{"id":"W4405906449","doi":"10.2196/59298","title":"Assessing the Data Quality Dimensions of Partial and Complete Mastectomy Cohorts in the <i>All of Us</i> Research Program: Cross-Sectional Study","year":2024,"lang":"en","type":"article","venue":"JMIR Cancer","topic":"Global Cancer Incidence and Screening","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Preprint; Cross-sectional study; Quality (philosophy); Psychology; Gerontology; Medicine; Computer science; Statistics; World Wide Web; Mathematics; Physics","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003567843,0.00008056511,0.0002124558,0.00005912362,0.0001148546,0.0001314145,0.000256833,0.00004151936,0.00004286104],"category_scores_gemma":[0.0001085537,0.00004019016,0.00003434026,0.0005068516,0.0003670692,0.0002029454,0.0002427848,0.0004316811,9.51717e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006191825,"about_ca_system_score_gemma":0.0002832707,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007526134,"about_ca_topic_score_gemma":0.001979758,"domain_scores_codex":[0.9978348,0.0003456838,0.0003858702,0.0003156585,0.0008989152,0.0002190961],"domain_scores_gemma":[0.998761,0.000425982,0.00006355304,0.0005279108,0.0001827962,0.0000387435],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001264437,0.0001698209,0.9930511,0.000142911,0.00006663177,0.00002678559,0.001247447,0.0000181372,0.001109104,0.00009529637,0.001049835,0.002896537],"study_design_scores_gemma":[0.0004549105,0.0002205533,0.9831402,0.0002260261,0.00003467154,0.00001695801,0.001827315,0.0008086751,0.00005239413,0.00005588061,0.01311697,0.00004538777],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9957359,0.001564188,0.000007152022,0.0008444188,0.0001152024,0.001320151,0.00008596274,0.00001552906,0.0003114392],"genre_scores_gemma":[0.9992797,0.00004119859,0.00006230878,0.0001390083,0.0001491804,0.0002727795,0.00002047684,0.000007675157,0.00002772175],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01206713,"threshold_uncertainty_score":0.9990829,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.609466669380634,"score_gpt":0.6237969928738044,"score_spread":0.01433032349317043,"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."}}