{"id":"W4391812863","doi":"10.1186/s12911-024-02449-8","title":"Characterizing the limitations of using diagnosis codes in the context of machine learning for healthcare","year":2024,"lang":"en","type":"article","venue":"BMC Medical Informatics and Decision Making","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":26,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute for Clinical Evaluative Sciences; SickKids Foundation; Hospital for Sick Children","funders":"","keywords":"Medicine; Health informatics; Context (archaeology); Confidence interval; Medical diagnosis; Odds ratio; Gold standard (test); Diagnostic odds ratio; Concordance; Kappa; Internal medicine; Pediatrics; Emergency medicine; Public health; Pathology","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":[0.006195967,0.00009149149,0.0002816565,0.0001761469,0.0005255683,0.00002545592,0.0001883412,0.0001608389,0.000030859],"category_scores_gemma":[0.005868805,0.0000477926,0.00005976277,0.0002979003,0.00009529796,0.0001665969,0.00008461945,0.0006761791,0.000003324235],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003612223,"about_ca_system_score_gemma":0.0004760324,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008863994,"about_ca_topic_score_gemma":0.0003533059,"domain_scores_codex":[0.997207,0.0002600663,0.001591246,0.00006607181,0.0006279718,0.0002476691],"domain_scores_gemma":[0.9688975,0.03032658,0.0004154407,0.000135957,0.0001417093,0.00008284092],"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":[0.0001111345,0.00002449417,0.01687612,0.006573994,0.00001294887,8.101552e-7,0.07705339,0.0001494032,0.000002634506,0.02592118,0.0006012758,0.8726726],"study_design_scores_gemma":[0.0004181131,0.00007348598,0.001485906,0.006325476,0.00001624117,0.000004496552,0.03913788,0.9380621,0.000002318834,0.001453862,0.0129648,0.00005534888],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5832487,0.002150949,0.4102868,0.002083276,0.0005739416,0.001171118,0.00004338434,0.00003042889,0.0004114625],"genre_scores_gemma":[0.9788032,0.002032763,0.01577784,0.003165131,0.00008793564,0.0001040909,0.00001866568,0.000007380824,0.000003009399],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9379126,"threshold_uncertainty_score":0.7025932,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4599963482276016,"score_gpt":0.5098633561727007,"score_spread":0.04986700794509913,"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."}}