Bridging the gap: a practical step-by-step approach to warrant safe implementation of large language models in healthcare

Frontiers in Artificial Intelligence, Journal Year: 2025, Volume and Issue: 8

Published: Jan. 27, 2025

Large Language Models (LLMs) offer considerable potential to enhance various aspects of healthcare, from aiding with administrative tasks clinical decision support. However, despite the growing use LLMs in a critical gap persists clear, actionable guidelines available healthcare organizations and providers ensure their responsible safe implementation. In this paper, we propose practical step-by-step approach bridge support warranting implementation into healthcare. The recommendations manuscript include protecting patient privacy, adapting models healthcare-specific needs, adjusting hyperparameters appropriately, ensuring proper medical prompt engineering, distinguishing between (CDS) non-CDS applications, systematically evaluating LLM outputs using structured approach, implementing solid model governance structure. We furthermore ACUTE mnemonic; for assessing responses based on Accuracy, Consistency, semantically Unaltered outputs, Traceability, Ethical considerations. Together, these aim provide clear pathway practice.

Language: Английский

Artificial Intelligence and Patient Education: Examining the Accuracy and Reproducibility of Responses to Nutrition Questions Related to Inflammatory Bowel Disease by GPT-4

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Oct. 30, 2023

ABSTRACT Background and Aims Generative Pre-trained Transformer-4 (GPT-4) is a large language model (LLM) trained on vast corpus of data, including the medical literature. Nutrition plays an important role in managing inflammatory bowel disease (IBD), with unmet need for nutrition-related patient education resources. This study examines accuracy, comprehensiveness, reproducibility responses by GPT-4 to nutrition questions related IBD. Methods Questions were obtained from adult IBD clinic visits, Facebook, Reddit. Two IBD-focused registered dieticians independently graded accuracy GPT-4’s while third senior dietitian arbitrated. Each question was inputted twice into model. Results 88 selected. The correctly responded 73/88 (83.0%), 61 (69.0%) as comprehensive. 15/88 (17%) mixed correct incorrect/outdated data. comprehensively 10 (62.5%) “Nutrition diet needs surgery”, 12 (92.3%) “Tube feeding parenteral nutrition”, 11 (64.7%) “General questions”, (50%) “Diet reducing symptoms/inflammation” 18 (81.8%) “Micronutrients/supplementation needs”. provided reproducible 81/88 (92.0%) questions. Conclusion answered most questions, demonstrating promising potential LLMs supplementary tools patients seeking information. However, 17% contained incorrect information, highlighting continuous refinement prior incorporation clinical practice. Future studies should emphasize leveraging enhance outcomes promoting healthcare professional proficiency using maximize their efficacy. Lay Summary that With validation, there enhancing health literacy this population.

Language: Английский