Federated learning (FL) has emerged as a promising solution for privacy-preserving machine learning in sensitive domains like healthcare. By enabling collaborative model training without sharing raw data, FL holds the potential to unlock high-quality and generalizable AI models across different institutions. Despite significant academic interest, the real-world deployment of FL in healthcare remains limited due to persistent challenges, including strict privacy requirements, robustness to failures and adversarial attacks, regulatory compliance, and infrastructural constraints. These barriers are particularly critical in medical contexts, where errors can have life-threatening consequences and trust in AI systems must be exceptionally high.

This Methods Collection aims to advance the development of practical and reliable FL methods tailored to the healthcare domain. It invites contributions that address the full spectrum of challenges in deploying FL for medical applications, including privacy-preserving algorithms, robustness against malicious clients, handling heterogeneous data distributions, compliance with data protection regulations, and fault-tolerant system designs. By focusing on methods bridging the gap between research prototypes and production-ready healthcare systems, this collection will serve as a valuable resource for researchers and practitioners.

This Methods Collection will help accelerate the development of federated learning systems that are technically sound and deployable in high-stakes medical environments, ultimately contributing to safer, fairer, and more effective AI-driven healthcare solutions.