{"id":"W4411259709","doi":"10.1101/2025.06.13.25329541","title":"Automation of Systematic Reviews with Large Language Models","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Ontario; University Health Network; Ottawa Hospital; University of Alberta; St. Michael's Hospital; University of British Columbia; Mount Sinai Hospital; University of Calgary; McGill University; University of Ottawa; Vector Institute; Wilfrid Laurier University; University of Waterloo; University of Toronto","funders":"","keywords":"Automation; Computer science; Systems engineering; Engineering; Mechanical engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001228437,0.0002415937,0.001028464,0.0002996799,0.00003176462,0.0000516342,0.001539202,0.0001235059,0.000004702901],"category_scores_gemma":[0.0001747477,0.0001711798,0.0001904861,0.000482504,0.0000205896,0.0002564962,0.0008822593,0.0002477948,0.00001020798],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000611522,"about_ca_system_score_gemma":0.00006774404,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001260063,"about_ca_topic_score_gemma":0.00002087793,"domain_scores_codex":[0.9979208,0.0003220331,0.0007493263,0.000471603,0.0003618193,0.000174467],"domain_scores_gemma":[0.9967693,0.0001119133,0.0008740939,0.002047116,0.0001612995,0.00003626108],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001337095,0.0006220863,0.001617758,0.4984612,0.0008779691,0.00009034546,0.01629764,0.01983959,0.001749038,0.4491812,0.001392903,0.009856931],"study_design_scores_gemma":[0.0001655652,0.0000386017,0.00009788044,0.03897478,0.0002540906,0.000004615091,0.00004408657,0.9292596,0.005394591,0.02516383,0.0001302721,0.0004720755],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.002132013,0.003465275,0.9907868,0.00005830746,0.00006276121,0.001299882,0.000007604832,0.0003938107,0.001793544],"genre_scores_gemma":[0.4476429,0.0005356172,0.5496737,0.0001100173,0.00002335061,0.0007046095,0.00001820843,0.00001482701,0.001276786],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.90942,"threshold_uncertainty_score":0.6980509,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03430894453732251,"score_gpt":0.3234808861615266,"score_spread":0.2891719416242041,"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."}}