Sage Journals: Discover world-class research

Abstract

Background and purpose. Augmenting changes in recovery is core to the rehabilitation process following a stroke. Hence it is essential that outcome measures are able to detect change as it occurs, a property known as responsiveness. This article critically reviewed the responsiveness of functional outcome measures following stroke, specifically examining tools that captured upper-extremity (UE) functional recovery. Methods. A systematic search of the literature was undertaken to identify articles providing responsiveness data for 3 types of change (observed, detectable, and important). Results. Data from 68 articles for 14 UE functional outcome measures were retrieved. Larger percentage changes were required to be considered important when obtained through anchor-based methods (eg, based on patient opinion or comparative measure) compared with distribution methods (eg, statistical estimates). Larger percentage changes were required to surpass the measurement error for patient-perceived functional measures (eg, Motor Activity Log) compared with laboratory-based performance measures (eg, Action Research Arm Test). The majority of rehabilitation interventions have similar effect sizes on patient-perceived UE function and laboratory-based UE function. Conclusions. The magnitude of important change or change that surpasses measurement error can vary substantially depending on the method of calculation. Rehabilitation treatments can affect patient perceptions of functional change as effectively as laboratory-based functional measures; however, larger sample sizes may be required to account for the larger measurement error associated with patient-perceived functional measures.

Keywords

stroke systematic review recovery of function responsiveness treatment outcome rehabilitation UE

Introduction

Functional recovery following stroke is complex, with wide variation in natural recovery and response to treatment across individuals. Optimizing or augmenting changes in recovery is core to the rehabilitation process following a stroke. Hence it is essential that outcome measures are able to detect change as it occurs, a property known as responsiveness.^1,2

Detecting change over time or from an intervention is one of the most critical requisites of an outcome measure; it is necessary information for selecting the best instrument for practice or research and for determining sample size for clinical studies. Furthermore, there is growing recognition that traditional reporting of statistical significance tests and effects sizes should be accompanied by methods for determining meaningful or important change.³

Beaton et al¹ argue that responsiveness is a context-specific characteristic that is influenced by factors such as the specific sample, treatment, and the type of change captured by an instrument. The authors outline 3 major types of change in their responsiveness taxonomy. They are as follows: (1) observed change, (2) important change, and (3) detectable change. Figure 1 depicts the interrelationship between the 3 types of change and the common metrics used to quantify them. Observed change is the amount of change observed in a population in which change is expected to occur (ie, after a treatment of known efficacy or a specific period within the natural recovery pattern). Traditional methods have captured this type of change with an effect size.¹ Important change is the observed change estimated to be meaningful and is often quantified as a minimally clinically important difference (MCID) value. For example, important change may reflect the value that patients, clinicians, or society places on the recovery (or partial recovery) of a task, such as the ability to bring a spoon to the mouth. Finally, detectable change takes into consideration the measurement error associated with a tool and is often quantified as a minimal detectable change score (MDC) or limit of agreement. Calculation of the MDC value associated with a measure varies depending on the confidence level selected. The most common confidence levels selected are 95% and 90% and are denoted by the subscripts MDC₉₅ and MDC₉₀. Confidence that true functional change has been observed in a clinical study is increased when the observed change is equal to or surpasses a measure’s detectable change values.¹

Figure 1.

Graphic representation of observed, detectable, and important change. Abbreviations: MDC₉₀, minimal detectable change (with 90% confidence interval); MDC₉₅, minimal detectable change (with 95% confidence interval); LOA, limits of agreement; PES, population effect size; SRM, standardized response mean; MCID, minimally clinically important difference.

The purpose of this article is to synthesize and critically review the research evidence that captures responsiveness as defined by 3 types of change (observed, important, and detectable). This systematic review provides an understanding of the responsiveness of outcome measures used in stroke research, specifically within the context of upper-extremity (UE) functional recovery.

Methods

Literature Search

This review targeted articles that capture 3 types of change (observed, important, and detectable) using UE functional outcome measures following a stroke. Articles were identified using a systematic search of electronic databases (MEDLINE, EMBASE, CINAHL, PsycINFO, Cochrane Central) from database inception through March 2012. The following keywords were used: stroke, cerebrovascular accident or hemiplegia or hemiparesis combined with upper extremity, function or activities of daily living and responsiveness, reliability, psychometrics, “minimally clinically important change or MCID,” “standard error of measurement or SEM,” “minimal detectable change or MDC,” “standardized response mean or SRM,” and effect size or outcome measurement. All terms were mapped onto subject headings. Articles were limited to the English language, human subjects, and adults. A hand search of reference lists from reviews and the gray literature (eg, StrokEngine Assess) was also conducted to ensure a thorough search.⁴

Inclusion Criteria

To be included in this review, articles had to meet the following inclusion criteria: they had to (1) provide a responsiveness index or sufficient information to calculate responsiveness index (eg, MDC or effect size); (2) use a sample of individuals with stroke; (3) use an outcome measure that assessed affected UE ability and included ≥50% functional activities (to ensure scale reflected International Classification of Functioning, Disability, and Health activity domain); and (4) be published in a peer-reviewed journal. Articles were excluded if they were (1) conference proceedings or abstracts; (2) pre-post studies or randomized control trials (RCTs), which calculated a responsiveness index without using the control group information to minimize bias for inflation of the effect sizes⁵; (3) studies with a sample size less than 10; and (4) studies that used measures that captured UE function as a single action (eg, box and blocks test, peg test). Single actions were not considered representative of the many actions involved in UE function. RCTs were excluded if they (1) used <2 UE functional outcome measures or (2) found no significant effects for UE function.

Data Extraction and Organization

Responsiveness indices were extracted or calculated from the included articles and subsequently organized into the 3 categories of change outlined by Beaton et al¹: (1) observed change, (2) important change, and (3) detectable change (Figure 1). Observed change was further subdivided into (1) change over natural recovery (categorized into <3 months postinjury and ≥3 months postinjury) and (2) change in response to an effective treatment. Effect sizes were calculated based on the change score divided by the baseline standard deviation, and MDC values (MDC₉₀, MDC₉₅) were calculated based on the test-retest coefficient (intraclass correlation coefficient) and the baseline standard deviation.⁶ In addition, MDC₉₅%, which are independent of measurement units, were calculated to compare MDC values across measures.⁷ We present the MDC₉₅% using the 2 methods commonly cited in the rehabilitation literature: (1) MDC₉₅/maximum score for the scale (eg, 57 points for Action Research Arm Test [ARAT]) and (2) MDC₉₅/baseline mean of the sample.

RCTs that used at least 2 different functional outcome measures were used to estimate observed change in response to a treatment of known efficacy. This allowed for comparison of observed change across different measures within the same study by controlling for variation in treatment and sample characteristics.⁸ To ensure that we were capturing observed change in response to an effective treatment, only RCTs that found a significant effect for at least one of the UE functional measures were included. Scatter plots of effect sizes were generated for a visual representation of one outcome measures’ relative ability to capture change compared with another measure in response to the same treatment with the same sample. When only median and range scores were provided, mean and standard deviation values were estimated using the method suggested by Hozo et al⁹ to calculate an effect size.

Results

The search strategy yielded 1770 titles of which 68 met the inclusion criteria (Figure 2). The articles provided responsiveness data for 14 functional outcome measures.^10-77 Table 1 displays the number of articles that provided responsiveness estimates for each measure. The ARAT, Motor Activity Log (MAL), Wolf Motor Function test (Wolf), and Stroke Impact Scale (SIS) were the 4 most frequently used measures among the included articles.

Figure 2.

Flow diagram of process to select final list of included articles and outcome measures. Abbreviations: OM, outcome measure; AMAT, Arm Motor Activity Test; ARAT, Action Research Arm Test; CAHAI, Chedoke Arm and Hand Activity Inventory; Duruoz, Duruoz Hand Index; Frenchay, Frenchay Arm Test; FTHUE, Functional Test for the Hemiplegic Upper Extremity; Jebsen, Jebsen Hand Function Test; MAL, Motor Activity Log; SIS, Stroke Impact Scale; TEMPA, Upper Extremity Performance Scale for the Elderly; Wolf, Wolf Motor Function Test.

Table 1.

MCID Values Calculated by Distribution and Anchor-Based Methods^a

	ABILHAND	Acceler	ARAT	MAL QOM	MAL AOU	SIS	Wolf FAS	Wolf Time
MCID obtained by distribution-based methods^b
Effect size	0.26⁶⁵ (in logits)					5.8⁶⁶ (5.8%)	0.14⁶⁷ (2.8%)	1.4⁶⁷ (1.1%)
MCID obtained by anchor-based methods^c
Fugl-Meyer							0.33⁶⁷ (6.6%)	1.64⁶⁷ (1.4%)
Percentage recovery	0.35⁶⁵ (In logits)				NS⁶³	17.8⁶⁶ (17.8%)		NS⁶³
Global		NS⁶⁴	12-17⁶⁴ (21%-30%)	1.0-1.1⁶⁴ (20%-22%)			1.0-1.2⁶⁴ (20%-24%)	19.0⁶⁴ (16%)

Abbreviations: MCID, minimally clinically important difference; Acceler, Accelerometry; ARAT, Action Research Arm Test; MAL, Motor Activity Log; QOM, Quality of Movement Scale; AOU, Amount of Use Scale; SIS, Stroke Impact Scale; Wolf, Wolf Motor Function Test; FAS, Functional Ability Scale; NS, MCID values could not be calculated because there was a relationship between change scores and global recovery scales; logits, log odds scale units that allow Likert scale scores (ie, ABILHAND raw scores) to be interpreted as interval scores.⁷⁸

MCID values are displayed in raw scores (percentage of scale maximum score).

An effect size of 0.2 was used as the distribution method to determine MCID.

Anchor-based methods used to calculate MCID consisted of: Fugl-Meyer, change of 6 to 10 points on the Fugl-Meyer upper-extremity scale; percentage recovery, 10% to 15% or 50% recovery on the SIS global recovery item; global, perception of important change on a global rating scale.

Observed Change

Observed change over natural recovery

Effect sizes that captured change over time of participants in standard care were extracted or calculated from 25 studies.^10-34 It was found that 14 studies followed participants up to 3 months poststroke, and 10 studies followed participants up to 6 months poststroke. Because only 1 study followed participants past 6 months poststroke, we compared the effect sizes for participants at less than 3 months with those at greater than or equal to 3 months postinjury. The majority of studies had observation periods from 2 to 5 months.

Effect sizes calculated at a baseline of 1 to 3 months postinjury were larger and showed greater variance than effect sizes calculated at a baseline ≥3 months postinjury (Figure 3). This finding was evident when the duration of follow-up was similar between the 2 groups. Two studies^17,25 considered the effect of stroke severity on observed changes over recovery; effect sizes were 2.0 to 2.6 times larger for individuals with less severe impairments at 1 to 2 months poststroke at study baseline. Three studies^13,31,33 calculated effect sizes using 2 methods: population effect size (based on the change score divided by baseline standard deviation) and standardized response mean (based on the change score divided by the change score standard deviation) for the same measure. In all 3 studies, the effect sizes calculated as a standardized response mean were larger than when calculated as a population effect size (Supplementary Table 2).

Figure 3.

Effect sizes by measure calculated at <3 months and ≥3 months poststroke. The full range of the effect sizes for the Frenchay is 0.2 to 5. The bars on the graph represent the range of effect sizes calculated from studies that measured UE function across time.

Observed change in response to a treatment of known efficacy

A total of 28 RCTs used more than 1 UE functional outcome measure and obtained a significant effect for at least 1 of these measures. ^35-62 These 28 RCTs were used to examine observed change of different measures in response to a treatment of known efficacy within the same study. Constraint-induced movement therapy (CIMT) was the most frequently studied intervention among the included RCTs (50%). Also, the majority of these studies (n = 23) used a laboratory-based performance measure (eg, ARAT, Wolf) in addition to a measure that captured perceived function in one’s own environment (eg, MAL, SIS). The most common outcome measures used together were the MAL with the ARAT (used in 7 studies)^35-41 or MAL with the Wolf (used in 10 studies).^42-51

The majority of effect sizes demonstrated close to a 1:1 relationship between the laboratory-based and patient-perceived functional measures within the same study (Figure 4). Effect sizes from studies that investigated CIMT^36-38,46,47 and 1 study that examined the effect of mirror therapy for individuals with Complex Regional Pain Syndrome⁴³ did not demonstrate this 1:1 relationship. Effect sizes for the perceived effect (MAL) were 1.6 to 6.2 times larger than the functional changes (measured by ARAT or Wolf) in these studies.

Figure 4.

Comparison of observed change captured by lab-based versus patient-perceived functional measures. Points on the graph represent the effect sizes obtained from a single study. Lines on the graph represent a 1:1 relationship between the laboratory-based and patient-perceived functional measures. Laboratory-based measures are located on the x axes (ie, Wolf, ARAT). Patient-perceived functional measures are located on the y axes (ie, MAL).

Important Change

Five studies established important change (MCID) values for 6 functional outcome measures (see Table 1).^63-67 All but 1 study calculated MCID values from CIMT trials.^63,64,66,67 The other study investigated change following robotic therapy.⁶⁵ Four studies^63,65-67 used individuals with chronic stroke, and 1 study sample consisted of individuals with acute stroke.⁶⁴

Three studies^65-67 used a combination of anchor and distribution-based techniques, whereas 2 studies^63,64 used only anchor-based methods to define important change. Anchor-based methods compare the change scores on the measure of interest with a comparative measure or anchor of important change.⁷⁹ The following anchors were used in the 5 studies: (1) a predetermined level on a global rating scale in which participants were asked to rate their perception of functional change⁶⁴; (2) a predetermined level of recovery on the SIS global recovery item^63,65,66; and (3) a change score of 6 to 10 points on the UE portion of the Fugl-Meyer assessment.⁶⁷ Distribution-based methods determine important change based on the statistical distribution of the results.⁷⁹ Three studies used 0.2 times the standard deviation of the sample (ie, effect size of 0.2) to determine MCID values.^65-67

Important change values displayed large variation, with values spanning from 1.1% to 30% of the tests’ maximum scores. The largest values were observed in the study that used a sample of individuals in the acute stage poststroke.⁶⁴ In addition, MCID values calculated using statistically derived distribution-based methods were substantially smaller (15%-88% less) than the values determined by anchor-based methods (Table 1). For example, the MCID for the SIS-hand ranged from a value of 5.8 using an effect size method (distribution method) to 17.8 using perceived amount of recovery on the global recovery question (participant-perceived anchor-based method) for the same sample of individuals.⁶⁶

Three studies were not able to establish MCID values because of nonsignificant relationships between the global recovery/rating scales and functional changes using the MAL,^63,76 Wolf (time component),^63,64 and accelerometry⁶⁴ (Table 1).

Detectable Change

Values needed to surpass measurement error, which are considered to represent true functional change (SEM, MDC₉₀, MDC₉₅, and limit of agreement values), were extracted or calculated from 16 studies for 9 measures.^{12,14,15,27,66-77} All studies used a test-retest methodology in which time frames in between assessments ranged from 1 day to 2 weeks; 9 studies used individuals with subacute stroke,^{14,15,27,68-70,72-74} and 7 studies used patients with chronic stroke.^{12,66,67,71,75-77} Also, 9 studies used a subsample of individuals from CIMT/forced-use trials^66-72,76,77 of which 4 were from the EXCITE trial.^68-70,72

MDC values at the 90% and 95% confidence levels ranged from 1.0% and 1.2% of the maximum score for the AMAT time subscale to 21.9% and 25.9% of the maximum score for the SIS hand, respectively. Relative to their sample means, MDC values at the 90% and 95% confidence levels ranged from 11.5% and 13.7% for the Wolf functional ability subscale to 72.5% and 86.7% for the MAL amount of use subscale, respectively. Detailed SEM, MDC₉₀, and MDC₉₅ estimates and their respective relative percentage values (ie, relative to the sample mean or relative to the scale maximum score) can be found in Supplementary Table 3 available online. Among the measures with multiple estimates, the values needed to surpass measurement error for tools that capture patient-perceived function (eg, MAL, SIS) were larger than laboratory-based performance measures (eg, Wolf, Chedoke Arm and Hand Inventory; Figure 5). This observation was present whether the MDC values were considered relative to the scale maximums or the sample means.

Figure 5.

Comparison of detectable change (calculated at 95% confidence level) relative to the sample means and scale maximums. The bars beside each measure represent the range of MDC₉₅% values extracted or calculated from different studies. *Two studies analyzed different subsets of the same sample to obtain multiple estimates. †Estimates for the ARAT, Wolf (FAS), and SIS are missing from this graph because sample means were not provided in 2 studies.^12,66

Also, the values needed to surpass measurement error for patient-perceived performance measures were larger or on par with important change values. For instance, MDC and important change values for the SIS were as follows: 17.1 to 21.9 (MDC₉₀) and 20.4 to 25.9 (MDC₉₅) versus 5.8 (distribution-based MCID) or 17.8 (anchor-based MCID). MDC and important change values for the MAL were as follows: 0.56 to 1.06 (MDC₉₀) and 0.67 to 1.27 (MDC₉₅) versus 1.0 to 1.1 (anchor-based MCID).

Discussion

Examining the measurement of change in UE function served as a framework for understanding the measurement of functional recovery following stroke and revealed several novel findings related to the ability of outcome measures to capture change as it occurs.

Observed Change

For equal duration of follow-up, the effect sizes for natural recovery calculated at 1 to 3 months poststroke were substantially larger than those calculated at 3 months or later poststroke. The observed differences in effect sizes between these time phases likely reflect the higher degree of neuroplasticity, which has been documented early after stroke.⁸⁰ The effect sizes obtained in the RCTs using a population at >6 months poststroke ranged from 0.05 to 4.28, demonstrating that individuals are still capable of change at later time periods poststroke when receiving treatment. Of importance, our collective data demonstrate that rehabilitation treatments can affect patient perceptions of functional change as effectively as laboratory-based functional measures. In fact, CIMT is an exemplary treatment model where patient perceptions of change (MAL) were 1.6 to 6.2 times larger than effect sizes obtained with laboratory-based functional performance measures (ARAT or Wolf). A likely explanation is that CIMT was specifically designed to overcome learned nonuse, thereby targeting functional change in one’s own environment. Additionally, the higher relative effect sizes observed in the MAL following CIMT could also reflect patient bias. This bias could be present because of the large investment of time and effort required from the CIMT program. Finally, it should be noted that our observations concerning the relative ability of treatments to affect laboratory-based versus patient-perceived function may be limited to studies able to capture statistically significant effects.

Real-time accelerometry monitoring of UE activity in the home and community is a promising technology to objectively capture function in one’s own environment without the necessity of self-report.⁸¹ However, only 1 study provided an estimate of true or important change for accelerometry measures. More studies are needed to better understand the usefulness of real-time activity monitors for capturing the effectiveness of UE rehabilitation interventions.

The influence of the method used to calculate effect sizes (ie, methods based on baseline standard deviation versus standard deviation of the change scores) was also revealed. Methods based on the change score standard deviation (ie, standardized response mean) produced estimates that were up to 1.2 to 1.9 times larger than those calculated based on the baseline standard deviation (ie, population effect size). The method of effect size calculation is an important contextual factor that must be considered when designing and interpreting research. This has particular importance for interpreting treatment effects and when performing sample size calculations. The observed influence of the calculation method also highlights the difficulty of using an effect size in isolation when making a judgment about a measure’s general responsiveness.⁸²

Important Change

A key finding was that important change values obtained through anchor-based methods (eg, based on patient opinion or comparative measures) were higher than those for the distribution methods (eg, statistical estimates) among the studies that used both approaches.^65-67 In fact, the MCID of 1 measure (SIS) tripled in magnitude from the distribution-based to the anchor-based approach using the same participants.⁶⁶ Our results indicate that distribution methods result in smaller MCID, and researchers may be tempted to interpret their findings in light of this MCID, especially if the intervention has small effects. However, statistically driven distribution methods have been criticized for their lack of meaning to participants.^1,83 The MCID values from anchor-based approaches are larger and may provide a challenge in finding therapies that can achieve this effect. Some have questioned whether it is realistic to achieve MCID values derived from patient-perceived global rating scales (anchor-based) because many factors (eg, recall bias, baseline characteristics, expectations of treatment, and question format) can affect patients’ perception of change.^64,84 It has been suggested that a combination of anchor-based methods from patient and clinical perspectives be used to determine a MCID value, and distribution-based methods should only supplement this information.⁸³

Minimal Detectable Change

This synthesis highlighted important differences between laboratory-based and participant-perceived functional measures. Measures that capture perceived function in one’s own environment (eg, SIS, MAL) required larger values to surpass their measurement error than laboratory-based performance measures (eg, Wolf, ARAT). Although incorporating patient perspectives of functional change is an important component of capturing meaningful outcomes in neurorehabilitation research,⁸⁵ researchers should be aware of the larger sample size required to be able to capture true change using perceived function measures. Quality criteria guidelines recommend that the values needed to surpass measurement error calculated at a 95% confidence interval (ie, true change captured by MDC₉₅) should be less than the minimum values considered to be important (ie, MCID).⁸⁶ In contrast, our study found that the MDC₉₅ and MDC₉₀ for patient-perceived functional measures (eg, MAL, SIS) were similar or greater than the MCID. There remains debate in the literature, however, concerning the best estimate of MDC values for self-report measures. Some researchers argue that MDC₉₀ and MDC₉₅ produce overly conservative estimates and recommend setting the MDC to 1 SEM.^87,88

One of the limitations of this article is that there is no one standard approach for conceptualizing responsiveness. We used the broad taxonomy described by Beaton et al.¹ One international group (COSMIN) defined a narrower conceptualization of responsiveness, which used a longitudinal validity approach such as correlating change with an external criterion or determining the area under the receiver operating characteristic curve to distinguish between known groups.⁸² In addition, whereas there are established appraisal guidelines for intervention studies (eg, PeDRO Score), consensus has yet to be reached with regard to assessing the rigor of studies that measure an outcome measure’s psychometric properties. Application of the Beaton taxonomy, which focuses on the nature of the change, provided a useful framework for understanding the state of the stroke literature in regard to the broader topic of measuring functional change. Effect sizes provided in this study can inform hypotheses for future responsiveness testing. Finally, ceiling and floor effects were also not considered in this review. Large ceiling and/or floor effects may indicate that there is a subsection of the population for which measures are not as responsive and is another factor that can bias estimates of important change values.⁸⁹ Given the contextual nature of responsiveness, the dominance of CIMT among the included studies should be noted. Samples from CIMT trials made up 80% of the important change articles, 56% of the detectable change articles, and 50% of the RCTs used to examine observed change in response to an effective treatment. Thus, the literature informing the ability of measurement tools to capture functional change in the UE poststroke may be biased in favor of those individuals who are eligible for CIMT trials.

Conclusions

In summary, this synthesis revealed important findings that have implications for the measurement and interpretation of UE functional recovery following stroke. The magnitude of important change or change that surpasses measurement error varied substantially depending on the method of calculation used. Our findings suggest that rehabilitation treatments can affect patient perceptions of functional change as effectively as laboratory-based functional measures; however, research studies may require higher sample sizes to account for the larger measurement error associated with patient-perceived functional measures. Future studies examining meaningful change in UE function in varied subgroups of individuals (ie, at different levels of stroke severity and stage of recovery) are needed.

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is supported by funding from the Canadian Health Institute of Research (CIHR MSH-63617) and Michael Smith Foundation of Health Research (JJE).

References

Beaton

Bombardier

Katz

Wright

. A taxonomy for responsiveness. J Clin Epidemiol. 2001;54:1204-1217.

Mokkink

Terwee

Patrick

. The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health-related patient-reported outcomes. J Clin Epidemiol. 2010;63:737-745.

Kain

MacLaren

. P less than .05: what does it really mean? Pediatrics. 2007;119:608.

Hopewell

Clarke

Lefebvre

Scherer

. Handsearching versus electronic searching to identify reports of randomized trials. Cochrane Database Syst Rev. 2007;(2):MR000001.

Morris

DeShon

. Combining effect size estimates in meta-analysis with repeated measures and independent-groups designs. Psychol Methods. 2002;7:105-125.

Stratford

. Getting more from the literature: estimating the standard error of measurement from reliability studies. Physiother Can. 2004;56:27-30.

Flansbjer

Holmbäck

Downham

Patten

Lexell

. Reliability of gait performance tests in men and women with hemiparesis after stroke. J Rehabil Med. 2005;37:75-82.

Angst

. The new COSMIN guidelines confront traditional concepts of responsiveness. BMC Med Res Methodol. 2011;11:152.

Hozo

Djulbegovic

Hozo

. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13.

10.

Beebe

Lang

. Relationships and responsiveness of six upper extremity function tests during the first six months of recovery after stroke. J Neurol Phys Ther. 2009;33:96-103.

11.

Hsueh

Hsieh

. Responsiveness of two upper extremity function instruments for stroke inpatients receiving rehabilitation. Clin Rehabil. 2002;16:617-624.

12.

Lin

Hsu

Sheu

. Psychometric comparisons of 4 measures for assessing upper-extremity function in people with stroke. Phys Ther. 2009;89:840-850.

13.

Rabadi

. Comparison of the Action Research Arm Test and the Fugl-Meyer assessment as measures of upper-extremity motor weakness after stroke. Arch Phys Med Rehabil. 2006;87:962-966.

14.

Barreca

Stratford

Masters

Lambert

Griffiths

McBay

. Validation of three shortened versions of the Chedoke Arm and Hand Activity Inventory. Physiother Can. 2006;58:148-156.

15.

Barreca

Stratford

Lambert

Masters

Streiner

. Test-retest reliability, validity, and sensitivity of the Chedoke Arm and Hand Activity Inventory: a new measure of upper-limb function for survivors of stroke. Arch Phys Med Rehabil. 2005;86:1616-1622.

16.

Barreca

Stratford

Masters

Lambert

Griffiths

. Comparing 2 versions of the Chedoke Arm and Hand Activity Inventory with the Action Research Arm Test. Phys Ther. 2006;86:245-253.

17.

Pandyan

Cameron

Powell

Stott

Granat

. 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.

18.

Brunner

Skouen

Strand

. Recovery of upper extremity motor function post stroke with regard to eligibility for constraint-induced movement therapy. Top Stroke Rehabil. 2011;18:248-257.

19.

Au-Yeung

Hui-Chan

. Predicting recovery of dextrous hand function in acute stroke. Disabil Rehabil. 2009;31:394-401.

20.

Rand

Eng

. Disparity between functional recovery and daily use of the upper and lower extremities during subacute stroke rehabilitation. Neurorehabil Neural Repair. 2012;26:76-84.

21.

Roiha

Kirveskari

Kaste

. Reorganization of the primary somatosensory cortex during stroke recovery. Clin Neurophysiol. 2011;122:339-345.

22.

Blennerhassett

Avery

Carey

. The test-retest reliability and responsiveness to change for the Hand Function Survey during stroke rehabilitation. Aust Occup Ther J. 2010;57:431-438.

23.

Rehme

Fink

von Cramon

Grefkes

. The role of the contralesional motor cortex for motor recovery in the early days after stroke assessed with longitudinal FMRI. Cereb Cortex. 2011;21:756-768.

24.

Filiatrault

Arsenault

Dutil

Bourbonnais

. Motor function and activities of daily living assessments: a study of three tests for persons with hemiplegia. Am J Occup Ther. 1991;45:806-810.

25.

Duncan

Wallace

Lai

Johnson

Embretson

Laster

. The Stroke Impact Scale version 2.0. Evaluation of reliability, validity, and sensitivity to change. Stroke. 1999;30:2131-2140.

26.

Kamel

Ghani

Zaiton

El-Motayam

El-Fattah

. Health related quality of life in stroke survivors measured by the Stroke Impact Scale. Egypt J Neurol Psychiatry Neurosurg. 2010;47:267-274.

27.

Sezer

Yavuzer

Sivrioglu

Basaran

Koseoglu

. Clinimetric properties of the Duruoz Hand Index in patients with stroke. Arch Phys Med Rehabil. 2007;88:309-314.

28.

Bouffioulx

Arnould

Thonnard

. Satisfaction with activity and participation and its relationships with body functions, activities, or environmental factors in stroke patients. Arch Phys Med Rehabil. 2011;92s:1404-1410.

29.

Rand

Gottlieb

Weiss

. Recovery of patients with a combined motor and proprioception deficit during the first six weeks of post stroke rehabilitation. Phys Occup Ther Geriatr. 2001;18:69-87.

30.

Wittenberg

Bastings

Fowlkes

Morgan

Good

Pons

. Dynamic course of intracortical TMS paired-pulse responses during recovery of motor function after stroke. Neurorehabil Neural Repair. 2007;21:568-573.

31.

Higgins

Mayo

Desrosiers

Salbach

Ahmed

. Upper-limb function and recovery in the acute phase poststroke. J Rehabil Res Dev. 2005;42:65-76.

32.

Mayo

Wood-Dauphinee

Ahmed

. Disablement following stroke. Disabil Rehabil. 1999;21:258-268.

33.

Desrosiers

Malouin

Richards

Bourbonnais

Rochette

Bravo

. Comparison of changes in upper and lower extremity impairments and disabilities after stroke. Int J Rehabil Res. 2003;26:109-116.

34.

Feydy

Carlier

Roby-Brami

. Longitudinal study of motor recovery after stroke: Recruitment and focusing of brain activation. Stroke. 2002;33:1610-1617.

35.

Hsu

Wang

Yip

Chiu

Hsieh

. Dose-response relation between neuromuscular electrical stimulation and upper-extremity function in patients with stroke. Stroke. 2010;41:821-824.

36.

Page

Levine

Leonard

. Modified constraint-induced therapy in acute stroke: a randomized controlled pilot study. Neurorehabil Neural Repair. 2005;19:27-32.

37.

Sun

Hsu

Sun

Hwang

Yang

Wang

. Combined botulinum toxin type A with modified constraint-induced movement therapy for chronic stroke patients with upper extremity spasticity: a randomized controlled study. Neurorehabil Neural Repair. 2010;24:34-41.

38.

Myint

Yuen

. A study of constraint-induced movement therapy in subacute stroke patients in Hong Kong. Clin Rehabil. 2008;22:112-124.

39.

van der Lee

Wagenaar

Lankhorst

Vogelaar

Deville

Bouter

. Forced use of the upper extremity in chronic stroke patients: results from a single-blind randomized clinical trial. Stroke. 1999;30:2369-2375.

40.

Harris

Eng

Miller

Dawson

. A self-administered graded repetitive arm supplementary program (GRASP) improves arm function during inpatient stroke rehabilitation: a multi-site randomized controlled trial. Stroke. 2009;40:2123-2128.

41.

Shindo

Fujiwara

Hara

. Effectiveness of hybrid assistive neuromuscular dynamic stimulation therapy in patients with subacute stroke. Neurorehabil Neural Repair. 2011;25:830-837.

42.

Dahl

Askim

Stock

Langorgen

Lydersen

Indredavik

. Short- and long-term outcome of constraint-induced movement therapy after stroke: a randomized controlled feasibility trial. Clin Rehabil. 2008;22:436-447.

43.

Cacchio

De Blasis

Santilli

Spacca

. Mirror therapy in complex regional pain syndrome type 1 of the upper limb in stroke patients. Neurorehabil Neural Repair. 2009;23:792-799.

44.

Kowalczewski

Gritsenko

Ashworth

Ellaway

Prochazka

. Upper-extremity functional electric stimulation-assisted exercises on a workstation in the subacute phase of stroke recovery. Arch Phys Med Rehabil. 2007;88:833-839.

45.

Pang

Harris

Eng

. 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.

46.

Taub

Uswatte

King

Morris

Crago

Chatterjee

. A placebo-controlled trial of constraint-induced movement therapy for upper extremity after stroke. Stroke. 2006;37:1045-1049.

47.

Wittenberg

Chen

Ishii

. Constraint-induced therapy in stroke: magnetic-stimulation motor maps and cerebral activation. Neurorehabil Neural Repair. 2003;17:48-57.

48.

Chuang

Lin

Chen

Tsay

. Randomized trial of distributed constraint-induced therapy versus bilateral arm training for the rehabilitation of upper-limb motor control and function after stroke. Neurorehabil Neural Repair. 2011;25:130-139.

49.

Gauthier

Taub

Perkins

Ortmann

Mark

Uswatte

. Remodeling the brain: plastic structural brain changes produced by different motor therapies after stroke. Stroke. 2008;39:1520-1525.

50.

Khan

Oesch

Gamper

Kool

Beer

. Potential effectiveness of three different treatment approaches to improve minimal to moderate arm and hand function after stroke: a pilot randomized clinical trial. Clin Rehabil. 2011;25:1032-1041.

51.

Tariah

Almalty

Sbeih

Al-Oraibi

Bernhardt

Rowe

. Constraint induced movement therapy for stroke survivors in Jordon: a home-based model. Int J Ther Rehabil. 2010;17:638-646.

52.

Dromerick

Lang

Birkenmeier

. Very early constraint-induced movement during stroke rehabilitation (VECTORS): a single-center RCT. Neurology. 2009;73:195-201.

53.

Church

Price

Pandyan

Huntley

Curless

Rodgers

. Randomized controlled trial to evaluate the effect of surface neuromuscular electrical stimulation to the shoulder after acute stroke. Stroke. 2006;37:2995-3001.

54.

Chen

Tsai

Lin

Chou

. A randomized controlled trial of modified constraint-induced movement therapy for elderly stroke survivors: changes in motor impairment, daily functioning, and quality of life. Arch Phys Med Rehabil. 2007;88:273-278.

55.

Kimberley

Lewis

Auerbach

Dorsey

Lojovich

Carey

. Electrical stimulation driving functional improvements and cortical changes in subjects with stroke. Exp Brain Res. 2004;154:450-460.

56.

Lin

Chang

Chen

. Effects of constraint-induced therapy versus bilateral arm training on motor performance, daily functions, and quality of life in stroke survivors. Neurorehabil Neural Repair. 2009;23:441-448.

57.

Lin

Liu

Chen

Hsu

. Constraint-induced therapy versus dose-matched control intervention to improve motor ability, basic/extended daily functions, and quality of life in stroke. Neurorehabil Neural Repair. 2009;23:160-165.

58.

Ertelt

Small

Solodkin

. Action observation has a positive impact on rehabilitation of motor deficits after stroke. Neuroimage. 2007;36:T164-T173.

59.

Conroy

Whitall

Dipietro

. Effect of gravity on robot-assisted motor training after chronic stroke: a randomized trial. Arch Phys Med Rehabil. 2011;92:1754-1761.

60.

Liao

Hsieh

Lin

Chang

. Effects of robot-assisted upper limb rehabilitation on daily function and real-world arm activity in patients with chronic stroke: a randomized controlled trial. Clin Rehabil. 2012;26:111-120.

61.

Page

Levin

Hermann

Dunning

Levine

. Longer versus shorter daily durations of electrical stimulation during task-specific practice in moderately impaired stroke. Arch Phys Med Rehabil. 2012;93:200-206.

62.

Hsieh

Liao

Lin

Lee

. Effects of treatment intensity in upper limb robot-assisted therapy for chronic stroke: a pilot randomized controlled trial. Neurorehabil Neural Repair. 2011;25:503-511.

63.

Fritz

George

Wolf

Light

. 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.

64.

Lang

Edwards

Birkenmeier

Dromerick

. Estimating minimal clinically important differences of upper-extremity measures early after stroke. Arch Phys Med Rehabil. 2008;89:1693-1700.

65.

Wang

Lin

Chung

Pei

Teng

. Validity, responsiveness, and clinically important difference of the ABILHAND questionnaire in patients with stroke. Arch Phys Med Rehabil. 2011;92:1086-1091.

66.

Lin

. Minimal detectable change and clinically important difference of the Stroke Impact Scale in stroke patients. Neurorehabil Neural Repair. 2010;24:486-492.

67.

Lin

Hsieh

Chen

Jang

Liu

. Minimal detectable change and clinically important difference of the Wolf Motor Function Test in stroke patients. Neurorehabil Neural Repair. 2009;23:429-434.

68.

Fritz

Blanton

Uswatte

Taub

Wolf

. Minimal detectable change scores for the Wolf Motor Function Test. Neurorehabil Neural Repair. 2009;23:662-667.

69.

Wolf

Thompson

Morris

. The EXCITE trial: attributes of the Wolf Motor Function Test in patients with subacute stroke. Neurorehabil Neural Repair. 2005;19:194-205.

70.

Uswatte

Giuliani

Winstein

Zeringue

Hobbs

Wolf

. 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.

71.

Uswatte

Taub

Morris

Vignolo

McCulloch

. Reliability and validity of the upper-extremity Motor Activity Log-14 for measuring real-world arm use. Stroke. 2005;36:2493-2496.

72.

Uswatte

Taub

Morris

Light

Thompson

. The Motor Activity Log-28: assessing daily use of the hemiparetic arm after stroke. Neurology. 2006;67:1189-1194.

73.

Kopp

Kunkel

Flor

. The Arm Motor Ability Test: reliability, validity, and sensitivity to change of an instrument for assessing disabilities in activities of daily living. Arch Phys Med Rehabil. 1997;78:615-620.

74.

Richards

Stoker-Yates

Pohl

Wallace

Duncan

. Reliability and validity of two tests of upper extremity motor function post-stroke. Occup Ther J Res. 2001;21:201-219.

75.

Carod-Artal

Coral

Trizotto

Moreira

. The Stroke Impact Scale 3.0: evaluation of acceptability, reliability, and validity of the Brazilian version. Stroke. 2008;39:2477-2484.

76.

van der Lee

Beckerman

Knol

De Vet

Bouter

. Clinimetric properties of the Motor Activity Log for the assessment of arm use in hemiparetic patients. Stroke. 2004;35:1410.

77.

van der Lee

Beckerman

Lankhorst

Bouter

. The responsiveness of the Action Research Arm Test and the Fugl-Meyer assessment scale in chronic stroke patients. J Rehabil Med. 2001;33:110-113.

78.

Streiner

Norman

. Health Measurement Scales: A Practical Guide to Their Development and Use. 4th ed. New York, NY: Oxford University Press; 2008.

79.

Lydick

Epstein

. Interpretation of quality of life changes. Qual Life Res. 1993;2:221-226.

80.

Biernaskie

Chernenko

Corbett

. Efficacy of rehabilitative experience declines with time after focal ischemic brain injury. J Neurosci. 2004;24:1245-1254.

81.

Dobkin

Dorsch

. The promise of mHealth: daily activity monitoring and outcome assessments by wearable sensors. Neurorehabil Neural Repair. 2011;25:788-798.

82.

Mokkink

Terwee

Knol

. The COSMIN checklist for evaluating the methodological quality of studies on measurement properties: a clarification of its content. BMC Med Res Methodol. 2010;10:22.

83.

Revicki

Hays

Cella

Sloan

. Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes. J Clin Epidemiol. 2008;61:102-109.

84.

Guyatt

Norman

Juniper

Griffith

. A critical look at transition ratings. J Clin Epidemiol. 2002;55:900-908.

85.

Salter

Jutai

Teasell

Foley

Bitensky

Bayley

. Issues for selection of outcome measures in stroke rehabilitation: ICF activity. Disabil Rehabil. 2005;27:315-340.

86.

Terwee

Bot

SDM

De Boer

. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol. 2007;60:34-42.

87.

Wyrwich

Tierney

Wolinksy

. Further evidence supporting standard error of measurement based criterion for identifying meaningful intra-individual change in health-related quality of life. J Clin Epidemiol. 1999;52:861-873.

88.

Turner

Schunemann

Griffith

. The minimal detectable change cannot reliably replace the minimal important difference. J Clin Epidemiol. 2010;63:28-36.

89.

Stucki

Daltroy

Katz

Johannesson

Liang

. Interpretation of change scores in ordinal clinical scales and health status measures: the whole may not equal the sum of the parts. J Clin Epidemiol. 1996;49:711-717.

Functional Recovery Following Stroke

Abstract

Keywords

Introduction

Methods

Literature Search

Inclusion Criteria

Data Extraction and Organization

Results

Observed Change

Observed change over natural recovery

Observed change in response to a treatment of known efficacy

Important Change

Detectable Change

Discussion

Observed Change

Important Change

Minimal Detectable Change

Conclusions

Footnotes

Declaration of Conflicting Interests

Funding

References