Randomised trial of AI-generated lay summaries before neurosurgical consultations
In a single-centre randomised trial of 300 outpatients in Dresden, giving patients a quality-checked AI-generated lay summary of their medical findings before consultation improved patient-rated quality of the physician-patient interaction (4.42 against 4.25 on the FAPI scale) and shortened median consultations from 33 to 25 minutes.
- Model
- Simply Onno (underlying model not named in the abstract)
- Field
- Medicine
- Date
- 2026-10-15
- Human collaborators
- Elida Benker-Hasani, Witold H. Polanski
What was found
Adults attending an outpatient neurosurgical clinic were randomised 1:1 to standard care with or without a layperson summary of their findings, produced by the Simply Onno tool and quality-checked before use. The primary endpoint, patient-reported interaction quality, improved by 0.17 (95% CI 0.05 to 0.29), driven by shared decision-making, with no difference in perceived empathy. Patients rated autonomy support higher while physicians rated their own slightly lower, and the ratio of geometric mean consultation times was 0.79 (0.70 to 0.88). Evidence for an increase in encounter difficulty was weak, and secondary intervals were not adjusted for multiplicity.
Novelty check
Europe PMC, 2026-10-29: TITLE_ABS:"summaries" AND TITLE_ABS:"randomized" AND TITLE_ABS:"consultation" returned 20 records; the related ones are a 2025 JMIR protocol on layperson-friendly AI translation of medical documents and a 2026 trial of computer-assisted pre-visit history taking. No earlier completed randomised trial of AI-generated lay summaries before consultations was found.
Caveats and known objections
Single centre, the neurosurgery department of University Hospital Carl Gustav Carus in Dresden; the abstract names the tool but not its underlying model or whether its maker had a role, and states no funding. Randomised, but the endpoints are patient-reported scores and consultation time rather than clinical outcomes, and the authors say effects on outcomes and workload need confirmation. The effect on the primary scale is small (Hedges' g 0.33). Autonomy is ai-assisted: the AI drafted summaries that were quality-checked before patients saw them.
No independent check is recorded on this entry 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. (2026). Randomised trial of AI-generated lay summaries before neurosurgical consultations. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-10-15-ai-infocare-previsit-summaries-rctBibTeX
@misc{whataifound-tudresden-2026-rct,
title = {Randomised trial of AI-generated lay summaries before neurosurgical consultations},
author = {{whataifound.org}},
year = {2026},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by TU Dresden (University Hospital Carl Gustav Carus). Verification: Peer reviewed. Autonomy: AI-assisted.},
url = {https://whataifound.org/finding/2026-10-15-ai-infocare-previsit-summaries-rct}
}