Protocol Choices Matter More Than the Aggregation Algorithm: Federated Learning for In-Hospital Mortality Prediction in Care-Unit and Multi-Hospital Federations
This study compared seven federated learning strategies for predicting hospital mortality across two different data settings: one within a single hospital and one across multiple hospitals. The results show that the specific evaluation protocol used, such as how data is coded or which validation round is chosen, often has a larger impact on prediction accuracy than the choice of aggregation algorithm itself.