<p>This review delves into the evolving landscape of privacy-preserving federated learning, with a spotlight on its pivotal role and applications in healthcare. Federated learning, a decentralised machine learning approach, has emerged as a cornerstone for enabling model training across multiple devices or servers while maintaining data privacy. In the healthcare sector, preserving the confidentiality of sensitive patient data is paramount, necessitating innovative privacy-preserving techniques such as homomorphic encoding, differential privacy, and secure multi-party computation. The paper explores the integration of these techniques with disease prediction models, highlighting their potential to enhance remote patient monitoring, clinical decision support, and personalised medicine. Despite the promising advancements, the review identifies key challenges, including scalability, communication overhead, and ethical considerations, that need addressing to foster wider adoption. The paper concludes by projecting future directions, emphasising the continual development of privacy-preserving methods and their integration with emerging technologies to expand the applications of federated learning in healthcare. Through a comprehensive exploration, this review aims to shed light on the advancements, applications, and challenges, offering insights and recommendations for harnessing the potential of privacy-preserving federated learning in healthcare.</p>