Validation of a machine learning algorithm for early severe sepsis prediction: a retrospective study predicting severe sepsis up to 48 h in advance using a diverse dataset from 461 US hospitals
Source check: Source checked. Not independent scientific confirmation.
Finding: A machine-learning algorithm predicted severe-sepsis onset from retrospective records across 461 health centers, reporting AUROC 0.827 at 48 hours before onset and external-dataset AUROC 0.752 at that horizon. This is retrospective validation; prospective clinical impact was not tested.
Technology: A machine-learning algorithm predicts severe-sepsis onset from electronic health-record variables up to 48 hours ahead.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- A machine-learning algorithm predicts severe-sepsis onset from electronic health-record variables up to 48 hours ahead.
- Test and result
- A 461-hospital retrospective cohort evaluates early machine-learning prediction of severe sepsis. Interpretation is limited to the source's cohort, data or test setting.
- Limitation / still unknown
- The paper reports the relevant cohort/image/field dataset, comparator or reference, and validation/test evaluation; the judgment does not imply prospective clinical benefit.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC7590695/ · https://pubmed.ncbi.nlm.nih.gov/33109167/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
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 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.