Smartphone-based Biofeedback System for Post-Stroke Rehabilitation Tele-Physiotherapy

Thesis Abstract

This study addresses the persistent gap in access to intensive post-stroke rehabilitation by evaluating a smartphone-based biofeedback system designed for tele-physiotherapy, aiming to enhance motor recovery, adherence, and self-management in home-based settings. The overarching objective is to determine whether real-time kinematic biofeedback delivered via a mobile application can improve upper and lower limb function, reduce disability, and sustain therapeutic engagement beyond traditional clinic visits. Specific objectives include (1) to assess the feasibility and usability of the smartphone biofeedback platform among stroke survivors and clinicians; (2) to evaluate changes in motor outcomes using standardized measures (Fugl-Meyer Assessment for Upper and Lower Extremities, Action Research Arm Test, and Gait Assessment scale) over a 12-week intervention; (3) to examine adherence, motivation, and self-efficacy as mediators of functional recovery; (4) to identify predictive factors associated with clinically meaningful gains; and (5) to explore clinicians’ and patients’ experiences through qualitative inquiry to inform system refinement. The study employs a mixed-methods design, integrating a randomized controlled trial with an embedded qualitative process. The population comprises adults aged 40–80 years, 6–24 months post-ischemic stroke, recruited from three urban rehabilitation centers. A sample of 120 participants will be randomized to either the smartphone-based biofeedback tele-physiotherapy group (n=60) or standard home-based exercise guidance (n=60) for 12 weeks, with follow-up at 24 weeks. The intervention arm utilizes a smartphone application that captures movement via the device’s inertial measurement unit (IMU) sensors, provides real-time visual and haptic biofeedback, prescribes personalized exercise regimens, and transmits progress data to a remote physiotherapist. Data collection instruments include the Fugl-Meyer Assessment (FMA), Wolf Motor Function Test (WMFT), Timed Up and Go (TUG), adherence logs from the app, the Stroke Self-Efficacy Scale, and System Usability Scale (SUS). Qualitative data will be gathered through semi-structured interviews with a purposive subsample of 20 participants and 10 physiotherapists, analyzed via thematic analysis. Quantitative analysis will involve intention-to-treat principles. Primary outcomes (FMA and WMFT scores) will be analyzed using repeated-measures ANOVA and linear mixed-effects models to account for intra-subject correlations and missing data. Mediation analysis will explore whether adherence and self-efficacy mediate functional outcomes, using PROCESS macro for bootstrapping. Regression analyses will identify predictors of clinically meaningful improvement, defined as a ?10-point improvement on the FMA and/or a 15% improvement in WMFT. The qualitative data will be analyzed inductively to generate themes related to usability, perceived effectiveness, barriers to adoption, and suggestions for enhancement, with triangulation against quantitative findings. Expected findings include statistically significant greater gains in motor function (FMA and WMFT) in the tele-physiotherapy group compared with controls at 12 weeks, sustained or further improved outcomes at 24 weeks, higher adherence rates, and enhanced self-efficacy. It is anticipated that higher engagement with real-time feedback, exercise intensity adherence, and timely remote supervision will mediate improved outcomes. The study also expects to identify subgroups (e.g., based on time since stroke, baseline impairment, or age) that benefit disproportionately, informing targeted deployment. The contribution to knowledge encompasses (1) empirical evidence on the efficacy of smartphone-based biofeedback to augment tele-physiotherapy after stroke; (2) a validated framework for remote motor rehabilitation that integrates objective kinematic data with patient-reported outcomes and clinician feedback; (3) insights into usability and implementation science for mobile rehabilitation platforms in routine care; and (4) theoretical advancement by testing mechanisms of action grounded in self-determination theory and the Technology Acceptance Model within a neurorehabilitation context. Overall conclusion expects that integrating real-time biofeedback with remote clinician oversight can enhance motor recovery, adherence, and self-management in post-stroke rehabilitation, thereby offering a scalable solution to improve access and outcomes. Recommendations will address optimization of feedback modalities, data privacy and security, integration with electronic health records, strategies to sustain long-term engagement, and considerations for broader deployment across diverse healthcare settings.