Abstract
BACKGROUND: Individualizing interventions is imperative to optimize physical activity in people with chronic stroke. This secondary analysis grouped individuals with chronic stroke into clinical profiles based on baseline characteristics and examined if these clinical profiles preferentially benefitted from a specific rehabilitation intervention to improve daily step-activity. METHODS: Participants had non-cerebellar strokes ≥6 months prior to enrollment, were 21-85 years old, had walking speeds of 0.3-1.0 m/s, and took <8,000 steps-per-day. Participants were randomized to 1 of 3 interventions: high-intensity treadmill training (FAST), a step-activity behavioral intervention (SAM), or a combined intervention (FAST+SAM). The primary outcome was the interaction of latent class (clinical profile) and intervention group (FAST, SAM, FAST+SAM) on a change in steps-per-day. Key clinical characteristics to identify the latent classes included walking speed, walking endurance, balance self-efficacy, cognition, and area deprivation. RESULTS: Of the 190 participants with complete pre- and post-intervention data (mean [SD] age, 64 [12] years; 93 females [48.9%]), 3 distinct profiles of people with chronic stroke were identified. Within our sample, class 1 had the lowest walking capacity (speed and endurance), lowest balance self-efficacy, and highest area deprivation, and had the greatest change in step-activity when enrolled in the SAM (mean[95%CI], 1624 [426 - 2821]) or FAST+SAM (1150 [723 - 1577]) intervention. Class 2 had walking capacity, baseline steps-per-day, and self-efficacy values between Class 1 and 3, and had the greatest change in step-activity when enrolled in the SAM (2002 [1193-2811]) intervention. Class 3 had the highest walking capacity, highest self-efficacy, and lowest area deprivation and the greatest change in step-activity when enrolled in the FAST+SAM (1532 [915-2150]) intervention. CONCLUSIONS: People with chronic stroke require different interventions to optimize a change in step-activity. Clinicians can use clinically relevant measures to personalize intervention selection to augment step-activity in people with chronic stroke. TRIAL REGISTRATION: NCT02835313; https://clinicaltrials.gov/ct2/show/NCT02835313.