LLMs Show Promise in Extracting Sex-Specific Blood Pressure Data from Scientific Literature

A new study from arXiv (ID: 2402.01826) explores the use of large language models (LLMs) to automate the extraction of blood pressure (BP) information from scientific literature, with a focus on biological sex differences. The research addresses the longstanding issue that current BP standards, established decades ago, do not account for demographic factors such as sex and age, potentially leading to inaccurate diagnoses. The study employs natural language processing (NLP) methods to extract mean and standard deviation values of BP from articles, distinguishing by biological sex. A Solr-based search engine was developed to retrieve relevant articles containing BP-related keywords and sex indicators. The findings suggest that LLMs can effectively automate this process, which could lead to more personalized and accurate BP assessments. The study was announced as a replace-cross update on arXiv, indicating a revised version. The research is significant for the medical community, as it could improve the reliability of BP diagnostics by incorporating demographic considerations.