Prospective multi-site study of the TREWS sepsis early-warning system
Across five hospitals that monitored 590,736 patients with a machine-learning sepsis alert, patients with sepsis whose alert a clinician confirmed within three hours had 3.3 percentage points lower adjusted in-hospital mortality, an 18.7% relative reduction, than patients whose alert was not confirmed in that time.
- Model
- TREWS (Targeted Real-time Early Warning System)
- Field
- Medicine
- Date
- 2022-07-21
- Human collaborators
- Roy Adams, Katharine E. Henry, Suchi Saria
What was found
TREWS runs on electronic health record data and alerts the care team, who then confirm or dismiss it. The analysis focused on 6,877 patients with sepsis who were flagged before antibiotics had started. Adjusting for presentation and severity, those whose alert was confirmed within three hours also had less organ failure and shorter stays, and the mortality difference was larger, 4.5 percentage points, among patients the system additionally flagged as high risk. A companion paper in the same issue studied how clinicians adopted the alerts. In April 2026 the FDA cleared Bayesian Health's sepsis flagging device, part of the TREWS platform, by 510(k).
Novelty check
Europe PMC, titles (machine learning OR artificial intelligence) AND sepsis AND (randomised OR prospective), before 21 July 2022: three hits, including a small single-centre randomised trial of a machine-learning sepsis predictor in intensive care (Shimabukuro et al., BMJ Open Respiratory Research, 2017). This study's novelty is scale, five hospitals and more than half a million monitored patients, not being the first prospective evaluation.
Caveats and known objections
Not randomised. The comparison is between alerts clinicians confirmed quickly and alerts they did not, and the patients clinicians chose to act on promptly may differ in ways the adjustment does not capture, so the mortality difference is an association rather than a measured effect of deploying the system. The senior author, Suchi Saria, founded Bayesian Health, which commercialises TREWS. Europe PMC, TREWS in titles and abstracts: the published evaluations come from the developers' group, and this search found no validation by an unaffiliated health system, unlike Epic's sepsis model.
Regulatory authorization
FDA AI-enabled devicesBayesian Health Sepsis Flagging Device
K250680
That the FDA authorized the device for marketing in the US for a stated indication, on the evidence the manufacturer submitted. Most AI devices reach the market by 510(k), which requires substantial equivalence to an earlier device rather than new clinical evidence, so an authorization is not evidence of clinical benefit.
Cleared 30 April 2026 and announced by the company on 12 May 2026 as part of the TREWS platform; the FDA record itself does not name TREWS.
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. (2022). Prospective multi-site study of the TREWS sepsis early-warning system. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2022-07-21-trews-sepsis-early-warningBibTeX
@misc{whataifound-johnshopkinsuniversity-2022-warning,
title = {Prospective multi-site study of the TREWS sepsis early-warning system},
author = {{whataifound.org}},
year = {2022},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by Johns Hopkins University. Verification: Peer reviewed. Autonomy: AI-assisted.},
url = {https://whataifound.org/finding/2022-07-21-trews-sepsis-early-warning}
}