{"id":"W4390745503","doi":"10.1038/s41746-023-00970-0","title":"Large language models to identify social determinants of health in electronic health records","year":2024,"lang":"en","type":"article","venue":"npj Digital Medicine","topic":"Chronic Disease Management Strategies","field":"Medicine","cited_by":254,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"U.S. National Library of Medicine; European Commission; ViewRay; NRG Oncology; Radiological Society of North America; National Cancer Institute; National Institutes of Health; Conquer Cancer Foundation; U.S. Department of Health and Human Services","keywords":"Social determinants of health; Health care; Medicine; Political science; Law","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007269428,0.0001976124,0.00070739,0.0004507752,0.00004073624,0.00003498684,0.0001316784,0.00004707681,0.0001523899],"category_scores_gemma":[0.0000763886,0.0001666103,0.00008403172,0.0006864818,0.00007092037,0.000286849,0.00008778335,0.0002273291,0.00003859103],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006079492,"about_ca_system_score_gemma":0.0008723428,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004553765,"about_ca_topic_score_gemma":0.000558645,"domain_scores_codex":[0.9977083,0.00004192713,0.0007099087,0.0003865807,0.0004429197,0.0007103635],"domain_scores_gemma":[0.9993598,0.00004934932,0.0001058092,0.0002480765,0.00003141762,0.000205577],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.0004762846,0.001045756,0.003819881,0.00946483,0.0002616765,0.0005782936,0.03300972,0.00002197065,0.0001557825,0.03746618,0.1230995,0.7906001],"study_design_scores_gemma":[0.04229123,0.03132416,0.3685255,0.05843661,0.000732964,0.0003536795,0.1480623,0.0274218,0.0002614264,0.09248691,0.2267371,0.003366361],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9157853,0.01628208,0.001977728,0.03495791,0.0006246662,0.001705146,0.0001822832,0.0003007805,0.02818414],"genre_scores_gemma":[0.9961209,0.0002866632,0.00001668636,0.001770757,0.000288714,0.00002628238,0.0001435591,0.00003976881,0.001306724],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7872337,"threshold_uncertainty_score":0.679417,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04766063648794992,"score_gpt":0.4148725473884519,"score_spread":0.3672119109005019,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}