Prediction of length of stay for stroke patients

Objectives - To examine the factors that influence acute and total length of stay (LOS) for stroke patients. Materials and methods - The basis of this investigation was a population-based cohort of first-ever stroke patients (n = 388). Subjects were survivors of the initial hospitalization (n = 295). Age, sex, social factors, risk factors, dementia, stroke type, and stroke severity, measured with the NIH stroke scale (NIHSS), were registered. Results - Mean acute LOS was 12 days and mean total LOS was 29 days. Independent predictors of acute LOS were stroke severity, lacunar stroke, prestroke dementia, and smoking. Independent predictors of total LOS were stroke severity and prestroke activities of daily living (ADL) dependency. The NIHSS items that best correlated with LOS were paresis, unilateral neglect and level of consciousness. Conclusions - Stroke severity is a strong and reliable predictor of LOS. The results of this study can be used as a baseline for evaluating cost-effectiveness of stroke care changes, e.g. organizational changes or evaluation of new drugs.