Effects of a Self-Exercise Program on Activities of Daily Living in Patients After Acute Stroke: A Propensity Score Analysis Based on the Japan Association of Rehabilitation Database

Abstract

ObjectiveTo investigate whether self-exercise programs for patients after stroke contribute to improved activities of daily living (ADL) at hospital discharge. DesignRetrospective, observational, propensity score (PS)–matched case-control study. SettingGeneral hospitals. ParticipantsParticipants included patients after stroke (N=1560) hospitalized between January 3, 2006, and December 26, 2012, satisfying the following criteria: (1) data on age, sex, duration from stroke to hospital admission, length of stay, FIM score, modified Rankin Scale (mRS) score, Glasgow Coma Scale score, Japan Stroke Scale score, and self-exercise program participation were available; and (2) admitted within 7 days after stroke onset, length of stay was between 7 and 60 days, prestroke mRS score was ≤2, and not discharged because of FIM or mRS exacerbation. A total of 780 PS-matched pairs were selected for each of the self-exercise program and no–self-exercise program groups. InterventionSelf-exercise program participation. Main Outcome MeasuresAt discharge, FIM motor score, FIM cognitive score, FIM motor score gain (discharge value − admission value), FIM motor score gain rate (gain/length of stay), a binary variable divided by the median FIM motor score gain rate (high efficiency or no–high efficiency), and mRS score. ResultsPatients were classified into a self-exercise program (n=780) or a no–self-exercise program (n=780) group. After matching, there were no significant between-group differences, except motor system variables. The receiver operating characteristic curve for PS had an area under the curve value of .71 with a 95% confidence interval of .68 to .73, and the model was believed to have a relatively favorable fit. A logistic regression analysis of PS-matched pairs suggested that the self-exercise program was effective, with an overall odds ratio for ADL (high efficiency or no–high efficiency) of 2.2 (95% confidence ratio, 1.75–2.70). ConclusionsSEPs may contribute to improving ADL.