{"id":"W4400771582","doi":"10.2196/56243","title":"Extraction of Substance Use Information From Clinical Notes: Generative Pretrained Transformer–Based Investigation","year":2024,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Mental Health via Writing","field":"Psychology","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Center for Advancing Translational Sciences; National Institute on Drug Abuse; National Heart, Lung, and Blood Institute","keywords":"Computer science; Preprint; Health care; Set (abstract data type); Informatics; Identification (biology); Health informatics; Information extraction; Artificial intelligence; Data science; Machine learning; Public health; Medicine","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001266842,0.0001858757,0.0003186825,0.000195081,0.00007103448,0.00007665142,0.0001776491,0.0004915062,0.001247137],"category_scores_gemma":[0.0003303185,0.0001653612,0.00012083,0.0004111437,0.0002519661,0.001985313,0.000007071075,0.0007559438,0.000325049],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001121614,"about_ca_system_score_gemma":0.0003156338,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001200903,"about_ca_topic_score_gemma":0.00005628464,"domain_scores_codex":[0.9962332,0.0001979769,0.002284979,0.000130094,0.0008535728,0.0003002193],"domain_scores_gemma":[0.9970546,0.001811494,0.0004016671,0.0002600343,0.0001252707,0.0003469648],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0008066687,0.0005283044,0.01635784,0.003642922,0.0003599036,0.00003886969,0.2535675,0.00006478058,0.0004827818,0.01638554,0.01847311,0.6892918],"study_design_scores_gemma":[0.008904088,0.001747808,0.1006648,0.005918804,0.000229416,0.00006517741,0.02309134,0.8049327,0.007117392,0.003076183,0.04292667,0.001325618],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8671494,0.00008087453,0.1278125,0.0006701351,0.001441766,0.0007594291,0.0001657967,0.0001864502,0.001733688],"genre_scores_gemma":[0.980858,0.00004538968,0.01442411,0.003375835,0.0003095035,0.0001599445,0.0007777608,0.00001856844,0.00003088004],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8048679,"threshold_uncertainty_score":0.9996659,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.108892163245482,"score_gpt":0.4401655376174755,"score_spread":0.3312733743719936,"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."}}