Henry Hoffman Saturday, June 9th, Life after a stroke can be challenging. Many patients wonder if they will ever fully recover their muscle coordination, or how long or difficult the process of recovery may be. Fortunately, the field of occupational and physical therapy has come a long way in developing approaches that help patients regain controlled muscle movements after a stroke. There are seven recognized stages of stroke recovery through which most patients progress.
|Published (Last):||13 February 2017|
|PDF File Size:||9.51 Mb|
|ePub File Size:||19.25 Mb|
|Price:||Free* [*Free Regsitration Required]|
The Brunnstrom recovery stages the BRS consists of 2 items assessing the poststroke motor function of the upper extremities and 1 assessing the lower extremities. The 3 items together represent overall motor function. Although the BRS efficiently assesses poststroke motor functions, a lack of rigorous examination of the psychometric properties restricts its utility.
We aimed to examine the unidimensionality, Rasch reliability, and responsiveness of the BRS, and transform the raw sum scores of the BRS into Rasch logit scores once the 3 items fitted the assumptions of the Rasch model. We used Rasch analysis to examine the unidimensionality and Rasch reliability of both upper-extremity items and the 3 overall motor items of the BRS.
The Rasch reliabilities of the upper-extremity items and overall motor items were high 0. The results of our study support the unidimensionality, Rasch reliability, and responsiveness of the BRS. Moreover, the BRS can be transformed into an interval-level measure, which would be useful to quantify the extent of poststroke motor function, the changes of motor function, and the differences of motor functions in patients with stroke. Motor recovery is one of the most important treatment goals for patients with stroke.
The Brunnstrom recovery stages BRS is a short and easily administered measure for assessing motor function. These items are usually used individually to describe the motor function i.
The BRS has good item-level psychometric properties. However, 3 weaknesses restrict the utility of the BRS. First, the unidimensionality of the BRS has not yet been investigated.
It is unknown whether the 3 items of the BRS assess the same construct and whether the scores of the items can be summed to represent overall motor function. Moreover, sum scores could be an outcome indicator because any progress made on each item by a patient could be detected, which is useful for monitoring a patient's overall change over time and determining the effects of intervention.
Therefore, validation of the unidimensionality of the BRS is warranted. Second, it is unknown whether the BRS is as responsive as lengthier measures of motor function, such as the stroke rehabilitation assessment of movement STREAM , which has moderate to large responsiveness.
However, previous studies have shown that the responsiveness of the short-form format of a measure can be comparable to that of the long-form format in a group of patients. Third, the BRS is rated on an ordinal scale rather than on an interval scale. An ordinal scale identifies the order of the values, but the differences between the values remain unknown.
Rasch analysis is based on a mathematical model that can estimate person ability the motor function in our study and item difficulty the level of difficulty of each item simultaneously, and then place the person ability and the item difficulty on the same interval scale. First, items that fit the assumptions of the Rasch model are unidimensional, which facilitates examination of the unidimensionality of a measure.
Second, Rasch analysis helps users transform an ordinal-level measure into an interval measure. Rasch scores, scored on an interval scale, can be used to represent the estimated person ability and item difficulty.
Third, Rasch analysis provides Rasch reliability, which is an indicator of measurement error of the BRS scores. Although BRS efficiently assesses motor functions, its utility is restricted because its psychometric properties have not been rigorously examined. This study had 2 purposes. First, we aimed to examine the psychometric properties of the BRS, including the unidimensionality, Rasch reliability, and responsiveness. Second, once the 3 items fitted the assumptions of the Rasch model, we aimed to transform the sum scores of the 3 items an ordinal scale into Rasch logit scores an interval scale.
We hypothesized that both the upper-extremity items and the overall motor items of the BRS were unidimensional, the responsiveness of the BRS was comparable with that of the STREAM, and the BRS could be transformed from an ordinal-level measure into an interval-level measure. We retrospectively retrieved a set of admission and discharge data from to from medical records of the occupational therapy OT department of a medical center. We selected medical records according to the following inclusion criteria: patients who had diagnoses of stroke, and patients who underwent BRS evaluations at admission.
Diagnoses of stroke were based on the International Classification of Disease, Ninth Revision, Clinical Modification Codes , including cerebral hemorrhage , cerebral infarction , or others , , , , and This study was approved by the institutional review board of the medical center.
The BRS was designed to describe a sequence of extremity motor recovery after stroke based on the synergy pattern of movement that develops during recovery from a flaccid limb to near-normal and normal movement and coordination. Higher levels represent better motor function. Clinicians rate a patient's stage based on the patient's spasticity and movement. It takes less than 10 minutes to complete the evaluation. In our study, levels I to VI were respectively recoded into scores of 0 to 5 for Rasch analysis.
The STREAM was designed to provide a comprehensive, objective, and quantitative evaluation of the motor functioning of patients with stroke. The STREAM contains 30 items divided into 3 subscales: 10 items for voluntary motor ability of the upper-extremity, 10 items for voluntary motor ability of the lower-extremity, and 10 items for basic mobility.
The upper-extremity and the lower-extremity subscales were used in our study. A 3-point ordinal scale is used for scoring voluntary movement of the extremities. Descriptive statistics were used to analyze the characteristics of the patients and the score distributions of the 3 BRS items. In addition, percentage floor and ceiling effects i. We calculated the percentages of patients obtaining the highest and the lowest raw sum admission scores on the upper-extremity items and overall motor items.
We examined the unidimensionality and Rasch reliability of the upper-extremity items and the overall motor items. We did not examine the lower extremity item BRS-L because at least 2 items were required for Rasch analysis. Rasch analysis with the partial credit model PCM was used for examining data—model fitting. Unidimensionality was examined using all BRS admission data.
To examine the unidimensionality of the BRS, we used infit and outfit statistics, and principal component analysis PCA. First, the infit and outfit statistics were used to examine whether the item responses fit the expectations of the PCM model. The acceptable ranges of both infit and outfit MNSQ values for each item are from 0.
PCA was further applied to examine the standardized residuals observed BRS scores minus expected scores. We also investigated the level of difficulty of the 3 items and the appropriateness of the response category of each item. The level of difficulty was calculated and expressed as a logit score along with Rasch analysis. The appropriateness of the response category was determined by the step difficulties in each item, which should be in order for the design of the response categories to be satisfactory.
Disordering of the step difficulties in an item indicates the need for adjustment of the response category. Notable gaps along the step difficulty continuum indicate that additional response categories or items are needed to distinguish patients falling in the gaps.
Moreover, we examined whether patients from different subgroups age groups or sex but at the same ability level had equal probabilities of responding positively to the three items. Item bias existed when the DIF contrast was greater than 0. We examined the Rasch reliability for BRS admission data. We considered Rasch reliability coefficients higher than 0. When the upper-extremity items and overall motor items fitted the assumptions of the Rasch model, the sum scores of the upper-extremity items and overall motor items were transformed to Rasch scores, respectively.
First, the sum scores of the upper-extremity and overall motor items on admission and at discharge were transformed into BRS Rasch scores using Rasch scores obtained by the quantification of the BRS. The paired t -test was used to determine the statistical significance of the change in scores. The ES is a measure of change obtained by dividing the mean change in scores between assessments on admission and at discharge by the standard deviation SD of the assessment on admission. The SRM is the mean change in scores between two measurements divided by the SD of the changes scores.
ES and SRM values of 0. Each bootstrap sample was the same size as the original sample recruited in the responsiveness analysis. The bootstrap samples produced pairs of differences in the ESs and the SRMs of both measures. After sorting these differences from lowest to highest, we examined whether the value 0 was included in the 26th and th observations i. The descriptive statistics and responsiveness analysis were analyzed using SPSS The Rasch analysis was performed using Winstep 3.
Admission medical records of patients with stroke were available. Patients who had better motor function levels V and VI in the 3 BRS items at admission tended to be lost to follow-up, since patients scoring in levels V and VI in group 1 outnumbered those in group 2. Only upper-extremity motor function showed a notable ceiling effect, with There were no obvious floor effects in either upper-extremity motor function or overall motor function.
PCA results showed that PCA showed that The biggest difference in difficulty among the 3 items was 0. There was no disordering of the step difficulties in the 3 items.
Similar results were found for the overall motor items. The person—item map of the upper-extremity items left and overall items right of the BRS. Accordingly, the 3 items of the BRS were unidimensional. In addition, the hierarchy of items was identified, which aids in the understanding of progress in the recovery of motor function. No significant DIFs due to age or sex were found. Rasch reliability coefficients of the upper-extremity motor function were 0.
Rasch reliability coefficients of overall motor function were 0. Higher scores indicated better motor function. In addition, because the BRS contains only 3 items, the standard errors of Rasch-transformed scores were large, ranging from 1.
Accordingly, the raw sum scores of the BRS have been transformed into Rasch scores to represent the motor function of the upper extremities and overall motor function.
Moreover, the standard errors of each Rasch score have also been identified. The results have 2 implications. First, the upper-extremity items and the overall motor items are both unidimensional, supporting the use of their raw sum scores to respectively represent upper-extremity motor function and overall motor function. Therefore, in addition to each item score describing the hand, arm, or leg motor function of a patient, users now have other choices to quantify the patient's poststroke motor function based on their needs such as presenting an overall motor function or investigating the treatment effect on overall motor function.
Therefore, the Rasch scores of the upper-extremity items and overall motor items appear more appropriate for reflecting the extent of upper-extremity motor function and overall motor function, monitoring progress, and comparing the differences of upper-extremity motor function and overall motor function.
The Rasch scores of the BRS are strongly recommended for future users.
Improving the utility of the Brunnstrom recovery stages in patients with stroke
When a stroke occurs, typically it affects one side of the body. The Brunnstrom approach describes the sequence of motor development and reorganization of the brain after stroke. You can think of it as a built in organizational system. Our brain automatically recruits lower functioning reflexes, just to get any movement. Then it begins to sort out what is useful and what connections we need to build at a higher level. It's also a way to know what movement actions to take advantage of during your post stroke rehabilitation. During this stage the muscles on your affected side aren't able to move and they might feel limp and floppy.
Clinical characteristics affecting motor recovery and ambulation in stroke patients
Brunnstrom hand and upper extremity values in females were lower than in males. Complex regional pain syndrome was observed at a level of Brunnstrom hand stage was lower in complex regional pain syndrome patients than in those without the syndrome. Shoulder pain was present in Brunnstrom lower extremity values and functional ambulation scale scores were higher in rehabilitated than in non-rehabilitated cases. Shoulder pain and lower Brunnstrom hand stages were related to the presence of complex regional pain syndrome. The World Health Organization WHO has defined stroke as a focal or generalized neurological deficit related to vascular causes that develop suddenly with the impairment of cerebral function, lasting for more than 24 h, which may result in death within this period 1.