{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.1938365,0.001483239,0.001099083,0.01216882,0.002459178,0.005567731,0.003912731,0.001704484,0.002978018],"category_scores_gemma":[0.3866277,0.0008987366,0.002032989,0.008889225,0.002303303,0.002994254,0.006752537,0.002122661,0.001871705],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.006763326,"about_ca_system_score_gemma":0.01994934,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03025748,"about_ca_topic_score_gemma":0.02980585,"domain_scores_codex":[0.8454638,0.06919545,0.02911188,0.01861777,0.03452352,0.003087572],"domain_scores_gemma":[0.5563561,0.2032307,0.03964838,0.05115841,0.1454477,0.004158807],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001188194,0.001143978,0.586468,0.002590029,0.0009039912,0.0003118734,0.01112599,0.008389332,0.005274828,0.006928151,0.04184724,0.3338284],"study_design_scores_gemma":[0.001027554,0.0009737973,0.6831223,0.005747555,0.0008629295,0.001034104,0.007627825,0.1342223,0.03098635,0.007355874,0.1265316,0.0005077499],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4489933,0.001717873,0.4435047,0.003439059,0.00112896,0.03368318,0.03641821,0.01014703,0.02096766],"genre_scores_gemma":[0.3775105,0.0003393621,0.5710793,0.00092478,0.0001597875,0.01530925,0.03197243,0.001010168,0.001694519],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1938365,"threshold_uncertainty_score":0.9941432,"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."}}