{"id":"W4407119832","doi":"10.1097/js9.0000000000002215","title":"AI-driven evidence synthesis: data extraction of randomized controlled trials with large language models","year":2025,"lang":"en","type":"article","venue":"International Journal of Surgery","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University; Impact","funders":"Fundamental Research Funds for the Central Universities; China Academy of Chinese Medical Sciences; National Natural Science Foundation of China","keywords":"Data extraction; Randomized controlled trial; Medicine; Meta-analysis; Systematic review; MEDLINE; Artificial intelligence; Data mining; Computer science; Natural language processing; 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":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0131057,0.0001024244,0.002034812,0.0006413619,0.00002734971,0.00003468696,0.0002454256,0.00007086833,0.0001719212],"category_scores_gemma":[0.03450507,0.00006247129,0.0005085145,0.0001445709,0.00006286364,0.000640596,0.00003056517,0.0002097275,0.000002788983],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009164691,"about_ca_system_score_gemma":0.001114742,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000186097,"about_ca_topic_score_gemma":0.00001833308,"domain_scores_codex":[0.9961708,0.0006576212,0.002263676,0.0001317625,0.0006609312,0.0001152056],"domain_scores_gemma":[0.9714074,0.02421216,0.001984773,0.0002743002,0.002049385,0.00007205595],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"randomized_trial","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.86983,0.001306156,0.01964158,0.0003337732,0.009162874,0.0003298606,0.001819349,0.001213004,0.007600541,0.00253464,0.01463564,0.07159255],"study_design_scores_gemma":[0.2222108,0.0008072366,0.008549327,0.1179888,0.03030903,0.003127972,0.03270004,0.2781326,0.2732168,0.0268339,0.004471414,0.001652147],"study_design_candidate":"randomized_trial","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8030081,0.01285832,0.1329672,0.04355226,0.005429273,0.001357169,0.00005725522,0.000026103,0.0007442975],"genre_scores_gemma":[0.9953103,0.002673275,0.0006110223,0.0006365114,0.000570847,0.00002504688,0.00001422797,0.000007324156,0.0001514532],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6476192,"threshold_uncertainty_score":0.9736277,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.310832927540387,"score_gpt":0.5076900762546582,"score_spread":0.1968571487142712,"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."}}