{"id":"W4398201968","doi":"10.1001/jamanetworkopen.2024.12687","title":"Assessing the Risk of Bias in Randomized Clinical Trials With Large Language Models","year":2024,"lang":"en","type":"article","venue":"JAMA Network Open","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":89,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University; Impact","funders":"","keywords":"Randomized controlled trial; Medicine; Systematic review; Reliability (semiconductor); Risk assessment; MEDLINE; Physical therapy; Internal medicine; Computer science","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":[0.06774247,0.00009494408,0.0009927652,0.00005143386,0.00007137919,0.0002345045,0.000178193,0.0001277744,0.0001291414],"category_scores_gemma":[0.006956445,0.00004595704,0.000189628,0.0004235909,0.00008989786,0.0003185233,0.00006607208,0.0005195786,0.000017065],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003105199,"about_ca_system_score_gemma":0.0004681538,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002811366,"about_ca_topic_score_gemma":0.0005528467,"domain_scores_codex":[0.9928401,0.005166857,0.001351682,0.0002239178,0.0001887385,0.0002286759],"domain_scores_gemma":[0.9813212,0.01788027,0.0003254794,0.0003136435,0.00008923648,0.00007012981],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.06919245,0.0004679182,0.1411823,0.0003138646,0.0008341188,0.0001088338,0.0209833,0.008970935,0.000008403736,0.0110876,0.01948371,0.7273666],"study_design_scores_gemma":[0.02567606,0.0003843631,0.01315756,0.007301139,0.001301146,0.00002906928,0.02318376,0.8906436,0.0002608209,0.03321481,0.004534942,0.0003127303],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9649183,0.005856235,0.01135539,0.007961389,0.0007968512,0.002715562,0.000005866322,0.00002934498,0.006361051],"genre_scores_gemma":[0.9949674,0.001155003,0.001293673,0.0006942992,0.001381013,0.0001085447,0.00001925232,0.00001732231,0.0003635069],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8816726,"threshold_uncertainty_score":0.9599553,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4904734455333462,"score_gpt":0.5865774049141176,"score_spread":0.09610395938077138,"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."}}