{"id":"W3157701965","doi":"10.1101/2021.04.22.21255659","title":"Predicting a diagnosis of ankylosing spondylitis using primary care health records – a machine learning approach","year":2021,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Spondyloarthritis Studies and Treatments","field":"Medicine","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute of Infection and Immunity","funders":"Engineering and Physical Sciences Research Council; Economic and Social Research Council; Chief Scientist Office, Scottish Government Health and Social Care Directorate; Health and Social Care Research and Development Division; Medical Research Council; Public Health Agency; Scottish Government; Department of Health and Social Care; UCB Pharma; British Heart Foundation; Wellcome Trust","keywords":"Ankylosing spondylitis; Medicine; Spondylitis; Test (biology); Population; Decision tree; Medical record; Artificial intelligence; Back pain; Physical therapy; Machine learning; Feature selection; Internal medicine; Computer science; Alternative medicine; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003349881,0.0004170923,0.0007837442,0.003172986,0.00040049,0.001239122,0.0005935577,0.001032429,0.001822413],"category_scores_gemma":[0.01098842,0.0003060586,0.0009142376,0.001707572,0.0002284816,0.0006583724,0.0004799494,0.001204325,0.0005596813],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006634911,"about_ca_system_score_gemma":0.0006599843,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009363282,"about_ca_topic_score_gemma":0.006494361,"domain_scores_codex":[0.9986904,0.0006986562,0.0001158649,0.0002207642,0.0001589347,0.0001154092],"domain_scores_gemma":[0.9896054,0.008272982,0.0009099693,0.0003202868,0.0006159139,0.0002754338],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000250257,0.0004886536,0.9563567,0.00006034599,0.0002633747,0.00015442,0.00008782685,0.01852771,0.000263875,0.0001353005,0.0009266947,0.02248496],"study_design_scores_gemma":[0.0000432513,0.0004639225,0.4429173,0.00009637945,0.0001644691,0.0002799993,0.0003560201,0.5531115,0.0004529597,0.001488502,0.0005962587,0.00002945042],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9801682,0.0004185631,0.01351843,0.001221329,0.00003411932,0.0001385638,0.003275738,0.0001414911,0.001083508],"genre_scores_gemma":[0.9917759,0.00008826303,0.006433047,0.00006164756,0.00002865022,0.00003726494,0.001378853,0.00000452219,0.0001919542],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009363282,"threshold_uncertainty_score":0.01861757,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03847327851158796,"score_gpt":0.2944224288130068,"score_spread":0.2559491503014188,"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."}}