Cross-Silo Federated Learning for Predicting Successful Mechanical Ventilation Weaning: A Study Across Five ICU Databases (Preprint)

BACKGROUND The prediction of weaning from mechanical ventilation (MV) can support clinical decision-making and help reduce the risk of weaning failure in intensive care units (ICUs). Cross-silo federated learning (FL) offers a promising approach to developing robust predictive models across multiple institutions without requiring the sharing of patient-level data. OBJECTIVE This study aimed to evaluate the feasibility and efficacy of FL for predicting successful weaning from MV across 5 diverse ICU databases and to compare its performance with local learning (LL) and centralized learning (CL) approaches that differ in their data-sharing requirements. METHODS We conducted a retrospective analysis using 5 ICU databases, namely the eICU Collaborative Research Database (eICU-CRD), Medical Information Mart for Intensive Care IV (MIMIC-IV), Universitätsklinikum Augsburg (UKA), High-Resolution ICU Dataset (HiRID), and Amsterdam University Medical Centers (AUMC), transforming clinical variables into the Observational Medical Outcomes Partnership (OMOP) Common Data Model. We defined successful weaning as a sustained reduction in positive end-expiratory pressure. We compared 3 learning approaches, FL, LL, and CL, using extreme gradient boosting (XGBoost). Performance was evaluated using the area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), precision, recall, and F1-score. All data use complied with local ethical regulations and institutional review board approvals, and all databases contained deidentified patient data accessed under institutional data use agreements and governed by applicable privacy regulations. RESULTS A total of 24,521 patients were included across 5 databases. The CL model achieved an AUROC of 0.81, AUPRC of 0.57, and F1-score of 0.54 on pooled test data. The FL model achieved a macroaveraged AUROC of 0.74, AUPRC of 0.56, and F1-score of 0.52. LL model performance varied across databases (AUROC=0.68-0.84, AUPRC=0.51-0.68, F1-score=0.49-0.67), reflecting differences in data distribution and class balance. CONCLUSIONS Our findings highlight performance differences between learning approaches for MV weaning prediction. LL models achieved the highest performance within their respective institutions (AUROC=0.68-0.84), CL achieved the highest performance on pooled test data (AUROC=0.81), and FL showed lower but reasonable performance while avoiding direct sharing of patient-level data (AUROC=0.74). The choice between approaches depends on institutional data-sharing constraints, local dataset characteristics, and acceptable performance thresholds. As privacy was not formally measured, these findings should be read as a performance comparison rather than a privacy-performance trade-off.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.