{"meta":{"query_hash":"5570edbb5698","filters":{"venue":"BMJ Digital Health & AI"},"cohort_total":4,"direct_labels_cover":0,"predictions_cover":4,"exported":4,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/5570edbb5698","api":"https://metacan.xera.ac/api/v1/cohort?venue=BMJ+Digital+Health+%26+AI"},"results":[{"id":"W4413252984","doi":"10.1136/bmjdhai-2025-000114","title":"Development and prospective evaluation of a machine learning model to predict serious cardiac outcomes among paediatric cardiac inpatients","year":2025,"lang":"en","type":"article","venue":"BMJ Digital Health & AI","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Canadian Patient Safety Institute; SickKids Foundation; Toronto General Hospital; Casey House; Institute for Clinical Evaluative Sciences; Hospital for Sick Children","funders":"","keywords":"Retrospective cohort study; Medicine; Receiver operating characteristic; Logistic regression; Prospective cohort study; Emergency medicine; Internal medicine","score_opus":0.02054444801677871,"score_gpt":0.3433876705748698,"score_spread":0.32284322255809106,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4413252984","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9684686,0.0012725834,0.021963183,0.00216192,0.0005878588,0.0040474487,0.00005653534,0.00023764589,0.0012042077],"genre_scores_gemma":[0.98826605,0.000018651068,0.010592691,0.00041393685,0.00003176865,0.0003860596,0.000032267108,0.000017786188,0.00024079483],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9967848,0.00032098498,0.0007626691,0.0006301279,0.0010292755,0.00047214585],"domain_scores_gemma":[0.9983469,0.00015017718,0.0003068553,0.00042684024,0.00048931665,0.0002799037],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0018703709,0.00024635455,0.0005595938,0.00038660126,0.00027355537,0.00014694284,0.00033243725,0.00008162929,7.0583803e-7],"category_scores_gemma":[0.00078048423,0.00023512499,0.00008322545,0.00081729714,0.000032838037,0.00054184813,0.0005655007,0.00036808263,0.0000055031996],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000124780045,0.000028932578,0.73639065,0.00023610747,0.00003126877,3.1137196e-7,0.0029329928,0.010692679,2.1159963e-7,0.0006066873,0.00009193121,0.24897574],"study_design_scores_gemma":[0.00036653408,0.00016695342,0.64106053,0.00011967632,0.000010967565,4.1619532e-7,0.00005978875,0.3571169,0.0000072139983,0.0004542295,0.00047805347,0.00015871934],"about_ca_topic_score_codex":0.00034567225,"about_ca_topic_score_gemma":0.000057352634,"teacher_disagreement_score":0.34642425,"about_ca_system_score_codex":0.0005816164,"about_ca_system_score_gemma":0.0017912858,"threshold_uncertainty_score":0.9588119},"labels":[],"label_agreement":null},{"id":"W4414189579","doi":"10.1136/bmjdhai-2025-000014","title":"Optimising large language models for clinical information extraction: a benchmarking study in the context of ulcerative colitis research","year":2025,"lang":"en","type":"article","venue":"BMJ Digital Health & AI","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"Clinical and Translational Science Institute, University of California, Los Angeles; Clinical and Translational Science Institute, University of California, San Francisco","keywords":"Benchmarking; Context (archaeology); Adaptation (eye); Oracle; Language model; Set (abstract data type); Predictive modelling; Colonoscopy","score_opus":0.11497063511652149,"score_gpt":0.5230203048782074,"score_spread":0.4080496697616859,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4414189579","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.94277376,0.0015668207,0.036486134,0.00918843,0.00038214168,0.0040428,0.00029889203,0.000021974522,0.0052390704],"genre_scores_gemma":[0.9976831,0.00006859194,0.00048165474,0.0013282567,0.0001058163,0.00012558732,0.00012388747,0.000003097287,0.00008002219],"study_design_codex":"design_other","study_design_gemma":"qualitative","domain_scores_codex":[0.9986291,0.0002508255,0.0005181264,0.00015601268,0.0001907834,0.00025514656],"domain_scores_gemma":[0.99913687,0.00035295167,0.00010567327,0.00019254362,0.00017309451,0.000038842725],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003135381,0.000067798224,0.00017273924,0.00008161768,0.00012513921,0.000068129266,0.00015533951,0.000107557615,0.0000012426618],"category_scores_gemma":[0.00075897406,0.000051151164,0.000054463973,0.00018602792,0.00010063686,0.000025928237,0.00007804556,0.00019462,7.020527e-7],"study_design_candidate":"design_other","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0012234973,0.00070796465,0.0042632744,0.00027020485,0.000059013673,0.00000408344,0.009693703,0.00003465832,0.000011500351,0.0026141426,0.02724208,0.9538759],"study_design_scores_gemma":[0.0149054825,0.016266048,0.1199334,0.0011573682,0.000031375712,0.00003709017,0.47401172,0.026353134,0.0005843525,0.0038786258,0.34215665,0.0006847409],"about_ca_topic_score_codex":0.00008527783,"about_ca_topic_score_gemma":0.00036498628,"teacher_disagreement_score":0.95319116,"about_ca_system_score_codex":0.00002795718,"about_ca_system_score_gemma":0.00037604143,"threshold_uncertainty_score":0.20858839},"labels":[],"label_agreement":null},{"id":"W7105645746","doi":"10.1136/bmjdhai-2025-000078","title":"“Trying to Do No Harm”: exploring clinician concerns towards the use of AI for risk prediction in psychiatry","year":2025,"lang":"en","type":"article","venue":"BMJ Digital Health & AI","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Waypoint Centre for Mental Health Care; University of Toronto","funders":"","keywords":"Mental health; Thematic analysis; Focus group; Health care; Reflexivity; Quality (philosophy); MEDLINE; Mental health care","score_opus":0.4641667768071848,"score_gpt":0.5330326401816575,"score_spread":0.0688658633744727,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W7105645746","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8753441,0.0005270722,0.0075027,0.10385175,0.007343113,0.0045205704,0.0002552373,0.00007538604,0.00058007526],"genre_scores_gemma":[0.9830016,0.00021664886,0.00086287834,0.013793377,0.0011634743,0.00050217035,0.000042150088,0.000018377448,0.00039933028],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99804866,0.000055463664,0.0010706623,0.00027841647,0.00019228303,0.00035449103],"domain_scores_gemma":[0.9986128,0.00036532545,0.00019491551,0.00033605032,0.00032342313,0.00016743665],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006814775,0.000119726814,0.0003001526,0.00018223474,0.00016644958,0.00006381853,0.00009329868,0.000084478765,0.000010710865],"category_scores_gemma":[0.0014689035,0.00009739971,0.00009733384,0.0004301325,0.00004456403,0.00040067555,0.00003235057,0.00033467673,0.000027044674],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00074192515,0.0001829526,0.3424937,0.0008288029,0.000020998115,4.1096934e-7,0.003330765,0.00022943308,0.0000022856734,0.0012282016,0.03294595,0.61799455],"study_design_scores_gemma":[0.0011913845,0.005873242,0.35915777,0.010751859,0.00012896118,0.0000122374895,0.024020312,0.015635587,0.0010030115,0.014513054,0.56713593,0.0005766504],"about_ca_topic_score_codex":0.0026839941,"about_ca_topic_score_gemma":0.00049603725,"teacher_disagreement_score":0.61741793,"about_ca_system_score_codex":0.00024267058,"about_ca_system_score_gemma":0.0018525758,"threshold_uncertainty_score":0.40574142},"labels":[],"label_agreement":null},{"id":"W7105695419","doi":"10.1136/bmjdhai-2025-000134","title":"Retrospective evaluation of a machine learning model to facilitate pharmacogenetic testing","year":2025,"lang":"en","type":"article","venue":"BMJ Digital Health & AI","topic":"Pharmacogenetics and Drug Metabolism","field":"Pharmacology, Toxicology and Pharmaceutics","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Toronto General Hospital; Institute for Clinical Evaluative Sciences; Hospital for Sick Children","funders":"Hospital for Sick Children","keywords":"Pharmacogenetics; Logistic regression; Medical prescription; Epidemiology; Retrospective cohort study; Cohort; Gradient boosting","score_opus":0.3195143293878861,"score_gpt":0.5259982311863415,"score_spread":0.20648390179845538,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W7105695419","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.92311895,0.01107543,0.0060424055,0.008592243,0.0015537392,0.004439671,0.0011404052,0.0002349301,0.043802224],"genre_scores_gemma":[0.9899352,0.00017449373,0.00081038417,0.006916897,0.00012293515,0.00021884807,0.00005136677,0.000028645534,0.0017412392],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99681664,0.000584036,0.0008725676,0.00051941106,0.00054374494,0.00066358305],"domain_scores_gemma":[0.9979783,0.00024920268,0.00028535756,0.0002230092,0.00083929836,0.00042483205],"candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0030173173,0.00029788498,0.0005072621,0.0003310627,0.0003126476,0.000040478004,0.00028888884,0.0001336885,0.00012287594],"category_scores_gemma":[0.0009700279,0.00031182292,0.00011842363,0.0008616209,0.00010013291,0.00016249955,0.00020377256,0.0007418166,0.00007085801],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00045383174,0.00034666847,0.01837206,0.000294786,0.00020104984,0.000003917029,0.0016149628,0.3080407,0.016464472,0.00062240753,0.00466164,0.6489235],"study_design_scores_gemma":[0.003041827,0.0002183802,0.0032892886,0.000083535706,0.00020250026,0.0000065675003,0.00006819222,0.9361272,0.007387252,0.0051678224,0.044095192,0.00031223954],"about_ca_topic_score_codex":0.000046357156,"about_ca_topic_score_gemma":0.000012965513,"teacher_disagreement_score":0.64861125,"about_ca_system_score_codex":0.00027966686,"about_ca_system_score_gemma":0.0011965057,"threshold_uncertainty_score":0.99993336},"labels":[],"label_agreement":null}]}