AI Science Tracker
EN

Deep learning predicts mortality in a multicenter severe COVID-19 ICU cohort

Chonnam National University Hospital, Gwangju, Korea · 16 Nov 2026 · medicine

Discovery

Deep learning predicts mortality in a multicenter severe COVID-19 ICU cohort

Source check: Source checked. Not independent scientific confirmation.

Finding: A deep learning model was assessed in registry data for 900 patients treated in intensive care at 22 Korean centers during 2020–2021. A reduced seven-variable model had an AUC of 0.8247, compared with 0.8500 for the 66-variable model. This is a retrospective prediction result; the study does not establish clinical utility or temporal external validation.

Technology: Deep learning mortality prediction model with feature reduction

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
The learned deep tabular model estimated mortality risk and supported feature reduction.
Test and result
The model predicted mortality from registry data for 900 ICU patients across 22 centers; the seven-variable and 66-variable models had AUCs of 0.8247 and 0.8500, respectively.
Limitation / still unknown
The retrospective 2020–2021 analysis does not establish prospective clinical benefit or temporal external validation.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://jkms.org/DOIx.php?id=10.3346%2Fjkms.2026.41.e275

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.

Open in the full catalog for filters, charts and related discoveries. Browse medicine discoveries.