Afshar ambient AI (well-being)

A Pragmatic Randomized Controlled Trial of Ambient Artificial Intelligence to Improve Health Practitioner Well-Being

Patient / Population Intervention / Exposure Comparison Outcome

In 66 ambulatory care practitioners, an ambient AI scribe drafting visit notes reduced work exhaustion and interpersonal disengagement compared with documentation without ambient AI.

N
66patients
Design
Stepped-wedge, individually randomised pragmatic trial (24 weeks)
Endpoint
Professional fulfilment and work exhaustion/interpersonal disengagement (Stanford PFI; co-primary)
Relevance
2Important — one of several pillars.
ResultWork exhaustion/interpersonal disengagement −0.44 points (95% CI −0.62 to −0.25; P<0.001); professional fulfilment +0.14 points (95% CI 0.004–0.28; P=0.04), reported as non-significant. Time on notes −0.36 h/day (95% CI −0.55 to −0.17); 38% of 71,487 notes drafted with ambient AI.
Afshar M, et al. NEJM AI. 2025;2(12). 10.1056/aioa2500945
Discussion & critique

A randomised implementation showing that ambient AI eases the exhaustion half of burnout and shortens documentation time without degrading note quality, billing codes or diagnosis capture — the safety question that matters to health systems. Small, unblinded and survey-based for its primary outcomes; the professional-fulfilment co-primary was null, the work-outside-work saving vanished once extreme days were trimmed, and 24 weeks is short for a well-being endpoint.