A detailed review of large language models in the healthcare sector

Abstract Large Language Models (LLMs) are advanced systems designed to process and generate human-like language by leveraging vast data. LLMs are increasingly used in healthcare, performing tasks such as question-answer generation, medical image analysis, and translation. They have significantly transformed healthcare through clinical decision support, patient care improvement, simplified administrative tasks such as documentation, medical billing, data provision and security, and medical research. They are also instrumental, but not limited to, radiology for abnormality detection and pharmacogenomics for reducing adverse drug effects. However, despite LLMs’ broad adoption, they face challenges related to data privacy, ethical usage, and interpretability, highlighting the need for comprehensive guidelines to ensure their optimal use in the healthcare sector. Therefore, this review’s objective is to investigate the potential applications, advantages, and barriers of LLMs in the healthcare sector. In this study, a systematic review was conducted on 39 studies from PubMed (PM), Cochrane Library (Cc), Science Direct (SD), Web of Science (WOS), and Scopus database between 2019 and 2024. The result shows that Large Language Models (LLMs) are widely used in healthcare, including medical education, cardiology, mental health, emergency treatment, radiography, pharmacogenomics, and patient interactions. However, challenges like data confidentiality, interpretability, and ethical approvals necessitate strict implementation criteria. In conclusion, LLMs have demonstrated significant potential to improve diagnostic accuracy, patient-clinician communication, and administrative efficiency, ultimately leading to better healthcare outcomes as well as patient satisfaction.