Answers are grounded in this paper and cite their sources.
Objective: To develop and internally validate a calibrated discharge model of 30-day within-system readmission in Medicare-insured adults, and to quantify how a future-information leak and outcome artifacts distorted performance, timing, and subgroup conclusions. Materials and Methods: Retrospective study of 236,906 Medicare-insured MIMIC-IV v3.1 admissions (87,750 patients; prevalence 21.4%). Predictors were audited for discharge-time availability; history features were tested for future information and recomputed from raw timestamps. A consensus gradient-boosted (XGBoost) model was selected inside patient-grouped folds; CIs used patient-cluster bootstrap. Results: A feature documented as a prior 180-day length-of-stay trend used each patient's whole record, including later admissions; removing it and two derived terms lowered test area under the receiver operating characteristic curve (AUROC) from 0.774 to 0.706. The corrected procedure reached an out-of-fold AUROC of 0.7057 (95% CI 0.7016 to 0.7102; primary); the 44-feature model scored 0.7056 on the historically exposed test partition (calibration slope 0.99), above reconstructed LACE and HOSPITAL (both P<.001). A day-1 discrimination peak largely reflected 2,368 probable administrative continuations within 1 hour of discharge; merging these into the preceding stay gave test AUROC 0.698 (0.702 after refitting). The White minus Black AUROC difference was 0.003 (95% CI -0.018 to 0.024); specificity was lower for Black patients (0.61 vs 0.69). Discussion: Documentation audits and patient-grouped validation missed a leak that inflated discrimination and created spurious subgroup and calendar-period differences; simple empirical checks found it. Conclusion: Empirical leakage checks should accompany availability audits. The corrected model is calibrated, moderately discriminating, and needs external validation before clinical use.
No takes yet. Share an insight, caveat, or question.
Bandeira et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: