Privacy-Preserving and Explainable Federated Learning for Multi-Hospital Clinical Decision Support: A Narrative Review and Proposed Framework
Abstract
Background: Artificial intelligence holds considerable promise for clinical decision support, yet adoption across hospitals remains constrained by two persistent obstacles — the legal and ethical barriers to sharing patient data, and the limited interpretability of many high-performing models. Federated learning (FL) has emerged as a means of training models across institutions without centralizing raw data, while explainable AI (XAI) offers a partial remedy to the "black box" problem that complicates clinical trust. Methods: This narrative review synthesizes peer-reviewed literature, predominantly published between 2018 and 2023, addressing the intersection of FL, privacy-preserving computation (differential privacy, secure aggregation, homomorphic encryption), explainability methods (SHAP, LIME, attention mechanisms), and blockchain-based auditability in healthcare contexts. Sources were identified through targeted database searches and citation tracking rather than a fully systematic protocol, consistent with the narrative review format. Results: The reviewed literature suggests that FL can achieve competitive predictive performance while keeping patient data institution-bound, that privacy-preserving mechanisms introduce measurable but often manageable accuracy trade-offs, and that XAI methods meaningfully improve clinician engagement with model outputs. Comparatively few studies, however, integrate all three elements — privacy, explainability, and blockchain-based…
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.