{"id":"W4387963398","doi":"10.1093/jamia/ocad205","title":"Systematic replication of smoking disease associations using survey responses and EHR data in the <i>All of Us</i> Research Program","year":2023,"lang":"en","type":"article","venue":"Journal of the American Medical Informatics Association","topic":"Health, Environment, Cognitive Aging","field":"Environmental Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"The Scarborough Hospital; University of Toronto","funders":"National Human Genome Research Institute; National Institutes of Health","keywords":"Phenome; Bonferroni correction; Biobank; Replication (statistics); Meta-analysis; Medicine; MEDLINE; Disease; Bioinformatics; Phenotype; Internal medicine; Biology; Genetics; Statistics","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":["metaresearch"],"category_scores_codex":[0.05365726,0.00007052154,0.0003072149,0.0001368507,0.0001306916,0.00003376627,0.0007691839,0.00004721984,0.000004837966],"category_scores_gemma":[0.05578537,0.00004485229,0.00004547292,0.001347361,0.0002906062,0.0003547684,0.0004242242,0.0004681232,0.000002334857],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005529172,"about_ca_system_score_gemma":0.0001514465,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008366142,"about_ca_topic_score_gemma":0.000156177,"domain_scores_codex":[0.9925172,0.00324901,0.001235058,0.0001049494,0.002651172,0.0002425853],"domain_scores_gemma":[0.9905483,0.005490684,0.003162294,0.0006011814,0.00009257454,0.0001050061],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00002213816,0.0001184115,0.9922929,0.0003680395,0.00004941884,9.162582e-7,0.002718808,0.0003514393,0.0000892562,0.000007163222,0.0004269849,0.003554557],"study_design_scores_gemma":[0.0001733872,0.00005985974,0.952568,0.0006075326,0.00005343865,0.000002560926,0.00158097,0.04468205,0.00001600614,0.0001571785,0.00005671658,0.0000422914],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9976948,0.00001723726,0.000189716,0.00145286,0.0000364507,0.0004930669,0.00005471307,0.000004404023,0.00005671843],"genre_scores_gemma":[0.9988362,0.0002621952,0.0004619271,0.0003791704,0.00001921728,0.000008193464,0.00001918631,0.000006623457,0.000007338417],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04433061,"threshold_uncertainty_score":0.974459,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1401989339252414,"score_gpt":0.4433557366599765,"score_spread":0.3031568027347351,"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."}}