What are the valued attributes and perceived risks of harm in digital technologies and AI-enabled digital coaches among youth living with chronic musculoskeletal pain? An exploratory, mixed-methods study.

Youth (16-24 years) experiencing chronic musculoskeletal pain (CMP) value access to digitally-supportive care. Adopting a three-phase design, we aimed to identify and validate youth's preferences for digital health solutions (DHSs) to support their CMP self-care and measure their preferences for artificial intelligence (AI)-enabled digital coaches and perceived risks of harm. Phase 1 involved focus groups with 20 youth experiencing CMP, applying content analysis to explore the attributes they consider when choosing digitally-supportive care. Phase 2 involved a two-round eDelphi (n=20 youth) to validate attributes identified in Phase 1, rank the importance of digital coach-related attributes, and identify risks of harm with digitally-supportive CMP care. Phase 3 involved two discrete choice experiments (n=41 youth) to measure preferences for digital coach-related attributes and risks of harm. Phase 1 produced 31 discrete attributes of DHSs (11 digital coach-related). Phase 2 validated 31 attributes and identified six risks of harm. Phase 3 identified three most preferred digital coach-related attributes (sample mean[SD] weights representing relative importance): 'Generates a pain self-care plan' (19·6[14·5]%); 'Learns my preferences for my pain self-care' (15·9[12·9]%); and 'Supports my healthcare visits' (11·1[12·4]%). The three most concerning risks of harm were: 'Quality of pain care advice and guidance' (31·8[17·3]%); 'Information being consistent with advice from my healthcare team' (25·8[17·5]%); and 'Data security protections' (14·8[17·6]%). Youth consider a range of attributes and potential harms of DHSs to support their CMP care; a subset appears to be most important. These attributes and risks of harm should be considered in co-design of digitally-supportive CMP care for youth. PERSPECTIVE: The results identify that a subset of digital coach-related attributes and risks of harm dominate youth's preferences for digitally-supportive chronic musculoskeletal pain care. These attributes and risks of harm should be considered in co-design of digital health solutions, and in positioning a solution within a clinical care ecosystem.