Factors predicting stroke disability at discharge: A study of 793 Chinese

Objective: To identify factors predicting stroke disability at discharge in a Chinese population. Design: Retrospective analysis of data collected from stroke patients. Setting: A 25-bed stroke rehabilitation unit in Hong Kong. Participants: A total of 793 Chinese patients with acute stroke consecutive ly admitted for inpatient rehabilitation. Interventions: All patients received traditional rehabilitation therapies i ncluding physical, occupational, and speech therapies when appropriate. Main Outcome Measures: Disability was measured with the Barthel index (BI), and mild disability at discharge was defined as a BI score of greater than or equal to 15. Odds ratios (ORs) and 95% confidence intervals (CIs) were computed. Results: Logistic regression analysis revealed that a BI score of greater t han or equal to 15 points at admission strongly predicted that the patient' s BI at discharge would be greater than or equal to 15. For those whose BI score at admission was less than 15, these factors correlated negatively wi th a discharge BI of greater than or equal to 15: BI at admission of <5 (OR .08, CI .04-.17); National Institutes of Health stroke scale at admission of >7 (OR .23, CI .12-.43); urinary incontinence at admission (OR .35, CI . 21-.60); age greater than or equal to 65 years (OR .44, CI .25-.77); and ab breviated mental test at admission of <7 (OR .56, CI .33-.94). Conclusions: For Chinese stroke patients, the disability at admission is th e most important predictor for disability at discharge. Patients with very severe disability, severe neurologic impairment, urinary incontinence, old age, and impaired cognition at admission are less likely to recover to mild disability at discharge. Although hemorrhagic stroke is more common among Chinese populations, it is not an independent predictor for disability at d ischarge.