Reducing sedentary behaviour in patients after stroke (RECREATE): exploratory economic evaluation from an external pilot cluster randomised trial of the Get Set Go intervention

Background: Sedentary behaviour is common among stroke survivors and contributes to poor long-term health outcomes and high societal costs. The Get Set Go (GSG) intervention was developed to reduce sedentary behaviour through a whole-service approach spanning inpatient and community stroke care. This paper reports an exploratory economic evaluation conducted alongside the external pilot cluster randomised trial of GSG within the RECREATE programme. Objectives: The evaluation examined the potential cost-effectiveness of GSG compared to usual care and assessed the feasibility of alternative methods for capturing economic data. Specific aims included estimating costs and outcomes from NHS/personal social services (PSS) and societal perspectives, assessing the agreement between self-reported and routinely collected Hospital Episode Statistics (HES) data, and exploring the implications of different data sources and follow-up periods for future evaluations. Methods: Stroke services across England and Scotland were randomised (eight intervention, seven control), recruiting 334 stroke survivors. Costs were derived from self-reported service use (Client Service Receipt Inventory) and HES data, and outcomes were measured using the Nottingham Extended Activities of Daily Living Scale (NEADL) and EQ-5D-5L-derived quality-adjusted life years (QALYs) over 12 and 24 months. Analyses followed an intention-to-treat approach using bootstrap mixed-effects models. Results: From both NHS/PSS and societal perspectives, mean costs were higher and outcomes (QALYs and NEADL) were lower in the GSG arm compared to usual care, suggesting that GSG was unlikely to be cost-effective. Self-reported data underestimated secondary care use compared with HES, producing lower total cost estimates. While HES-derived costs were higher overall, they did not replicate the higher costs observed for GSG using self-report data. Substantial missing data and reduced follow-up due to the COVID-19 pandemic limited interpretation of long-term cost-effectiveness. Conclusions: This exploratory analysis indicates that GSG is unlikely to be cost-effective and highlights methodological lessons for future trials. Routine data, such as HES, may provide a more reliable and less burdensome source for estimating secondary care costs, whereas self-report remains important for capturing unpaid care and community-based resource use. A hybrid data collection approach is recommended for future economic evaluations in stroke rehabilitation.