The abstract is read at the publisher; the summary is JClub's.
In 6,366 stroke ICU stays in MIMIC IV, an hourly simulator matched a gradient-boosting model for predicting 24-hour alive ICU exit (Brier 0.1238 vs 0.1245).
- Why it matters: ICU time after stroke is partly unexplained, and it is unclear whether modeling patient condition, transfer readiness, and ICU exit separately can predict outcomes as well as direct approaches.
- What they did: They built an hourly simulator using stays up to 2016, then validated it on 2017–2019 (n=1,714) and 2020 onward (n=1,521), comparing 24-hour alive exit Brier scores to gradient boosting.
- The result: Unexplained hours were 34.1% in standard ICUs and 87.6% in neuro intermediate units; post-2016 errors mainly occurred after readiness, with calibration better at 48–72 hours but lower discrimination.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.