LO
PilotCognitive automation2024-03-04
At Stanford, physicians, advanced practice practitioners, nurses and pharmacists used AI-drafted replies for 20% of patient messages, with no change in reply time
Nurse practitioneroccupation page →
Event date / reported
2024-03-04
Evidence stage
PilotSmall-scale trial in a real setting. Tells us the deployment conditions are being tested, not that they hold — so one pilot is never enough on its own; two independent ones are.
Tasks this bears on
Patient messages and education
Answering patients' portal messages and explaining conditions and treatment.
Being augmented✓ Evidence-backed
Where this applies
Stanford Health Care, United States. A five-week quality improvement study in 2023 enrolled all attending physicians, advanced practice practitioners, clinic nurses and clinical pharmacists in primary care and gastroenterology; 162 clinicians were analysed. A language model integrated in the health record drafted replies to patient portal messages. The mean draft utilisation rate was 20%; reply, write and read times did not change, while reported task load and work exhaustion fell. It does not report results for nurse practitioners separately.
What this means
Clinicians including advanced practice practitioners chose AI drafts for a fifth of patient replies; it eased reported burden but did not save time.
What it does not yet show
One medical centre for five weeks; nurse practitioners are not reported separately.
What you can check
Check whether your patient portal offers AI-drafted replies and how often you would keep one.
Does it change the assessment?
No. The impact index is never moved by a single event, and this stage does not move one on its own: a "Pilot" record counts toward a judgement but needs a second, independent record before the judgement rests on evidence. This one is counted; on its own it changed nothing.
Source
Garcia P, Ma SP, Shah S, et al. "Artificial Intelligence-Generated Draft Replies to Patient Inbox Messages." JAMA Network Open 2024;7(3) (published March 4, 2024; PMC10955355) · verified 2026-10-07 · Claude (VOLO agent) · interpreted 2026-10-07 · Claude (VOLO agent)
Primary source — published by the party that did this, or the authority of record. No co-signature needed.