{"id":"W3011708395","doi":"10.1136/bmjopen-2019-034305","title":"Constructing treatment episodes from concomitant medication logs: a prospective observational study","year":2020,"lang":"en","type":"article","venue":"BMJ Open","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vancouver Coastal Health Research Institute; University of British Columbia; International Collaboration On Repair Discoveries; Vancouver Coastal Health","funders":"Avid Radiopharmaceuticals; Sanofi Genzyme; Allergan; Genentech; H. Lundbeck A/S; Servier; Voyager Therapeutics; Biogen; Celgene; Verily Life Sciences; Teva Pharmaceutical Industries; Sanofi; GlaxoSmithKline; Pfizer; Eli Lilly and Company; Bristol-Myers Squibb; Michael J. Fox Foundation for Parkinson's Research","keywords":"Medicine; Observational study; Concomitant; Intensive care medicine; Family medicine; Emergency medicine; Internal medicine","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":[0.0113964,0.0004372112,0.0006853277,0.002260705,0.0009113997,0.001229903,0.0008427584,0.0007487993,0.001313905],"category_scores_gemma":[0.03092766,0.0006893234,0.001053101,0.003430013,0.0004346643,0.00136741,0.001242531,0.001183258,0.0004006746],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006195706,"about_ca_system_score_gemma":0.002168266,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007064581,"about_ca_topic_score_gemma":0.009939075,"domain_scores_codex":[0.9877359,0.007338969,0.002191664,0.001045857,0.001317139,0.0003704692],"domain_scores_gemma":[0.9680132,0.01101885,0.01463255,0.003143242,0.002143793,0.001048227],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003567694,0.0003870029,0.9921852,0.0001005041,0.0001472467,0.0000934439,0.0004328046,0.0001514775,0.0001159652,0.00007648717,0.0003463807,0.005606746],"study_design_scores_gemma":[0.0002004559,0.002228084,0.9892985,0.0001872748,0.0003062668,0.0005241859,0.00166961,0.003308658,0.0002426356,0.0002784831,0.001704003,0.00005193926],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9828934,0.0006849994,0.0095186,0.0001652605,0.00002559483,0.00126093,0.004756405,0.00003671291,0.0006580642],"genre_scores_gemma":[0.9712622,0.0005822362,0.01942718,0.0002020505,0.00005698848,0.001723752,0.006492066,0.00001619607,0.0002373634],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0113964,"threshold_uncertainty_score":0.06027061,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2750850702495213,"score_gpt":0.453766932212457,"score_spread":0.1786818619629357,"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."}}