{"id":"W4412938129","doi":"10.1186/s44247-025-00166-4","title":"Post-processing methods for mitigating algorithmic bias in healthcare classification models: An extended umbrella review","year":2025,"lang":"en","type":"article","venue":"BMC Digital Health","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"York University; Schmidt Futures","keywords":"Health care; Computer science; Data science; Risk analysis (engineering); Medicine; Political science","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002117282,0.0001995772,0.0005488405,0.0002671825,0.0002331355,0.00008241426,0.0001299785,0.0001374327,0.000003205611],"category_scores_gemma":[0.001323906,0.0001942334,0.000108636,0.0007023146,0.00005585198,0.0005658423,0.00002187152,0.0002830356,0.00000756148],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006484753,"about_ca_system_score_gemma":0.004144466,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001409418,"about_ca_topic_score_gemma":0.0004831887,"domain_scores_codex":[0.9971495,0.0002649968,0.001259829,0.0005559223,0.0001821955,0.0005876186],"domain_scores_gemma":[0.9977849,0.0004995322,0.0003250415,0.0004043638,0.0006418204,0.0003443427],"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":[0.00008315091,0.0001971383,0.001652687,0.01568701,0.000003544801,3.629557e-7,0.0005550216,0.000009265876,0.00002150776,0.002882026,0.0002050523,0.9787033],"study_design_scores_gemma":[0.0008793394,0.003266601,0.01983031,0.06210731,0.0001135512,0.00009335512,0.0209377,0.7217104,0.001076913,0.1601587,0.008755045,0.001070812],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04145015,0.1600752,0.6086398,0.1741059,0.001444882,0.01107743,0.0001220329,0.0007123883,0.002372227],"genre_scores_gemma":[0.8038464,0.00559617,0.1600375,0.02847087,0.0003033918,0.0006080858,0.0006240149,0.00007793961,0.0004355981],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9776324,"threshold_uncertainty_score":0.792061,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4986411230848035,"score_gpt":0.5740969237019772,"score_spread":0.07545580061717372,"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."}}