{"id":"W2050091824","doi":"10.1136/bmj.c4226","title":"Importance of accurately identifying disease in studies using electronic health records","year":2010,"lang":"en","type":"article","venue":"BMJ","topic":"Chronic Disease Management Strategies","field":"Medicine","cited_by":161,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute for Clinical Evaluative Sciences; Ottawa Hospital; University of Ottawa; University of Toronto; Statistics Canada","funders":"","keywords":"Health records; Electronic health record; Disease; Data science; Computer science; Data mining; Medicine; Health care; Political science; 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":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.4940192,0.001416419,0.004600752,0.0106851,0.002779579,0.0132577,0.004736038,0.006015596,0.001945645],"category_scores_gemma":[0.8146713,0.003035053,0.002872875,0.01136632,0.005361356,0.01515407,0.005495262,0.004702935,0.001077654],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002789937,"about_ca_system_score_gemma":0.004774033,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006935294,"about_ca_topic_score_gemma":0.009676922,"domain_scores_codex":[0.3025701,0.4838029,0.152521,0.01010876,0.04912936,0.001867997],"domain_scores_gemma":[0.07695229,0.8013467,0.04396985,0.03890542,0.03759352,0.001232299],"domain_codex":"methods","domain_gemma":"methods","domain_candidate":"methods","domain_consensus":"methods","study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001969292,0.0006121995,0.4500445,0.0429542,0.0111334,0.00158603,0.01387121,0.003118193,0.001847604,0.03059278,0.03869794,0.4035725],"study_design_scores_gemma":[0.001792956,0.002475813,0.4806187,0.0689415,0.01528982,0.009479347,0.01444904,0.02005145,0.009962529,0.1398808,0.2358138,0.001244307],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1718683,0.205043,0.3534701,0.1744518,0.02883421,0.01135511,0.007546504,0.0008671281,0.0465639],"genre_scores_gemma":[0.6484101,0.02705073,0.2684056,0.0344338,0.01010468,0.006918174,0.001780565,0.0002570174,0.002639286],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5059807,"threshold_uncertainty_score":0.6239644,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1610533120942863,"score_gpt":0.4712441419890614,"score_spread":0.3101908298947751,"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."}}