A groundbreaking study from researchers at Mass General Brigham reveals that large language models can help identify social determinants of health from electronic health records (EHRs), potentially improving real-world evidence and addressing health disparities. The research, published in npj Digital Medicine, highlights the underdocumentation of social determinants in structured EHR data, which hinders comprehensive research and clinical care.
The study explores the use of natural language processing to automate the extraction of social determinant information from clinical texts, focusing on six categories: employment, housing, transportation, parental status, relationship, and social support. By addressing the challenge of incorporating these critical factors into research databases, the researchers aim to create a more holistic understanding of the impact of social determinants on health outcomes.
This innovative approach could help alleviate the bottleneck in understanding health disparities, particularly among underrepresented populations, and ultimately contribute to more equitable healthcare.
See “AI can help identify social determinants of health” by Todd Shryock on the Medical Economics website (January 15, 2024)
