Reinforcement learning based automated anesthesia system for gastrointestinal endoscopy with a multicenter randomized trial.
Source check: Source checked. Not independent scientific confirmation.
Finding: In a multicenter randomized trial of 418 adults undergoing endoscopy, AAS-GE and clinician-managed anesthesia had similar observed hypoxemia rates (14.42% vs 14.29%; OR 1.01, 95% CI 0.59–1.75; P=.968). Induction was shorter, while body movement was more frequent with AAS-GE. The published protocol powered a test of a difference based on expected rates of 20% versus 10%; it did not specify a noninferiority margin, so the trial does not establish that a prespecified noninferiority criterion was met.
Technology: Offline RL learned the dose-control policy; the model was selected and refined before the randomized evaluation
Publication: Peer-reviewed paper · Test context: prospective or independent check. Catalog categories, not an evidence grade.
- AI contribution
- Offline RL learned the dose-control policy; the model was selected and refined before the randomized evaluation.
- Test and result
- The primary report gives hypoxemia rates of 14.42% vs 14.29% (OR 1.01, 95% CI 0.59–1.75; P=.968), shorter induction, and more body movement with AAS-GE. The published protocol's sample-size calculation tests a difference assuming expected rates of 20% vs 10%; no noninferiority margin is specified.
- Limitation / still unknown
- A prespecified noninferiority margin and corresponding noninferiority analysis were not identified in the published protocol or primary report; the observed nonsignificant difference does not itself establish noninferiority.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.nature.com/articles/s41746-026-02657-8 · https://ia.intanes.org/protocol/25001.html
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 4/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.