A Comprehensive Survey on the Trustworthiness of Large Language Models in Healthcare

The application of large language models (LLMs) in healthcare has the potential to revolutionize clinical decision-making, medical research, and patient care. As LLMs are increasingly integrated into healthcare systems, several critical challenges must be addressed to ensure their reliable and ethical deployment. These challenges include truthfulness, where models generate misleading information; privacy, with risks of unintentional data retention; robustness, requiring defenses against adversarial attacks; fairness, addressing biases in clinical outcomes; explainability, ensuring transparent decision-making; and safety, mitigating risks of misinformation and medical errors. Recently, researchers have begun developing benchmarks and evaluation frameworks to systematically assess the trustworthiness of LLMs. However, the trustworthiness of LLMs in healthcare remains underexplored, lacking a systematic review that provides a comprehensive understanding and future insights into this area. This survey bridges this gap by providing a comprehensive overview of the recent research of existing methodologies and solutions aimed at mitigating the above risks in healthcare. By focusing on key trustworthiness dimensions including truthfulness, privacy and safety, robustness, fairness and bias, and explainability, we present a thorough analysis of how these issues impact the reliability and ethical use of LLMs in healthcare. This paper highlights ongoing efforts and offers insights into future research directions to ensure the safe and trustworthy deployment of LLMs in healthcare.