{"id":"W4389036121","doi":"10.23889/ijpds.v8i1.2176","title":"Generating synthetic data from administrative health records for drug safety and effectiveness studies","year":2023,"lang":"en","type":"article","venue":"International Journal for Population Data Science","topic":"Advanced Causal Inference Techniques","field":"Mathematics","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; Manitoba Health; McGill University; University of Manitoba","funders":"Canadian Institutes of Health Research","keywords":"Real world data; Population; Medical prescription; Computer science; Medicine; Data science; Pharmacology; Environmental health","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true,"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":[0.03475294,0.0008954019,0.0007303579,0.003280087,0.0005703774,0.001852492,0.002334914,0.001134332,0.002137017],"category_scores_gemma":[0.108987,0.0006774883,0.002348001,0.003681832,0.000801752,0.0009901838,0.001832878,0.001452938,0.000509216],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002354303,"about_ca_system_score_gemma":0.003616134,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0165402,"about_ca_topic_score_gemma":0.01342595,"domain_scores_codex":[0.9827911,0.01173446,0.001617934,0.001914278,0.001585301,0.000356915],"domain_scores_gemma":[0.8845395,0.08250091,0.01021972,0.01429001,0.007429683,0.00102012],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001357322,0.0006500478,0.431515,0.001416298,0.001668131,0.0006868237,0.000905706,0.4409963,0.0008954389,0.02237058,0.03551169,0.06202665],"study_design_scores_gemma":[0.001114406,0.0007048874,0.09730221,0.0008220474,0.0005310792,0.0005157219,0.0008619733,0.809751,0.002565169,0.04684735,0.03875735,0.0002267823],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.5877399,0.001477115,0.2058225,0.005103218,0.000648599,0.003950485,0.1884856,0.001968532,0.004803956],"genre_scores_gemma":[0.6684603,0.0005276545,0.1905967,0.0009066283,0.0002082004,0.002841617,0.1357208,0.0001138391,0.0006242156],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.03475294,"threshold_uncertainty_score":0.1837933,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5555152502297888,"score_gpt":0.5974494260819023,"score_spread":0.04193417585211356,"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."}}