{"id":"W4414396056","doi":"10.1093/ajeadv/uuaf010","title":"ICD-10 code embedding for patient matching: a first step toward applications in causal inference from hospital discharge data","year":2025,"lang":"en","type":"article","venue":"AJE Advances Research in Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"University of Bern","keywords":"Propensity score matching; Causal inference; Matching (statistics); Set (abstract data type); Inference; Observational study; Population; Pooling; Code (set theory)","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.009293783,0.0001640302,0.0006226986,0.0004291576,0.0007579833,0.000007638815,0.000836557,0.0003331539,0.0001370545],"category_scores_gemma":[0.02226752,0.0001373994,0.00003220555,0.0006138931,0.0002116557,0.0004653879,0.0008258173,0.001768063,0.00009927426],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004498854,"about_ca_system_score_gemma":0.0007248083,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003780035,"about_ca_topic_score_gemma":0.009203741,"domain_scores_codex":[0.9941403,0.001942191,0.001663159,0.0006474545,0.0002591097,0.001347803],"domain_scores_gemma":[0.9554317,0.04291419,0.0003284188,0.0008491666,0.0002295546,0.0002470026],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004172788,0.0003314935,0.8013672,0.003108213,0.00002888456,0.000003302477,0.008701796,0.00131388,0.000006989252,0.09081522,0.03380957,0.06009616],"study_design_scores_gemma":[0.001481729,0.00026983,0.04445863,0.001554773,0.000005944095,1.208298e-7,0.008715744,0.08478221,0.000002398844,0.1235262,0.7349771,0.0002253066],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2018753,0.008348104,0.6547312,0.1094202,0.002304094,0.01286612,0.002741574,0.0002217406,0.007491667],"genre_scores_gemma":[0.9600184,0.003854652,0.02425905,0.002453034,0.0003166455,0.006390412,0.002367516,0.00002254898,0.0003177164],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7581431,"threshold_uncertainty_score":0.9859684,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4973699652473476,"score_gpt":0.628304726194639,"score_spread":0.1309347609472914,"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."}}