F-3-12-03
PREDICTING DISABILITY IN STROKE - A CRITICAL REVIEW OF LITERATURE
Gert Kwakkel, Robert C Wagenaar, Boudewijn J. Kollen, Gustaaf J. Lankhorst.
Dept. Physical Therapy, University Hospital Vrije Universiteit, Amsterdam, The Netherlands.
PURPOSE: Studies on prognosis of functional recovery after stroke meting methodological criteria were identified, in order to find predictors of functional outcome.
METHODS: Data sources comprised a computer-aided search of published prognostic studies and references of literature used in prognostic studies. Seventy eight studies were tested for adherence to the following key methodological criteria, i.e.: reliability and validity of assessment instruments for dependent and independent variables; inclusion of inception cohort; adequate and uniform endpoint of observation; control for drop outs; applied statistics for relations between dependent and independent variables; appropriate sample size in relation to number of determinants; control for multicolinearity; specification of the patient characteristics age, type, recurrent stroke and localization of stroke; description of interfering treatment effects; and, cross-validation of the prediction model in a second independent group of patients.
RESULTS: Only 3 studies satisfied 9 out of 11 criteria and 10 studies 8 criteria for the determination of valid prognostic research. Combining results of these studies indicates that the following variables are important predictors for functional recovery after stroke, i.e.: age; previous stroke; admission ADL score; urinary continence; consciousness at onset; disorientation in time and place; severity of paralysis; sitting balance; impact of social support; and metabolic rate of glucose outside the area infarct of hypertensive patients.
CONCLUSION: Despite heterogeneity of patient population and methodological shortcommings of the included prognostic studies a number of determinants of functional recovery could be pinpointed.
F-3-12-04
ACCURACY OF RECOVERY EVALUATING SYSTEM-4 (RES-4) FOR THE PREDICTION OF STROKE PATIENTS
Tobimatsu Y. Nakamura R. (National Rehab. Center) Isagoda A. (Sendai University)
We examined retrospectively the accuracy of prediction of Birthel Index (BI) score of stroke patients, 4, 8 and 12 weeks after admission by RES-4, applying it to 360 patients of seven hospitals. Using 31 items such as BI, Manual Function Score (MFS) and demographic data including neurological status at the admission, RES-4 computed the estimated BI score and others of 4, 8 and 12 weeks later. The difference between the measured score and the corresponding estimated score (△ D) of 4, 8 and 12 weeks was calculated for each patient. Regarding |△ D|<12.5 as an accurate prediction, RES-4 revealed accurate prediction in more than two-thirds of the patients. |△ D| were rather large in patients admitted within 2 weeks since stroke onset, and also those with BI score below 30. In conclusion, the prediction of BI score by RES-4 could be practically useful in stroke rehabilitation, although caution is necessary when applying RES-4 to the patients within 2 weeks since the stroke onset.
