| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Reports |
V.S. Mercer, PT, PhD, is Associate Professor, Division of Physical Therapy, Department of Allied Health Sciences, University of North Carolina at Chapel Hill, CB 7135, Bondurant Hall, Ste 3022, Chapel Hill, NC 27599-7135 (USA).
J.K. Freburger, PT, PhD, is Research Associate and Fellow, Cecil G. Sheps Center for Health Services Research, and Research Scientist, Institute on Aging, University of North Carolina at Chapel Hill.
S-H. Chang, PT, PhD, is Postdoctoral Research Scholar, Graduate Program in Physical Therapy and Rehabilitation Science, The University of Iowa, Iowa City, Iowa.
J.L. Purser, PT, PhD, is Assistant Professor, Division of Geriatrics, Department of Medicine, and Division of Physical Therapy, Department of Community and Family Medicine, and Senior Fellow, Center for the Study of Aging and Human Development, Duke University Medical Center, Durham, North Carolina.
Address all correspondence to Dr Mercer at: vmercer{at}med.unc.edu
Submitted July 29, 2008;
Accepted March 18, 2009
 | Abstract |
|---|
Objective: The purpose of this study was to assess convergent validity of the Step Test (ST) and the knee extension component of the Upright Motor Control Test (UMCe) as measures of paretic-limb loading and of the Repetitive Reach Test (RR) as a measure of weight transfer in the first 6 months after stroke.
Design: This was a prospective cohort study of 33 adults with lower-extremity motor impairment following unilateral, noncerebellar stroke. Participants were tested one time per month from 1 to 6 months poststroke.
Results: Scores on the ST (performed with the nonparetic leg as the stepping leg) and UMCe were positively correlated with peak vertical ground reaction forces (GRFs) beneath the paretic limb during functional tasks (R2=.35–.76 for the ST, pseudo R2=.21–.34 for the UMCe). Scores on the RR were positively correlated with change in vertical GRF beneath the paretic limb during the diagonal reach task (R2=.45) and with weight-transfer time during stepping with the nonparetic limb (R2=.15).
Conclusions: The ST, performed with the nonparetic leg as the stepping leg, is a valid measure of paretic-limb loading during stroke recovery. Of the clinical measures tested, the ST correlated most strongly with the force platform measures.
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussionConclusionReferences |
|---|
Several researchers have provided evidence that impaired lower-extremity loading9–13 and, to a lesser extent, weight-transfer8,14,15 abilities after stroke are associated with functional deficits. In most of these studies, the ability to load the paretic leg or to load both legs symmetrically was measured using force platforms or digital scales to record forces under the feet. Loading on the paretic lower extremity has been shown to relate to performance of functional tasks such as reaching in sitting,9,16 rising from a chair,10,11 standing,17,18 walking,12,19 and climbing curbs and stairs.11,13 Greater weight-bearing symmetry during the sit-to-stand (STS) task is associated with faster STS performance under both self-paced and fast-paced conditions.10 Lee et al20 reported that the maximum weight-bearing difference between the 2 lower extremities during the STS task was highest for the subjects with stroke who had the lowest scores on the Functional Independence Measure.21 Kim and Eng19 found that the greater the asymmetry of various temporal-distance and force platform measures of gait, the lower the gait speed in subjects with chronic stroke. The relationship was strongest for asymmetry of vertical ground reaction forces (GRFs), indicating that reduced dynamic loading of the paretic leg significantly affects gait performance.
Research also suggests that interventions designed to improve paretic–lower-extremity loading improve functional performance in individuals with stroke. Dean and Shepherd9 demonstrated that practice of seated reaching tasks over a 2-week period increased paretic–lower-extremity loading and improved task performance in subjects with stroke. After training, subjects in the experimental group showed increased paretic–lower-extremity loading when reaching forward and toward the paretic side and were able to reach faster and farther than subjects in the control group. Experimental group subjects, but not control group subjects, also exhibited a significant increase after intervention in paretic–lower-extremity loading during the STS task. In a study by Cheng et al,22 patients with stroke who received symmetrical standing training and repetitive STS training in addition to usual care showed more-symmetrical body-weight distribution during STS training and fewer falls over the 6-month follow-up period than patients who received only usual care.
A few researchers have focused specifically on the ability to transfer weight from one leg to the other after stroke.4,8,14,15,23 The functional importance of this ability, however, is not as well-established as that of paretic–lower-extremity loading. Both spatial and temporal features of weight transfer have been examined using force platform and kinematic data. Weight-transfer ability, as indicated by movement of the center of pressure or center of mass during voluntary weight shifts in a standing position, has been reported to relate to various measures of standing balance14 and gait performance.4 Pai and colleagues,15 however, found only a weak correlation (rho=.40) between successful weight transfer during a single-leg flexion task in a standing position and gait speed.
Paretic–lower-extremity loading and weight-transfer abilities are a major focus of rehabilitation training for patients with hemiparesis.24–27 Unfortunately, most previous studies of these abilities have relied heavily on laboratory measurements, such as GRFs and medial-lateral impulses calculated from force platform data. Force platforms are expensive and require extra time and a high level of technical support for data collection and reduction.28 Valid and practical clinical measures of paretic–lower-extremity loading and weight transfer have not been identified. In many settings, the only practical way to measure weight bearing is with digital scales.7,13 Because these scales cannot record dynamic load changes, they are useful only for measuring weight-bearing performance during certain tasks, such as static standing and prolonged voluntary weight shifting. Even with selection of relatively static tasks, important components of performance may be missed with use of such systems. The lack of clinically accessible measures makes it difficult for clinicians to make sound decisions about interventions directed toward improving symmetry and weight transfer and, ultimately, increasing function.
The Step Test (ST)29 and the knee extension component of the Upright Motor Control Test (UMCe)30,31 are clinical tests that require single-limb stance on the paretic lower extremity and, therefore, have face validity as measures of paretic-limb loading. The Repetitive Reach Test (RR)28 is a clinical test that requires stability in bipedal stance during rapid, repetitive nonparetic–upper-extremity reaching movements across midline and forward beyond arm's length. Because the individual must control movement of the body's center of mass toward both the nonparetic and paretic sides during these dynamic reaches, we viewed the RR as a potential measure of weight-transfer abilities.
The aims of this study were: (1) to determine whether the ST and the UMCe are valid measures of paretic–lower-extremity loading and (2) to determine whether the RR is a valid measure of weight transfer during the first 6 months of stroke recovery. The type of validity assessed in this study was convergent validity, which is a subtype of construct validity. Convergent validity reflects the ability of an instrument to measure an abstract concept, or construct.32 The basis for convergent validity is that 2 measures thought to reflect the same underlying construct (eg, a clinical measure and a laboratory measure of paretic–lower-extremity loading) should correlate highly.
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
Inclusion criteria were: (1) a primary diagnosis of unilateral noncerebellar stroke; (2) medically stable and free of major cardiovascular conditions (eg, recent myocardial infarction, unstable angina, ventricular tachycardia) and musculoskeletal problems (eg, fracture, sprain, strain); (3) able to follow 3-step commands; (4) able to reach in all directions to touch a target with the nonparetic hand while sitting without support; (5) lower-extremity motor impairment, as indicated by a score of 
28 on the lower-extremity motor scale of the Fugl-Meyer Assessment33; (6) adequate vision and hearing for completing the study protocol, as indicated by the ability to see targets for reaches and to follow oral instructions during screening; and (7) residence within an 80-km (50-mile) radius, with willingness to return to our laboratory for testing at monthly intervals from 1 to 6 months poststroke.
Potential participants were excluded from the study if they: (1) had a history of previous strokes or other neurologic diseases or disorders, such as Parkinson disease or Alzheimer disease; (2) were unable to live or ambulate independently in the community prior to the stroke; (3) had a terminal illness; or (4) had pain, limited motion, or weakness in the nonparetic lower extremity that affected performance of daily activities (by self-report). Informed consent was obtained from all participants prior to testing.
Procedure
Participants were recruited and baseline testing was completed during the time period from hospital admission to 1 month poststroke. At baseline, we examined paretic-limb motor function using the lower-extremity motor scale of the Fugl-Meyer Assessment33 and collected quantitative data on visuospatial neglect. Tests for neglect, including the letter and star cancellation subtests of the Behavioral Inattention Test (BIT),34,35 were included in the study primarily because of the strong association between spatial neglect and postural disorders such as asymmetrical weight bearing in standing.36
Testing at the Center for Human Movement Science at the University of North Carolina at Chapel Hill began at 1 month poststroke and continued at monthly intervals through 6 months poststroke. Participants were weighed at the beginning of each test session. Measurements of height and right foot length were recorded at the first (1-month) test session only and were used to determine standardized positions for force platform testing. The clinical and force platform tests described below were administered by the same examiner at each session.
Clinical tests.
The ST, UMCe, and RR were selected because of their face validity and because they: (1) simulate functional movements that challenge dynamic stability in standing, (2) were developed for and tested with people with stroke, (3) have evidence of reliability, and (4) can be easily administered in a variety of clinical settings.
The ST assesses an individual's ability to place one foot onto a 7.5-cm-high step and then back down to the floor repeatedly as fast as possible for 15 seconds.29 The step is placed 5 cm in front of the individual's feet. The test is scored by recording the number of steps completed in the 15-second period for each leg. Participants wore any customary orthoses but were not permitted to use an assistive device during testing. They performed the test first with the nonparetic foot and then with the paretic foot placed on and off the step. Because we were interested in loading of the paretic leg, only ST scores for stepping with the nonparetic leg were analyzed. The ST has high test-retest reliability in people poststroke (intraclass correlation coefficient [3,1]=.94 for performance with the nonparetic leg as the stepping leg)29 and is responsive to change during stroke rehabilitation, with standardized response means (SRMs)37,38 of 0.92 and 0.95 for the nonparetic and paretic legs, respectively.37,39 In a sample of older adults who were healthy and individuals undergoing inpatient rehabilitation after stroke, ST scores for stepping with the nonparetic leg were significantly correlated with scores on the Functional Reach Test,40 gait speed, and stride length, with Pearson correlation coefficients (r) of .73, .83, and .83, respectively.29
The UMCe assesses strength (force-generating capacity) and control of the knee extensor muscles during single-limb stance on the paretic side.31 In standing, the participants flexed both knees to approximately 30 degrees and then lifted the nonparetic foot off the ground. They then attempted to extend the knee on the paretic side while still holding the nonparetic foot off the ground. Knee extension was graded on a 3-point scale: 1=unable to bear full weight on a flexed knee, 2=able to support full weight on the flexed knee but unable to extend further, and 3=able to complete full range of knee extension. This test has evidence for intertester reliability and predictive validity.31 Knee extension and flexion scores on the UMCe have been shown to predict home versus community walking ability prior to hospital discharge, with a score of 3 on either component predicting community ambulation.31
Repetitive reach assesses how rapidly an individual is able to reach back and forth between 2 targets with the nonparetic arm while maintaining bipedal stance.28 The number of reaches performed during a 30-second period is recorded.39 Participants reached with the nonparetic upper extremity from a target placed on the nonparetic side opposite the greater trochanter to a target placed 15 cm beyond arm's length in front of their paretic hip (at the same height as the side target). Participants maintained their feet in a step stance position with the paretic foot forward while reaching. Intrasession and intersession reliability coefficients (Pearson product moment correlation coefficients) for RR ranged from .90 to .99 in patients tested at a mean time interval of 2 months after stroke.28 Intersession reliability was slightly higher for testing in step stance (r=.94) compared with parallel stance (r=.90) positions. Like the ST, the RR has been shown to be responsive to changes occurring during stroke rehabilitation (SRMs of 0.75 and 0.86 when performed in parallel stance and step stance, respectively).39
Force platform tests.
Two Bertec (N60501, Type 4060A, 40- x 60-cm) force platforms* mounted side-by-side in the floor were used to measure GRFs during performance of 4 functional tasks: diagonal reach in standing, STS transfer, stepping up onto a step with the nonparetic leg leading, and stepping up onto a step with the paretic leg leading. Foot tracings were used to facilitate consistency of foot position during testing for each participant. The participant's feet were positioned so that one foot was on each force platform and the distance between the midpoints of the heels was equal to right foot length.23 This positioning served to standardize the length and width of the base of support with respect to each participant's foot length. Participants performed 2 practice trials and 4 test trials of each task, as described below. Peak Motus software
on a personal computer was used to acquire force platform data online at a sampling rate of 500 Hz.
For diagonal reach in standing, participants used their nonparetic upper extremity to pick up an unopened 355-mL (12-fl oz) soft drink can from a starting point on the nonparetic side (15 cm directly lateral to the greater trochanter) and set it down on a target located 1.4 arm-lengths away on the paretic side at a 45-degree angle from midline.9 They were instructed to move as fast as they could without feeling unsafe or dropping or knocking over the can. Pressure switches at the start location and at the target were used to detect the beginning and end of the reach, respectively. The switch at the start location was "on" at the beginning of the trial and was released when a participant lifted the can. The switch at the end location was overlaid by a 12- x 12-cm sheet of hard plastic that served as the target. This switch closed when the can contacted the plastic sheet, signaling the end of the reaching movement.
For the STS transfer, participants started in a sitting position in a standard wooden chair without armrests (seat height=44.4 cm) with feet positioned in the tracings and the chair just behind the force platforms. Participants were instructed to come to a standing position as fast as they could without feeling unsafe. They were not allowed to use an assistive device and were asked to try not to use either upper extremity to push up from the chair. Because the legs of the chair were not in contact with the force platforms, a participant's use of upper-extremity support on the chair resulted in a proportional decrease in the forces recorded beneath the feet.
For the stepping tasks, participants started in a standing position with their feet in the tracings and stepped up onto an 8-cm-high step with both feet. The step, which did not contact the force platform, was positioned with the front edge 1.5 foot-lengths from the back of each participant's heels. Participants were instructed to step as fast as they could without feeling unsafe. The stepping task was performed first with the nonparetic lower extremity leading (4 test trials) and then with the paretic lower extremity leading (4 test trials).
Data Reduction
Force platform data were exported from Motus to customized software programs (MotionSoft 3D version 6.5 and MotionSoft Discrete Data Reader version 6.0)
for processing and reduction. The GRF signals were calibrated and converted to newtons. The force platform measures used in the analyses are described below. For each force platform measure, the mean of the 4 test trials was used. All GRF measurements were normalized by the participant's body weight recorded at each test session.
The peak vertical GRF through the paretic lower extremity was determined for each of the following tasks: diagonal reach, STS transfer, and stepping up onto a step with the nonparetic limb leading. These were the 3 biomechanical measures of paretic-limb loading.
The biomechanical measures of weight transfer were: (1) change in vertical GRF, (2) weight-transfer time, and (3) medial-lateral impulse. Change in vertical GRF was determined for the diagonal reach task only and was measured as the difference between the maximum and minimum values of the vertical GRF beneath the paretic limb during the reach (Fig. 1). The time period of the reach was indicated by the pressure switch signals. Weight-transfer time41 was determined for stepping up with the nonparetic limb leading (Fig. 2) and for stepping up with the paretic limb leading. This measure of weight transfer was defined as the duration from the first change in vertical GRF data (change for either limb of more than 2% of body weight from the mean baseline measurement during quiet standing) to when the force beneath the leading limb reached zero, indicating that the limb had lost contact with the force platform. Medial-lateral impulse42 also was determined for the 2 stepping tasks (stepping up with the nonparetic limb leading and stepping up with the paretic limb leading) and was calculated as the integral of the medial-lateral GRF beneath the paretic limb during the weight-transfer time.
|
|
Data Analysis
All analyses were conducted using Stata version 9.2.
Descriptive statistics were generated for the baseline characteristics of the sample and for the clinical and biomechanical measurements of paretic-limb loading and weight transfer collected at each session. The clinical and biomechanical data for each session and for the sessions combined also were plotted to examine visually the relationships between the measures. Bivariate linear regression analyses were conducted to determine the relationship between the ST scores (nonparetic limb) and each of the paretic–limb-loading measures (ie, peak vertical GRF through the paretic lower extremity for the diagonal reach and STS tasks and stepping with the nonparetic lower extremity leading) and between the RR scores and each of the weight-transfer measures (ie, change in vertical GRF, weight-transfer times during the stepping tasks, and medial-lateral impulse produced by the paretic limb during the stepping tasks). Because lower scores on weight-transfer times indicated less impairment, weight-transfer times of 0 (indicating individuals who could not attempt the test without assistance) were eliminated from the regression analyses. A bivariate linear regression analysis describes the nature (eg, as ST scores increase, limb-loading measurements increase) and strength (eg, the degree of correlation between ST scores and limb-loading measurements) of a linear relationship between a continuous dependent variable and a continuous or discrete independent variable. Linear regression analysis is one analytic approach for establishing convergent validity. The R2 value generated from a linear regression analysis specifically describes the strength of the relationship between the dependent and independent variable. For example, an R2 value of .30 indicates that 30% of the variation in the dependent variable is explained by variation in the independent variable.
Ordinal logistic regression analyses were conducted to examine the relationship between UMCe and peak vertical GRF through the paretic lower extremity for the diagonal reach and STS tasks and stepping with the nonparetic lower extremity leading. An ordinal logistic regression analysis is like a linear regression analysis in that it assesses the nature and strength of the relationship between variables. The dependent variable in an ordinal logistic regression analysis, however, is categorical and ordered (eg, UMCe score) rather than continuous.
Based on preliminary regression analyses conducted on data from each session, we determined that the strength of the relationships between the clinical and laboratory measurements (ie, R2 values) did not follow any consistent patterns (eg, increasing, decreasing) over time. Consequently, we combined the data from all sessions to increase the size of the data set and to account for the non-independence of measures in our analyses.43 Data on the same individuals over time are not independent of one another. If this non-independence is unaccounted for in regression analyses, the standard errors of the parameter estimates (eg, the measure of slope in linear regression) may be underestimated, thereby increasing the likelihood of statistically significant findings. To address this issue, we estimated the standard errors using the Huber/White/sandwich estimator of variance.44 This estimator gives accurate assessments of the sample-to-sample variability of the parameter estimates even when the statistical model is not correctly specified.
For the linear regression analyses, statistical tests45 were conducted to verify that a linear model was appropriate (eg, that the relationship between the variables was linear and not exponential). Based on these tests and plots of the data, the independent variables for the linear regression analyses were transformed, if necessary. Transformation of the independent variable is one approach to take when conducting a linear regression analysis if the relationship between the dependent and independent variables is not linear. For example, one way to linearize a curvilinear relationship between a dependent variable and an independent variable is to square the independent variable.
For the ordinal logistic regression analyses, statistical tests were conducted to verify that the analysis was appropriate for the data and to assess the strength of the relationship between the dependent variable (UMCe) and the independent variable (force platform measure).43,46 Although the ordinal logistic regression analysis does not have an exact analog to the R2 generated in a linear regression analysis, various tests and pseudo R2 values are recommended to describe the strength of the relationship between the dependent and independent variables. We used the McFadden's pseudo R2 and the Bayesian Information Criterion (BIC) to assess model fit.46 We interpret these statistics in the "Results" section.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
|
|
|
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
Mean ST scores in our sample improved steadily from 1 to 6 months poststroke, with no evidence of reaching a plateau after 3 months poststroke. The ST may offer sufficient challenge to make it a useful measure for monitoring progress beyond 6 months poststroke. However, 9 (27%) of our participants received the lowest possible score on this test at all time points. If meaningful change was actually occurring in these patients, then floor effects on the ST may have been present. Because the ST is performed in standing without upper-extremity support, some patients, especially those in the earliest stages of recovery, may have balance problems or other impairments that result in a score of 0 on this test.
The other measure of paretic-limb loading investigated in this study (ie, UMCe) also was associated with peak vertical GRF beneath the paretic limb during functional tasks. The strongest association (pseudo R2=.34) was for the task of stepping up onto a step with the nonparetic limb leading, again supporting the idea that similar tasks may require similar abilities (eg, the ability to load the paretic limb). The correlations were much lower than for the ST, however, possibly because the UMCe was less similar to the biomechanical test. Another explanation for the weaker correlations is that the UMCe is a less-sensitive measure than the ST. The UMCe uses a 3-level scoring system, whereas the ST is a continuous measure, with scores that ranged from 0 to 20 in our study.
As with ST scores, UMCe scores generally improved from 1 to 6 months poststroke. Both floor and ceiling effects, however, were observed for UMCe. Nine participants (27%) received the lowest-possible score on the test for all 6 sessions, and 12 participants (36%) received the highest possible score on the test before the sixth session.
Our expectation that RR scores would be correlated with force platform measures of weight transfer generally was not supported. An important consideration in interpreting this result is the lack of consensus in the literature about appropriate biomechanical measures of weight transfer. Unlike paretic-limb loading, for which peak or mean vertical GRF is widely accepted for biomechanical measurement, weight transfer has been measured in a number of different ways. We chose force platform measurements of the magnitude and temporal characteristics of the transfer of weight from one leg to the other leg. Evidence of the relationship between these characteristics and performance of functional activities is limited.
The strongest relationship in our study between RR scores and force platform measures was the relationship with the change in vertical GRF beneath the paretic limb during the diagonal reach task. Although diagonal reach was the functional task that bore the strongest resemblance to the RR task, even this association was not strong. This may be attributable at least partially to differences in stance position and target location. For the RR task, each participant stood in a step stance, and the target was placed 15 cm beyond arm's length in front of the participant's paretic hip. For the biomechanical test (diagonal reach task), participants stood with their feet side by side, and the target was placed 1.4 arm-lengths away on the paretic side at a 45-degree angle from midline. The difference in foot position means that the participants may have been able to achieve high RR scores by reliance on the nonparetic leg for support in the step stance position. Other researchers48,49 have reported that the step stance position may encourage individuals to bear more weight on the posterior (in this case, nonparetic) leg. The difference in target location in our study means that the participants did not have to reach as far across midline to perform the RR task and, consequently, may have been able to compensate to some degree for their weight-transfer difficulties during this task.
The possibility that participants could perform well on the RR task despite minimal weight transfer also may account for the weak or nonexistent relationships between RR scores and the weight-transfer measurements obtained for the stepping tasks. The RR task and the biomechanical measurements of weight transfer obtained during the stepping tasks (weight-transfer time and paretic-limb medial-lateral impulse) were similar in terms of emphasizing temporal components of limb movements in standing. The stepping tasks, however, required complete transition from bipedal to unipedal stance. In order to successfully lift one foot from the floor, the center of mass must move laterally at least as far as the medial border of the supporting foot.15 Although additional research would be needed to determine the extent of displacement of the center of mass during the RR task, the large, stationary base of support limits the amount of displacement that is necessary for successful task performance. These results suggest that the RR may be more clinically important as a measure of dynamic postural control, as originally intended, than as an indicator of weight transfer, as we hypothesized here.
The results of our study confirm previous reports in the literature of difficulties maintaining single-limb stance on the paretic side after stroke.15,50 As evidenced by their UMCe scores (Tab. 2), 24 (73%) of our subjects were unable to bear full weight on the paretic side with the knee flexed at 1 month poststroke, and 9 (31%) remained unable to do so at 6 months poststroke. The mean score (SD) on the ST (nonparetic limb) was 5.8±4.6 at 2 months poststroke, very similar to the score of 6.5±5.1 reported by Bernhardt and colleagues39 for subjects at the same time point during stroke recovery, but far below the normative value of 17.7±3.2 reported by Hill et al.29
Changes over time in the force platform measures of limb loading (Tab. 2) and weight transfer (Tab. 3) are more difficult to interpret. These data reflect inclusion of scores of 0 for participants who were unable to complete the various functional tasks. As a result, the values in the tables may underestimate the true means for each force platform measure. This issue is particularly salient for measures such as weight-transfer time, for which smaller values would indicate shorter durations (ie, better performance). The mean values for the force platform measures presented in Tables 2 and 3, therefore, should be interpreted with these caveats in mind.
Limitations of the study included a relatively small sample size and a preponderance of participants with left-sided hemiparesis. With regard to the latter issue, we have found that speech and language impairments may complicate the process of obtaining informed consent and thereby limit recruitment of subjects with right-sided hemiparesis. The sample included individuals with varying degrees of lower-extremity motor impairment, ranging from mild to severe, as well as those who showed evidence of unilateral visual neglect at baseline. Forty-five percent of the sample was nonwhite. Consequently, our results should be generalizable to the population of people poststroke with respect to these characteristics.
| Conclusion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
| Footnotes |
|---|
The study was approved by the Biomedical Institutional Review Board (IRB) at the University of North Carolina at Chapel Hill and by the WakeMed Institutional Review Board.
This study was supported by the National Institutes of Health/National Institute of Child Health and Human Development (grant R03 HD43907). Dr Purser's work on this study was supported, in part, by a Mentored Research Career Development Award from the National Institutes of Health/National Center for Medical Rehabilitation Research/National Institute of Child Health and Human Development (1K01HD049593–01A1).
* Bertec Corp, 6717 Huntley Rd, Columbus, OH 43229. 
Peak Performance Technologies, 7388 S Revere Pkwy, Ste 603, Englewood, CO 80112.
Bing Yu, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7135.
StataCorp LP, 4905 Lakeway Dr, College Station, TX 77845.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  V. S. Mercer, J. K. Freburger, S.-H. Chang, and J. L. Purser Step Test Scores Are Related to Measures of Activity and Participation in the First 6 Months After Stroke Physical Therapy, October 1, 2009; 89(10): 1061 - 1071. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
