Applications of Large Language Models (LLMs) in Healthcare

Abstract

In recent years, large language models (LLMs) have revolutionized language-related applications across various fields. These powerful models, trained on massive datasets, can understand and generate natural language for tasks like summarization, questionanswering, and information extraction. One notable application of LLMs is their integration into the healthcare sector. Although numerous clinical applications have been proposed, from extracting medication information to providing patient support through chatbots, widespread implementation is still in progress. This review aimed to identify clinical tasks that make use of LLMs and can be applied within the healthcare sector, from relevant literature. From the 1008 founded publications, a random subset was included in this review. After thorough screening, 129 clinical tasks were described within the resulting 89 publications. These clinical tasks were categorized into the overarching tasks: ‘Clinical workflow’, ‘Patient education and communication’ or ‘Healthcare management’. The categorization of these clinical tasks, as well as the identification of the underlying classical NLP tasks, aimed to provide a comprehensive understanding on the described clinical tasks and the potential capabilities of utilizing LLMs in healthcare. Although many utilities of LLMs in healthcare were described, the majority was not yet implemented within clinical settings. This indicates that the future holds promise for the widespread implementation of these clinical tasks, but further development and validation are essential for realizing their full potential in transforming healthcare services.