemeier M, et al. Cognitive assessment for CI therapy in the outpatient clinic. NeuroRehabilitation. 2006;21:139–146.
83. Shelton FD, Volpe BT, Reding M. Motor impairment as a predictor of functional recovery and guide to rehabilitation treatment after stroke. Neurorehabil Neural Repair. 2001;15:229–237.
84. Sonde L, Bronge L, Kalimo H, et al. Can the site of brain lesion predict improved motor function after low-TENS treatment on the post-stroke paretic arm? Clin Rehabil. 2001;15:545–551.
85. Jang SH, Kim YH, Chang Y, et al. The predictive value of cortical activation by passive movement for motor recovery in stroke patients. Restor Neurol Neurosci. 2004;22:59–63.
86. Wade D. Why physical medicine, physical disability and physical rehabilitation? We should abandon Cartesian dualism. Clin Rehabil. 2006;20:185–190.
87. Allen DD. Proposing 6 dimensions within the construct of movement in the movement continuum theory. Phys Ther. 2007;87:888–898, discussion 925–834.
88. Allen DD. Validity and reliability of the movement ability measure: a self-report instrument proposed for assessing movement across diagnoses and ability levels. Phys Ther. 2007;87:899–916, discussion 925–834.
89. Allen DD. Responsiveness of the movement ability measure: a self-report instrument proposed for assessing the effectiveness of physical therapy intervention. Phys Ther. 2007;87:917–924, discussion 925–934.
90. Bohannon RW, Smith MB. Upper extremity strength deficits in hemiplegic stroke patients: relationship between admission and discharge assessment and time since onset. Arch Phys Med Rehabil. 1987;68:155–157.
91. Jorgensen HS, Nakayama H, Raaschou HO. Neurologic and functional recovery the Copenhagen Stroke Study. Phys Med Rehabil Clin N Am. 1999;10:887–906.
92. Wolf SL, Binder-MacLeod SA. Electromyographic biofeedback applications to the hemiplegic patient. Changes in upper extremity neuromuscular and functional status. Phys Ther. 1983;63:1393–1403.
93. Paolucci S, Antonucci G, Grasso MG, et al. Functional outcome of ischemic and hemorrhagic stroke patients after inpatient rehabilitation: a matched comparison. Stroke. 2003;34:2861–2865.
94. Rosamond WD, Folsom AR, Chambless LE, et al. Stroke incidence and survival among middle-aged adults: 9-year follow-up of the Atherosclerosis Risk in Communities (ARIC) cohort. Stroke. 1999;30:736–743.
95. Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF Body Functions. Disabil Rehabil. 2005;27:191–207.
96. Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF Participation. Disabil Rehabil. 2005;27:507–528.
97. Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF activity. Disabil Rehabil. 2005;27:315–340.
98. Hendricks HT, Zwarts MJ, Plat EF, et al. Systematic review for the early prediction of motor and functional outcome after stroke by using motor-evoked potentials. Arch Phys Med Rehabil. 2002;83:1303–1308.
99. Winstein CJ, Wolf SL. Task-oriented training to promote upper extremity recovery. In: Stein J, Harvey RL, Macko RF, et al, eds. Stroke Recovery & Rehabilitation. New York: Demos Medical Publishing; 2009:267–290.
voluntary arm use; prediction; clinical meaningfulness
© 2009 Neurology Section, APTA
Highlight selected keywords in the article text.