{"id":"W4413052408","doi":"10.3233/shti251231","title":"An Iterative Composite Method for the De-Identification of Medical Notes","year":2025,"lang":"en","type":"article","venue":"Studies in health technology and informatics","topic":"Handwritten Text Recognition Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Princess Margaret Cancer Centre; University Health Network","funders":"","keywords":"Identification (biology); Computer science; Identifier; Matching (statistics); Security token; Medical record; Iterative method; Artificial intelligence; Residual; Machine learning; Data mining; Algorithm; Medicine; Statistics; Computer security; Mathematics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00375011,0.0009036496,0.001247521,0.005145166,0.0007799984,0.001586171,0.0016426,0.0008641383,0.003061072],"category_scores_gemma":[0.01627636,0.0003810762,0.001177423,0.0046442,0.0005756787,0.001854007,0.001633876,0.001222064,0.00350631],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005157531,"about_ca_system_score_gemma":0.003926617,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002810673,"about_ca_topic_score_gemma":0.00725326,"domain_scores_codex":[0.9956679,0.0009236557,0.0005066214,0.001124187,0.001617535,0.0001601323],"domain_scores_gemma":[0.9894433,0.004553175,0.001231395,0.001140546,0.003437227,0.0001943561],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0004205062,0.0001346795,0.003639027,0.0005211768,0.0001808035,0.0001661718,0.0003372977,0.004521185,0.02731659,0.00296819,0.00598484,0.9538094],"study_design_scores_gemma":[0.0002160334,0.001642515,0.02418617,0.0003748215,0.0007959133,0.003997829,0.001056154,0.6326628,0.183589,0.01759765,0.133403,0.0004782026],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01639326,0.001412581,0.9761321,0.000179931,0.0001429073,0.0004084315,0.0007867087,0.002725505,0.00181854],"genre_scores_gemma":[0.07540713,0.0005821141,0.9164219,0.0001321113,0.00009539098,0.0003839699,0.001933268,0.0002380365,0.004806226],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005145166,"threshold_uncertainty_score":0.01983273,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0406133108800299,"score_gpt":0.4391572423483992,"score_spread":0.3985439314683693,"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."}}