Predictors of Enrollment in an Exercise Study Among Stroke Survivors

Abstract

Background: Stroke is the fourth leading cause of death in the United States and major contributor to physical and cognitive disability among the older adult population. Exercise is an important aspect of stroke rehabilitation to reduce disability. Objective: To determine if functional disability, physical function, cognitive impairment, age or gender are predictive of enrollment in an exercise study among stroke survivors. Methods: Cross-sectional design using data from stroke survivors (n=166) who consented to be screened for the “Tai Chi for Stroke Survivors Study.” All participants completed a brief survey to determine eligibility, prior to screening. Screening tests include the Modified Rankin Scale (mRS, functional disability), Short Physical Performance Battery (SPPB, physical function), and Mini-Mental Status Exam (MMSE, cognitive function). First, we conducted binary logistic regression using mRS, SPPB-balance, gait speed and leg strength test times, MMSE, age and gender, as single predictors of study enrollment. Then, multiple logistic regression using significant predictors determined the overall model. Results: Participants (n=166) were 69±11 years old, and reported mild-moderate disability (mRS=2.0±0.8), had some impairments in physical functioning (SPPB=7.0±2.6), diminished leg strength (5-timed chair stands=19.3±7.2 seconds), and slow gait speed (0.76±0.3 meters/second) but, were without cognitive impairments (MMSE=27.8±3.3). Significant predictors of study enrollment were leg strength [X²(1)= 16.37, p<0.01] and gait speed [X²(1)= 6.89, p<0.01]. Multiple logistic regression indicated that leg strength and gait speed together were predictive of enrollment [X² (2)=22.38, p<0.01]. Prediction success was 77%. Leg strength was a significant predictor in the model (Wald=6.2, p=0.01), indicating that as test completion time increases, the odds of enrollment increases [Exp(B)=1.12, Cl=1.03-1.23]. Gait speed was significant predictor in the model (Wald=5.6, p=0.02), indicating that as gait speed increases, the odds of enrollment decreases [Exp(B)=0.14, Cl=0.03-0.71]. A total of 60% (100/166) of screened participants (women=46/100) enrolled in our study. Conclusion: Physical function has been shown to predict disability and overall decline in health and well-being in stroke survivors with mild-moderate disability. The SPPB is simple to administer, taking 5-10 minutes to complete. This test readily identified stroke survivors with diminished leg strength and slower gait speed, and was predictive of enrollment into the exercise study.*