{"id":"W4386533461","doi":"10.21203/rs.3.rs-3288515/v1","title":"Using a large language model (ChatGPT) to assess risk of bias in randomized controlled trials of medical interventions: protocol for a pilot study of interrater agreement with human reviewers","year":2023,"lang":"en","type":"preprint","venue":"Research Square","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Inter-rater reliability; Protocol (science); Psychological intervention; Randomized controlled trial; Psychology; Medicine; Clinical psychology; Medical physics; Psychiatry; Alternative medicine; Developmental psychology; Internal medicine; Pathology; Rating scale","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":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.3085962,0.007165108,0.008336275,0.008152573,0.004960866,0.005316569,0.005216366,0.009021068,0.06344141],"category_scores_gemma":[0.4325734,0.005830495,0.01315833,0.008513311,0.007556554,0.007488484,0.006007578,0.01168769,0.01785628],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01234007,"about_ca_system_score_gemma":0.03500834,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002846163,"about_ca_topic_score_gemma":0.005858705,"domain_scores_codex":[0.6901212,0.2154163,0.06091489,0.01282465,0.01548569,0.005237277],"domain_scores_gemma":[0.5213791,0.2526847,0.06383613,0.07460038,0.07995524,0.007544506],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"not_applicable","study_design_scores_codex":[0.2129648,0.009081952,0.005624969,0.2927328,0.009827838,0.002064053,0.0167981,0.0172606,0.01010252,0.02867753,0.1409441,0.2539208],"study_design_scores_gemma":[0.4146365,0.03307055,0.01567524,0.1116628,0.006409142,0.001120132,0.004604434,0.03409212,0.01376944,0.06133425,0.3013494,0.002275942],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"protocol","genre_gemma":"protocol","genre_scores_codex":[0.0005723117,0.0001089582,0.005263583,0.0001814399,0.0002004397,0.9923066,0.0008475558,0.0002051363,0.0003139811],"genre_scores_gemma":[0.0002808359,0.00001993117,0.003993884,0.00002788239,0.000007157288,0.9955622,0.00005150442,0.000007754757,0.0000488476],"genre_candidate":"protocol","genre_consensus":"protocol","teacher_disagreement_score":0.6914037,"threshold_uncertainty_score":0.8526239,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.8791265894664788,"score_gpt":0.707523575191614,"score_spread":0.1716030142748648,"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."}}