{"id":"W4399120869","doi":"10.1016/j.ijmedinf.2024.105508","title":"Development and external validation of tools for categorizing diagnosis codes in international hospital data","year":2024,"lang":"en","type":"article","venue":"International Journal of Medical Informatics","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"Sinai Health System; St. Michael's Hospital; University of Toronto","funders":"","keywords":"Computer science; Diagnosis code; Health informatics; Data mining; Data science; Medicine; Public health; Nursing","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005083982,0.00009987297,0.0002511513,0.000414379,0.0000697933,0.00008331148,0.0008586759,0.0001878371,0.0002439005],"category_scores_gemma":[0.004161054,0.00007838676,0.00004060894,0.00009473599,0.00005738579,0.001581474,0.0002876378,0.0006350247,0.0000137762],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002216065,"about_ca_system_score_gemma":0.001103667,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002467544,"about_ca_topic_score_gemma":0.00001461365,"domain_scores_codex":[0.9956583,0.00005525054,0.002360114,0.00007411597,0.001660017,0.0001922282],"domain_scores_gemma":[0.9966446,0.001789558,0.0007389421,0.00010377,0.0005375001,0.0001856518],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002217928,0.0002060421,0.01909621,0.002349973,0.0003417039,0.00004444917,0.05476506,0.00008467414,0.00001733413,0.01988762,0.01525054,0.8877346],"study_design_scores_gemma":[0.005797967,0.0003398557,0.007090983,0.01866872,0.00007695833,0.0001328599,0.02542024,0.2543574,0.0008598947,0.00314701,0.6836945,0.0004136453],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7947835,0.0005346252,0.1803336,0.01340246,0.009401221,0.0004696074,0.0001438186,0.00002810409,0.000902986],"genre_scores_gemma":[0.9567475,0.002095215,0.03854936,0.001033247,0.001208154,0.00004546307,0.0002743093,0.00001250407,0.00003427224],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8873209,"threshold_uncertainty_score":0.4981471,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.318970735756868,"score_gpt":0.5105087465303292,"score_spread":0.1915380107734612,"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."}}