User Needs for an Artificial Intelligence-Empowered Basic Activities of Daily Living Training System After Stroke: A Qualitative Study.

PURPOSE: To explore patients' and primary caregivers' willingness to use an artificial intelligence (AI)-empowered basic activities of daily living (BADL) training system and their needs, concerns, and expectations regarding the system. PATIENTS AND METHODS: A descriptive qualitative study was conducted in the Department of Neurology at the Second Affiliated Hospital of Fujian Medical University, Fujian, China. Patients with poststroke hemiplegia and primary caregivers were recruited using purposive sampling. Semistructured, face-to-face interviews were conducted with 11 patients with poststroke hemiplegia and 13 primary caregivers to capture both patient and caregiver perspectives. Data were collected on participant-reported BADL difficulties, willingness to use the AI-empowered BADL training system, perceived needs and concerns, and expectations for system design. Data were analyzed using inductive thematic analysis. RESULTS: Four themes were identified: 1) BADL limitations created shared difficulties for patients and caregivers; 2) acceptance of the AI-empowered BADL training system depended on usefulness and functional fit; 3) participants expected low-barrier, understandable, and sustainable training support; and 4) the system was viewed as a supplement to professional rehabilitation and family caregiving. CONCLUSION: An AI-empowered BADL training system may be acceptable to patients with poststroke hemiplegia and their primary caregivers if it is safe, reliable, easy to use, and appropriate to patients' functional status. These findings may guide the future development of task-oriented AI-empowered BADL training systems that support patient-caregiver collaboration, safety assurance, sustained engagement, and human-AI collaboration in rehabilitation.