Investigating the Applications of Large Language Models (LLMs) in Extracting Information from Medical Records, Summarizing Patient Records, and Providing Support in Clinical Decision Making

Background: The application of LLMs in medicine has great potential to improve clinical processes and patient care. However, the development of these technologies must be done with due regard to ethical and security challenges to protect patient privacy. Optimal use of these models can lead to increased physician productivity, improved quality of care, and increased patient satisfaction. Materials and Methods: This article is a systematic review study that was conducted by searching for the keywords LLMs, artificial intelligence, Information Mining, Clinical Decision Making, Medicine in English. In order to collect data, articles published in reputable databases such as PubMed, Scopus and sciencedirect were used. Finally, ۱۰ articles related to the topic and with a time limit of five years were selected. Results: The results show that LLMs have a significant ability to extract key information from medical records, accurately summarize patient records, and provide support for clinical decision-making. These models can Information extraction from medical records: LLMs are able to identify and extract key information from medical texts with high accuracy. For example, in one study, LLMs were able to extract information related to medications, symptoms, and medical history of patients with an accuracy of more than ۸۵%. This information extraction can help students and physicians perform rapid and effective clinical analyses, especially in cases where the data is unstructured. Patient record summarization: LLMs were able to reduce physicians' reading time by ۳۰-۵۰% by producing useful and accurate summaries of medical records. The results showed that the summaries produced were so useful that more than ۷۰% of physicians used them for clinical decision-making. This helps improve the workflow of physicians and increases the quality of patient care. Clinical decision support: LLMs help provide treatment options and predict outcomes based on available medical data. In one case study, LLMS narrowed the number of treatment options based on a patient's past medical history to ۵-۷, which facilitated clinical decision-making and increased its accuracy. LLMs were also used to analyze clinical data and provide evidence-based recommendations to physicians, and the results showed that this increased physicians' confidence in their decisions. Conclusion: The application of LLMs in medicine has great potential to improve clinical processes and patient care. However, the development of these technologies must be done with due regard to ethical and security challenges to protect patient privacy. Optimal use of these models can lead to increased physician productivity, improved quality of care, and increased patient satisfaction.

Keywords:

Large Language Models (LLMs) , Information Mining , Clinical Decision Making , Artificial Intelligence , Medicine