Development of a Functional Recovery-Factors Model for Post-Stroke Physiotherapy Outcomes

Thesis Abstract

Stroke rehabilitation remains inconsistently optimized across settings, with heterogeneous outcomes attributable to patient-, process-, and environment-related factors that are not adequately integrated into existing theoretical frameworks. This study develops and validates a Functional Recovery-Factors Model (FRFM) to predict and enhance post-stroke physiotherapy outcomes by synthesizing motor recovery theories with socio-ecological determinants. The aim is to construct a parsimonious, clinically usable model that maps modifiable factors to functional recovery trajectories and informs individualized intervention planning. Specific objectives are to (1) identify core determinants of functional recovery across the acute, subacute, and chronic phases; (2) integrate physiological, cognitive, psychosocial, and environmental variables into a cohesive theoretical framework grounded in the International Classification of Functioning, Disability and Health (ICF) and Neurorehabilitation principles; (3) evaluate the predictive power of the FRFM for activities of daily living (ADL) and gait independence; (4) examine the mediating and moderating roles of therapy dose, adherence, and motor learning strategies; and (5) develop a decision-support tool for therapists to tailor physiotherapy plans. A sequential, mixed-methods design is employed. In the qualitative phase, purposively sampled stroke survivors (n=40) and physiotherapists (n=20) across three tertiary hospitals will participate in semi-structured interviews and focus groups to elicit key recovery-factors and therapeutic mechanisms, analyzed via thematic analysis informed by grounded theory principles. In the quantitative phase, a multicenter cohort of 600 stroke patients within two weeks post-onset will be followed for six months. Data collection will combine standardized measures (Fugl-Meyer Assessment for upper and lower extremity motor function, functional independence measure, Timed Up and Go, and Stroke Impact Scale) with process indicators (therapy session frequency, duration, and specific interventions) and environmental variables (caregiver support, home setup). Instruments will be validated for construct and criterion validity, with reliability assessed via Cronbach’s alpha and inter-rater reliability where applicable. Analytical strategies include structural equation modeling (SEM) to test the FRFM’s fit and pathways linking determinants to functional outcomes, hierarchical linear modeling (HLM) to account for nested data across settings and over time, and multiple regression analyses to identify independent predictors of ADL and ambulation. Moderation analyses will explore how therapy dose and adherence alter the strength of relationships within the model. Mediation analyses will consider motor learning strategies as conduits between therapy exposure and functional change. A priori hypotheses posit that (a) motor impairment, motor learning capability, cognitive function, mood, social support, and environmental factors collectively predict functional recovery; (b) higher therapy dose with task-specific, bilateral training and feedback enhances motor gains beyond impairment-focused approaches; and (c) FRFM-derived profiles will yield superior predictive accuracy (R2 > 0.40 for ADL and gait outcomes) compared with conventional models. Expected findings include a robust, multi-constituent FRFM with validated structural paths demonstrating the relative contributions of biological, personal, and contextual factors to functional recovery; identification of key leverage points (e.g., early initiation, intensity, task-specific practice, caregiver involvement) that maximize outcomes; and evidence that personalized FRFM-informed plans lead to greater improvements in ADL and independence trajectories. The study will contribute to knowledge by operationalizing an integrative, theory-driven framework that transcends siloed rehabilitation paradigms, bridging motor recovery science with health services and social determinants of health. It will produce a practical decision-support toolkit and scoring algorithm for routine clinical use, with user guidelines and an implementation plan for scaling across rehabilitation networks. The main conclusion anticipated is that functional recovery after stroke is best optimized when physiotherapy is guided by the FRFM, which integrates motor, cognitive, psychosocial, and environmental determinants into individualized, dose-optimized intervention plans. Recommendations include incorporating FRFM training into physiotherapy curricula, embedding the model within electronic health records for real-time decision support, and conducting a pragmatic trial to evaluate the model’s impact on long-term functional independence and health-related quality of life.