{"id":"W4413831675","doi":"10.5489/cuaj.9243","title":"Assessing the methodologic quality of systematic reviews using generative large language models","year":2025,"lang":"en","type":"article","venue":"Canadian Urological Association Journal","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University","funders":"","keywords":"Generative grammar; Quality (philosophy); Computer science; Natural language processing; Linguistics; Artificial intelligence; Epistemology; Philosophy","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"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":[0.658528,0.002730684,0.005673398,0.01559801,0.002102684,0.007997049,0.003706115,0.002501157,0.002422142],"category_scores_gemma":[0.8123941,0.00226874,0.01579396,0.009516074,0.005720222,0.005771046,0.005533126,0.002310657,0.0003660833],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007222017,"about_ca_system_score_gemma":0.0114094,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0023389,"about_ca_topic_score_gemma":0.005820647,"domain_scores_codex":[0.2408587,0.5978396,0.09641978,0.01897245,0.04509721,0.0008121665],"domain_scores_gemma":[0.06781805,0.7921354,0.07098458,0.04357129,0.02478534,0.0007052714],"domain_codex":"methods","domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.008822097,0.0005161986,0.2216342,0.1921014,0.1200923,0.0006366674,0.01775843,0.02673382,0.004812052,0.01950911,0.008437334,0.3789464],"study_design_scores_gemma":[0.01148552,0.01060892,0.1701892,0.1277967,0.1628589,0.00297874,0.004571151,0.237203,0.02341949,0.1853894,0.06067029,0.002828628],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1185398,0.05721836,0.7758446,0.006051963,0.001251806,0.03037347,0.003490777,0.00227523,0.004954053],"genre_scores_gemma":[0.5399135,0.00410457,0.4289647,0.001338762,0.0003261925,0.02363045,0.001173695,0.0002926392,0.0002554744],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.341472,"threshold_uncertainty_score":0.4210958,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.866523665420036,"score_gpt":0.5981163913023423,"score_spread":0.2684072741176937,"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."}}