An AI-based solution for physical activity after stroke: co-production and evaluation
This project aims to co-design, develop and test the feasibility of an AI-based chatbot providing tailored support for self-management of physical activity among people who have experienced stroke or transient ischaemic attack (TIA). The project will explore how AI chatbot technology can be integrated into physical activity self-management support, including how it can respond to individual needs, barriers and changing circumstances over time.
About the project
The project builds on previous work by the research group developing and testing digital support for physical activity after stroke or TIA. AI creates new opportunities to provide personalised support for behaviour change that is continuously available and responsive to a person’s current needs and circumstances. AI-based chatbots are one potential approach, enabling interactive, individualised support through natural-language communication, responses to users’ questions, and tailored information and guidance.
The project will be conducted in three phases:
Phase 1 – Understanding needs: This phase explores what people post-stroke/TIA and physiotherapists perceive as important support needs and desirable features in an AI chatbot for self-management of physical activity. Focus groups and individual interviews examine how AI-based support could address individual needs, barriers and contextual factors.
Phase 2 – Co-production and development: Findings from Phase 1 inform the co-production of an AI chatbot through an iterative, user-centred process involving people post-stroke/TIA, physiotherapists, patient organisations, digital health experts and researchers. The project examines how an AI chatbot can be developed to provide relevant, safe and personalised support. Development includes defining its functions, content and safety boundaries, followed by prototyping, user testing and iterative adaptation.
Phase 3 – Feasibility testing: The feasibility of the AI-based support will be evaluated in a randomised controlled trial. The study assesses whether the intervention is feasible and acceptable to deliver and use, including engagement with the chatbot and its potential to support self-management of physical activity after stroke/TIA.
Significance
Physical activity is important for recovery and secondary prevention after stroke or TIA, yet many people remain inactive and access to long-term support is limited. AI chatbots could provide accessible, personalised and ongoing support between contacts with healthcare professionals, complementing rather than replacing physiotherapists and existing services. However, their appropriate role, safety, accessibility and acceptability in stroke care remain unclear.
This project will address these gaps by co-producing an AI chatbot with end-users and key stakeholders and evaluating its usability, safety and feasibility. If successful, the project could contribute to more person-centred and sustainable long-term support for physical activity after stroke or TIA.