Randomised trial of two ambient AI scribes in outpatient practice
In a pragmatic three-arm trial of 238 UCLA outpatient physicians, the Nabla scribe cut time spent writing notes by 9.5% against usual care while Microsoft's DAX Copilot made no significant difference; both were associated with modest improvements in burnout and task-load scores, and physicians reported occasional clinically significant inaccuracies with either.
- Lab
- UCLA Health
- Model
- Microsoft Dragon Ambient eXperience (DAX) Copilot; Nabla
- Field
- Medicine
- Date
- 2025-11-26
- Human collaborators
- Paul J. Lukac, John N. Mafi
What was found
Ambient scribes record the visit and draft the clinical note, which the physician then edits. Physicians from 14 specialties were assigned 1:1:1 by covariate-constrained randomisation to DAX, Nabla or usual care from November 2024 to January 2025. The primary outcome was the change from baseline in log time-in-note, measured from record audit logs. DAX was used in 33.5% of 24,696 visits and Nabla in 29.5% of 23,653. Secondary survey outcomes, the Mini-Z burnout score, physician task load and work exhaustion, improved for users of either scribe against control. One mild adverse event was reported. A preprint appeared on medRxiv in July 2025.
Novelty check
The authors state that no randomised trial of ambient scribes preceded theirs. Europe PMC, (ambient OR scribe) in the title with a randomised design, to 26 November 2025: only this trial and its July 2025 preprint; observational evaluations at large health systems came earlier, and a randomised crossover trial at Duke (JAMIA, 2026) followed.
Caveats and known objections
Two months at one health system, with the scribes used in only about a third of eligible visits, so the intention-to-treat effects are small partly because uptake was. The burnout findings are secondary survey endpoints that the authors say need confirmation in larger multicenter trials, and physicians knew their allocation. Time-in-note does not capture time spent checking drafts outside the note. Autonomy is ai-assisted: the scribe drafts and the physician edits and signs.
Trial registration
That a study was registered, with the outcomes it declared and the date it declared them, so a published endpoint can be checked against the one planned. It does not certify the results, their analysis, or any clinical benefit; registry records are submitted by the sponsor and are not peer reviewed.
Nobody outside the lab has checked this yet.
Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.
Or on GitHub: submit a check challenge the grade send a correction or send a pull request
Flag this for triage
Signals order the review queue and nothing else. They are never published, and they never move a grade: that takes a citation.
Entry history (1 event)
- AddedEntered the registry graded Peer reviewed and AI-assisted.
Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.
Graded peer reviewed for verification and ai-assisted for autonomy. What these mean.
Cite this entry
whataifound.org. (2025). Randomised trial of two ambient AI scribes in outpatient practice. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2025-11-26-ambient-ai-scribes-rctBibTeX
@misc{whataifound-uclahealth-2025-rct,
title = {Randomised trial of two ambient AI scribes in outpatient practice},
author = {{whataifound.org}},
year = {2025},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by UCLA Health. Verification: Peer reviewed. Autonomy: AI-assisted.},
url = {https://whataifound.org/finding/2025-11-26-ambient-ai-scribes-rct}
}