{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":679,"total_is_capped":false,"direct_labels_cover":6,"predictions_cover":679,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"query_hash":"68b1c0c81223","filters":{"topic":"Medical Coding and Health Information"}},"results":[{"id":"W2061326496","doi":"10.1097/01.mlr.0000182534.19832.83","title":"Coding Algorithms for Defining Comorbidities in ICD-9-CM and ICD-10 Administrative Data","year":2005,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":10649,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Calgary; University of Alberta","funders":"","keywords":"ICD-10; Algorithm; Medicine; Coding (social sciences); Diagnosis code; Comorbidity; Current Procedural Terminology; Data mining; Computer science; Statistics; Population; Internal medicine; Mathematics; Surgery","retraction":null,"screen_n_in":null,"score":{"opus":0.4388182480607153,"gpt":0.537183217276309,"spread":0.09836496921559368,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00197658,0.0001296759,0.0003057009,0.00008788841,0.0005827587,0.00001274154,0.0002916649,0.0003326833,0.001706898],"category_scores_gemma":[0.004014481,0.0001088117,0.00001982139,0.000104175,0.000115628,0.0002188869,0.0002187529,0.0008333655,0.0001251634],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001554791,"about_ca_system_score_gemma":0.0009343986,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000137219,"about_ca_topic_score_gemma":0.001551016,"domain_scores_codex":[0.9977705,0.0002001768,0.0007239121,0.0002597936,0.0005292639,0.0005162859],"domain_scores_gemma":[0.9972259,0.001770055,0.000154548,0.0002817989,0.0001276279,0.0004400583],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002768842,0.00008019509,0.02192427,0.005903879,0.00002809031,0.0000369266,0.1012313,0.000004014385,0.000001863321,0.01633288,0.3035918,0.550588],"study_design_scores_gemma":[0.004055376,0.0003237283,0.008146698,0.002209956,0.00002689568,0.00001035826,0.06628975,0.04358803,0.00001085165,0.000263817,0.8747505,0.0003240148],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4535385,0.01968724,0.04682188,0.220881,0.008115213,0.01073845,0.005697131,0.001353132,0.2331675],"genre_scores_gemma":[0.9795955,0.0003363031,0.006529239,0.009076497,0.001281534,0.0002738321,0.00178312,0.00002389676,0.0011001],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5711588,"threshold_uncertainty_score":0.9992056,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2121656227","doi":"10.1093/aje/kwq433","title":"Updating and Validating the Charlson Comorbidity Index and Score for Risk Adjustment in Hospital Discharge Abstracts Using Data From 6 Countries","year":2011,"lang":"en","type":"article","venue":"American Journal of Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":5897,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Calgary","funders":"Hospices Civils de Lyon; Tokyo Medical and Dental University; Alberta Innovates; Faculty of Medicine, Nursing and Health Sciences, Monash University; Monash University; University of Otago","keywords":"Medicine; Charlson comorbidity index; Comorbidity; Index (typography); Emergency medicine; Hospital discharge; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.432246237228286,"gpt":0.4849603682303741,"spread":0.05271413100208816,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0104858,0.0001033208,0.0005419207,0.00006681263,0.0004328897,0.000003550103,0.0002047771,0.00008364127,0.00004332277],"category_scores_gemma":[0.01013908,0.0000641915,0.00002131153,0.00005973408,0.000314946,0.0002564925,0.0001453618,0.000827688,0.000001724292],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006083892,"about_ca_system_score_gemma":0.0001844549,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006967623,"about_ca_topic_score_gemma":0.000254368,"domain_scores_codex":[0.9967071,0.001255575,0.001412806,0.0001538373,0.00009466191,0.000376037],"domain_scores_gemma":[0.9889162,0.007667585,0.002939803,0.0002088961,0.00009897401,0.0001685709],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001388687,0.00001944404,0.979825,0.00007337926,0.00002675866,0.000001141281,0.006591304,0.00001095087,0.000002507638,0.0005475397,0.0008476013,0.01191552],"study_design_scores_gemma":[0.0008680235,0.0004727558,0.9647405,0.0004756637,0.00005344622,0.00001027003,0.01464408,0.01420474,0.000001722123,0.002577977,0.001853828,0.00009706135],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9873673,0.0005521997,0.008505302,0.00261214,0.0004610149,0.0003163318,0.0001237261,0.000005255435,0.00005677267],"genre_scores_gemma":[0.9847726,0.001565605,0.01111413,0.002091978,0.0004089396,0.000007811854,0.00002992706,0.000007911321,0.000001146718],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01508456,"threshold_uncertainty_score":0.9996451,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2165037395","doi":"10.1016/j.jclinepi.2004.03.012","title":"New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality","year":2004,"lang":"en","type":"article","venue":"Journal of Clinical Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":2484,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"","keywords":"Generalizability theory; ICD-10; Medicine; Comorbidity; Charlson comorbidity index; Coding (social sciences); Diagnosis code; Population; Statistics; Internal medicine; Psychiatry","retraction":null,"screen_n_in":null,"score":{"opus":0.4828866030270053,"gpt":0.5947186039689906,"spread":0.1118320009419853,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.03103334,0.0001136487,0.00150846,0.00008887474,0.0001568419,8.996577e-7,0.0004045765,0.0007654236,0.0006192633],"category_scores_gemma":[0.09109595,0.00006683288,0.0003541831,0.0002081022,0.0002007557,0.0001386025,0.0001245976,0.003147545,0.00004886642],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001601066,"about_ca_system_score_gemma":0.002145845,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007799472,"about_ca_topic_score_gemma":0.0001458451,"domain_scores_codex":[0.9874949,0.004675962,0.006892046,0.0001389256,0.0003763135,0.0004218219],"domain_scores_gemma":[0.9793194,0.01339773,0.006083848,0.0003327865,0.0003625009,0.0005037694],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003353971,0.0001478549,0.9593135,0.0001480834,0.00002442835,0.000002615976,0.0003922533,0.0002658539,9.243175e-7,0.001514042,0.03519277,0.002662275],"study_design_scores_gemma":[0.003428593,0.0007400058,0.9658185,0.0006716225,0.00002385149,0.00000267323,0.0001384861,0.0004404386,0.000001887736,0.011322,0.01735968,0.00005226249],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9476891,0.00008648602,0.004944015,0.04049573,0.005218324,0.000315393,0.00001037982,0.00001113671,0.001229493],"genre_scores_gemma":[0.9920881,0.0005265938,0.0008383642,0.0051196,0.001293906,0.000002278726,0.000004703384,0.000006267193,0.0001201679],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.06006261,"threshold_uncertainty_score":0.9991522,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1582456500","doi":"10.1111/j.1475-6773.2007.00822.x","title":"Assessing Validity of ICD‐9‐CM and ICD‐10 Administrative Data in Recording Clinical Conditions in a Unique Dually Coded Database","year":2008,"lang":"en","type":"article","venue":"Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":892,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Alberta Health Services; University of Alberta; University of Calgary","funders":"Canadian Institutes of Health Research; Hochschule Niederrhein; Université de Lyon; Fondation pour la Recherche Médicale; Tokyo Medical and Dental University; Université de Lausanne; University of Toronto; Government of Canada","keywords":"Medicine; ICD-10; Chart; Positive predicative value; Database; Statistics; Predictive value; Internal medicine; Psychiatry; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.8304412398261176,"gpt":0.6963972829176036,"spread":0.134043956908514,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.03058983,0.0001551841,0.0006975923,0.0005983398,0.00130381,0.00002250808,0.0005622841,0.0003446808,0.0004834371],"category_scores_gemma":[0.002298519,0.0001462025,0.00002211087,0.001083113,0.0003125033,0.001075088,0.0007018379,0.00308309,0.00007482237],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003087703,"about_ca_system_score_gemma":0.006591225,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01646044,"about_ca_topic_score_gemma":0.02635756,"domain_scores_codex":[0.9874911,0.007424815,0.002541126,0.0005504512,0.0009043369,0.001088153],"domain_scores_gemma":[0.9889411,0.008080157,0.0008056254,0.0009323748,0.0005292,0.0007115941],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0006688444,0.0007299429,0.9226592,0.02378606,0.0000162934,0.0001056104,0.03401215,0.000005276965,0.00003044556,0.001516249,0.01161427,0.004855696],"study_design_scores_gemma":[0.00374393,0.000781893,0.9040841,0.01014055,0.000004964929,0.00001695389,0.03899887,0.03199801,0.00001152383,0.0007353241,0.009234769,0.0002490891],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9874098,0.0002144851,0.0001826504,0.006999689,0.0002239911,0.001773022,0.0004545183,0.00004201192,0.002699789],"genre_scores_gemma":[0.990321,0.003140465,0.002470023,0.002144827,0.0001794151,0.0001370582,0.001414336,0.00001861345,0.0001742472],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03199273,"threshold_uncertainty_score":0.9999964,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2171017045","doi":"10.1161/01.str.0000174293.17959.a1","title":"Coding of Stroke and Stroke Risk Factors Using <i>International Classification of Diseases</i> , Revisions 9 and 10","year":2005,"lang":"en","type":"article","venue":"Stroke","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":715,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Calgary","funders":"","keywords":"Medicine; Stroke (engine); Atrial fibrillation; Diabetes mellitus; Diagnosis code; Coding (social sciences); ICD-10; Coronary artery disease; Disease; Emergency medicine; Internal medicine; Population; Statistics","retraction":null,"screen_n_in":null,"score":{"opus":0.1642747832009984,"gpt":0.4362269011517596,"spread":0.2719521179507611,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005870573,0.0001099799,0.0002660604,0.0001662799,0.0003826205,0.000007444823,0.00009601647,0.0001389583,0.0005269525],"category_scores_gemma":[0.0007192135,0.00009433775,0.00004635044,0.00006539444,0.0001057383,0.0002233866,0.00007212922,0.000346943,0.00000865178],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007770553,"about_ca_system_score_gemma":0.000148048,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001286807,"about_ca_topic_score_gemma":0.00001815165,"domain_scores_codex":[0.9982987,0.0001745819,0.0007811072,0.0001575773,0.0003557719,0.0002322475],"domain_scores_gemma":[0.9981782,0.0005450522,0.0006966012,0.0001713314,0.0001989291,0.0002098763],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001637772,0.0000800094,0.9521807,0.0009734943,0.00006951254,2.263884e-7,0.002994295,0.00006555339,0.006121897,0.003636086,0.003877033,0.02983738],"study_design_scores_gemma":[0.002825301,0.0002171126,0.8460398,0.001126047,0.0002236027,0.000001613426,0.006918247,0.07452825,0.0006271679,0.00008955271,0.06710791,0.0002953626],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9878192,0.0002559116,0.002402169,0.0005739306,0.0002852897,0.0003309874,0.001443396,0.00002965583,0.006859442],"genre_scores_gemma":[0.9957542,0.0006611217,0.001692541,0.00009840363,0.0002002838,0.00000771788,0.00006020768,0.000009400056,0.001516067],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1061409,"threshold_uncertainty_score":0.5769761,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2991379615","doi":"10.2196/14325","title":"Mapping ICD-10 and ICD-10-CM Codes to Phecodes: Workflow Development and Initial Evaluation","year":2019,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":561,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"National Center for Advancing Translational Sciences; U.S. National Library of Medicine; National Heart, Lung, and Blood Institute; China Scholarship Council; National Institutes of Health; Cancer Research UK; Georgia Clinical and Translational Science Alliance; Vanderbilt University Medical Center; Vanderbilt University; National Institute of General Medical Sciences; American Heart Association","keywords":"ICD-10; Medicine; Systematized Nomenclature of Medicine; Biorepository; Health informatics; Diagnosis code; Informatics; Medicaid; Concordance; Medical classification; SNOMED CT; Electronic health record; Biobank; Health care; Internal medicine; Public health; Pathology; Terminology; Bioinformatics; Population","retraction":null,"screen_n_in":null,"score":{"opus":0.1801527069750413,"gpt":0.473251173470247,"spread":0.2930984664952057,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.005036595,0.0002184614,0.0004227724,0.0002566511,0.0006176769,0.0000421734,0.000189798,0.0005021158,0.006321845],"category_scores_gemma":[0.00144918,0.0001769965,0.00002482073,0.0002867219,0.00009114132,0.000438767,0.0002686996,0.0008362001,0.002566348],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002279036,"about_ca_system_score_gemma":0.001338993,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001179801,"about_ca_topic_score_gemma":0.00002707811,"domain_scores_codex":[0.9956657,0.0002170961,0.001727983,0.0001536384,0.001583171,0.0006523779],"domain_scores_gemma":[0.997306,0.0007438862,0.0003902767,0.0002471481,0.0003138225,0.0009988297],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001708077,0.00006414374,0.005343297,0.005917815,0.00004606635,0.000002029469,0.1805626,0.00001610908,0.000003990183,0.0009990542,0.06686356,0.7400106],"study_design_scores_gemma":[0.004889952,0.0003125053,0.01729879,0.004894572,0.00002629891,0.00001410963,0.03538783,0.1379372,0.00002184595,0.0004122267,0.7982081,0.0005966504],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9513605,0.0001004972,0.00464454,0.003135821,0.0007085172,0.002632802,0.00001063673,0.0001619928,0.03724466],"genre_scores_gemma":[0.92898,0.0002172792,0.02906704,0.03549918,0.0008715148,0.001263066,0.0004202019,0.00004636663,0.003635386],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7394139,"threshold_uncertainty_score":0.9982103,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1566902700","doi":"","title":"Accuracy of administrative databases in identifying patients with hypertension.","year":2007,"lang":"en","type":"article","venue":"PubMed","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":536,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Medicine; Confidence interval; Medical diagnosis; Predictive value; Population; Census; Health care; Pediatrics; Database; Demography; Internal medicine; Environmental health","retraction":null,"screen_n_in":null,"score":{"opus":0.5001758870461303,"gpt":0.4729391428250885,"spread":0.02723674422104178,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001624857,0.00006787266,0.000170622,0.0001373293,0.0001575272,0.000002011462,0.00007247501,0.00004702876,0.00005227003],"category_scores_gemma":[0.001918347,0.00004848773,0.00001188976,0.0002287521,0.0000295702,0.0002494259,0.00003680334,0.0003203562,0.00001774878],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007431127,"about_ca_system_score_gemma":0.0001369244,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002577736,"about_ca_topic_score_gemma":0.0005226186,"domain_scores_codex":[0.9984471,0.00009291442,0.0006279843,0.000104431,0.0002992696,0.0004282538],"domain_scores_gemma":[0.9981123,0.001047408,0.0003448053,0.0001538501,0.0001727935,0.0001689109],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003284914,0.0001035826,0.9167698,0.000492961,0.000005792775,0.000004064025,0.001529114,5.367109e-7,0.000001237378,0.0008253266,0.001386286,0.07855282],"study_design_scores_gemma":[0.001189634,0.00001913245,0.9934291,0.0001917082,0.000004771661,2.264008e-7,0.001124726,0.00001181161,0.00003652843,0.00002536752,0.003913475,0.00005356569],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9901391,0.00004129611,0.0009759781,0.0002713081,0.0002909361,0.00128572,0.00002103541,0.00002771804,0.006946878],"genre_scores_gemma":[0.9982151,0.00001869235,0.0004221131,0.0008713339,0.00005217908,0.0002510854,0.00006283858,0.000006038963,0.0001006667],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.07849925,"threshold_uncertainty_score":0.229658,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1967778685","doi":"10.1067/mhj.2002.123839","title":"A multicenter study of the coding accuracy of hospital discharge administrative data for patients admitted to cardiac care units in Ontario","year":2002,"lang":"en","type":"article","venue":"American Heart Journal","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":533,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Institute for Clinical Evaluative Sciences; University of Toronto","funders":"","keywords":"Medicine; Myocardial infarction; Medical diagnosis; Hospital discharge; Coronary care unit; Unstable angina; Chest pain; Angina; Diagnosis code; Heart failure; Emergency medicine; Coronary artery disease; Cardiology; Internal medicine; Intensive care medicine; Population; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.30024393590142,"gpt":0.4666259476293992,"spread":0.1663820117279792,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000569316,0.00009887839,0.0003807026,0.00009231958,0.0003996767,0.000006323835,0.0003063021,0.00003742307,0.0000983137],"category_scores_gemma":[0.002437156,0.00006724989,0.00004156388,0.0003252425,0.00004949623,0.0001867911,0.000155681,0.0007036191,0.00000627388],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002299037,"about_ca_system_score_gemma":0.0005097119,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01044017,"about_ca_topic_score_gemma":0.01464188,"domain_scores_codex":[0.9978153,0.0004576081,0.0008914142,0.0001329879,0.0003725179,0.0003302351],"domain_scores_gemma":[0.9975463,0.0007080356,0.0006595075,0.0003533472,0.0005262081,0.0002066356],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00007724144,0.0002236426,0.8678835,0.00006571043,0.0000178409,1.807532e-7,0.1264794,0.000001935224,0.000002741815,0.000003653775,0.004086392,0.001157813],"study_design_scores_gemma":[0.001425441,0.002143746,0.8804495,0.0004357817,0.00002340051,1.520173e-7,0.1072105,0.0001899697,0.000005054668,9.73237e-7,0.008040232,0.00007522709],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9961374,0.00001236377,0.00002521624,0.0008191604,0.0007288401,0.001966389,0.0001745329,0.000004579227,0.0001314706],"genre_scores_gemma":[0.999024,0.000009448479,0.0002644313,0.000490504,0.00007929085,0.00005406362,0.0000229797,0.000008727192,0.000046563],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0192689,"threshold_uncertainty_score":0.9961494,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2969218718","doi":"","title":"Charlson Comorbidity Index: <i>ICD-9</i> Update and <i>ICD-10</i> Translation.","year":2019,"lang":"en","type":"article","venue":"PubMed","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":526,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Office of the Chief Medical Examiner; Medicine Hat College","funders":"","keywords":"Medicine; ICD-10; Comparability; Comorbidity; Diagnosis code; Charlson comorbidity index; Medical record; Internal medicine; Population","retraction":null,"screen_n_in":null,"score":{"opus":0.1586748101747708,"gpt":0.3715498495472735,"spread":0.2128750393725027,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.001839352,0.0001398164,0.0002725042,0.00009456764,0.0004427157,0.00001629551,0.0001266578,0.0002598231,0.001025318],"category_scores_gemma":[0.0001313405,0.0001205902,0.00003542475,0.0001723773,0.00004351695,0.0003136543,0.00004874579,0.000674328,0.001306888],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008011037,"about_ca_system_score_gemma":0.0001113157,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009009599,"about_ca_topic_score_gemma":0.00003977391,"domain_scores_codex":[0.9979109,0.0002366772,0.000601952,0.0002212527,0.0003496503,0.0006796239],"domain_scores_gemma":[0.9987004,0.0003196404,0.000225802,0.0002627685,0.00008999996,0.000401388],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000320009,0.00006832035,0.1965366,0.002065418,0.00003018496,0.000001970602,0.002547577,0.000009358855,0.000009072854,0.004873847,0.0973525,0.6961851],"study_design_scores_gemma":[0.001577533,0.00001893041,0.3373942,0.0000521462,0.00001112864,0.000001226421,0.0001597303,0.0008348838,0.00001044472,0.0004646001,0.6593271,0.0001480652],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.786397,0.001230674,0.001668406,0.02954303,0.00348379,0.00653931,0.0001008782,0.000565551,0.1704714],"genre_scores_gemma":[0.9827675,0.0004828117,0.0001096096,0.01183682,0.0003382151,0.001004109,0.00008031577,0.00001885615,0.003361822],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6960371,"threshold_uncertainty_score":0.9998879,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1964304172","doi":"10.1186/1471-2288-8-1","title":"An administrative data merging solution for dealing with missing data in a clinical registry: adaptation from ICD-9 to ICD-10","year":2008,"lang":"en","type":"article","venue":"BMC Medical Research Methodology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":503,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Provincial Health Services Authority; University of British Columbia; University of Alberta; University of Calgary","funders":"Canadian Institutes of Health Research; Eli Lilly Canada; Fondation pour la Recherche Médicale; W. Garfield Weston Foundation; Eli Lilly and Company; University of Alberta; Government of Canada; Boston Scientific Corporation","keywords":"Missing data; Statistic; Medicine; Data mining; ICD-10; Computer science; Statistics; Mathematics; Machine learning","retraction":null,"screen_n_in":null,"score":{"opus":0.9723130920307268,"gpt":0.759938338111273,"spread":0.2123747539194538,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.0977209,0.0001852499,0.000752144,0.0002916022,0.00137579,0.00001693701,0.001882079,0.0007870065,0.0008114364],"category_scores_gemma":[0.3154645,0.0001488872,0.00003029524,0.0004497306,0.0005912038,0.0004941663,0.0008171002,0.002728362,0.00008889619],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002055471,"about_ca_system_score_gemma":0.01136106,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.007311806,"about_ca_topic_score_gemma":0.02659618,"domain_scores_codex":[0.969012,0.02432347,0.002140774,0.001173342,0.001978754,0.001371714],"domain_scores_gemma":[0.8529254,0.1400566,0.0006356674,0.00310677,0.0008730555,0.002402494],"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":[0.01884386,0.0006287436,0.03813229,0.002187704,0.0001147547,0.0002048792,0.02728314,0.0001032798,0.0002652163,0.006980167,0.06581361,0.8394424],"study_design_scores_gemma":[0.004616961,0.001775381,0.03351653,0.002074253,0.00003206243,0.00002108566,0.01018926,0.860641,0.00001139617,0.002047563,0.08474501,0.0003295165],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.04888191,0.0001944439,0.9358237,0.01215064,0.0004275734,0.001566561,0.0002311768,0.00007511709,0.0006488629],"genre_scores_gemma":[0.1115964,0.0004251165,0.8795316,0.002413274,0.001792592,0.0002665606,0.003752016,0.00003747765,0.0001850169],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8605377,"threshold_uncertainty_score":0.9999243,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W100013426","doi":"","title":"Coding accuracy of administrative drug claims in the Ontario Drug Benefit database.","year":2003,"lang":"en","type":"article","venue":"PubMed","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":493,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Centre for Advancing Health Outcomes; St. Paul's Hospital","funders":"","keywords":"Medicine; Pharmacy; Medical prescription; Database; Coding (social sciences); Audit; Family medicine; Logistic regression; Medical emergency; Statistics; Accounting; Nursing; Internal medicine; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.3582565692101561,"gpt":0.4382438260233484,"spread":0.07998725681319235,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004679634,0.0001048319,0.0002311921,0.00009193837,0.0003962457,0.000008853151,0.0002347011,0.00008272002,0.000320141],"category_scores_gemma":[0.001145908,0.00007427535,0.00003510989,0.000240627,0.00004792076,0.0002457701,0.00003430417,0.0009470678,0.00004382402],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002547733,"about_ca_system_score_gemma":0.0005291373,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.007239562,"about_ca_topic_score_gemma":0.06813954,"domain_scores_codex":[0.9976113,0.0004863529,0.0008188559,0.0001454429,0.0003759043,0.0005620985],"domain_scores_gemma":[0.9971081,0.00194517,0.0003859272,0.0003154627,0.00008344388,0.0001619164],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.0002992346,0.0003729354,0.3027708,0.002313603,0.00004016133,0.00001232663,0.3407155,0.00002410953,0.000004222674,0.2213646,0.08063613,0.05144635],"study_design_scores_gemma":[0.002103711,0.00002607063,0.7235239,0.0004042459,0.00002660293,0.00000319218,0.02831181,0.0001279141,0.0001333506,0.004091905,0.2410047,0.0002425292],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9092447,0.00006758374,0.00009412277,0.003802999,0.0006619741,0.002264044,0.00003143077,0.00002991167,0.08380324],"genre_scores_gemma":[0.9934627,0.00004839137,0.0001298692,0.002698774,0.00007412997,0.001497944,0.00004803611,0.000006749863,0.002033364],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4207531,"threshold_uncertainty_score":0.9993713,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2073015584","doi":"10.1016/s0895-4356(03)00246-4","title":"Validation of diagnostic codes within medical services claims","year":2004,"lang":"en","type":"article","venue":"Journal of Clinical Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":443,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"Medical Research Council","keywords":"Medicine; Medical diagnosis; Gold standard (test); Diagnosis code; Comorbidity; Medical record; COPD; Diabetes mellitus; Disease; Emergency medicine; Intensive care medicine; Internal medicine; Pathology; Environmental health","retraction":null,"screen_n_in":null,"score":{"opus":0.584328309000388,"gpt":0.639146006923215,"spread":0.05481769792282698,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.07874444,0.0001160182,0.001911865,0.0001227122,0.0001763337,0.000001267684,0.0003591402,0.0009523748,0.0005617896],"category_scores_gemma":[0.304565,0.00007550346,0.0002725606,0.0001339886,0.0002632299,0.0001624215,0.00006882955,0.002544268,0.0001299074],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006501274,"about_ca_system_score_gemma":0.001775767,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001500075,"about_ca_topic_score_gemma":0.00005543696,"domain_scores_codex":[0.9815,0.00662581,0.01077463,0.0001369706,0.000564768,0.0003978071],"domain_scores_gemma":[0.8289063,0.1585486,0.01050453,0.0002636459,0.0008024225,0.0009745488],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0006013223,0.0003124755,0.9322724,0.001880385,0.000110926,0.0000237599,0.00228386,0.001832569,0.000005322401,0.04457037,0.004935252,0.01117134],"study_design_scores_gemma":[0.01108516,0.003739743,0.6839794,0.009243315,0.0002032539,0.0001104352,0.002178965,0.003489742,0.00006249627,0.2655158,0.02009245,0.0002992796],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9379686,0.000251323,0.01911717,0.03740977,0.004108391,0.0002065366,0.000005068872,0.00001895545,0.0009142158],"genre_scores_gemma":[0.9705896,0.001949499,0.005575176,0.01992175,0.001907331,0.000006123279,0.00001159684,0.000008853902,0.00003002635],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2482931,"threshold_uncertainty_score":0.9997569,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1988984458","doi":"10.1097/01.mlr.0000132391.59713.0d","title":"Validity of Procedure Codes in International Classification of Diseases, 9th revision, Clinical Modification Administrative Data","year":2004,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":366,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Alberta Health Services; University of Calgary","funders":"","keywords":"Coding (social sciences); Data quality; Medicine; Computer science; Data mining; Statistics; Operations management; Mathematics; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.6237739865793028,"gpt":0.6062330047988461,"spread":0.01754098178045671,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.002669344,0.00009844525,0.0003612877,0.000102445,0.0001126937,0.000002918047,0.0005976374,0.0003636842,0.0003427419],"category_scores_gemma":[0.01383583,0.00007931457,0.00004506765,0.000229179,0.0001983419,0.000188173,0.0001488756,0.0007323469,0.00002692613],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001825027,"about_ca_system_score_gemma":0.002797302,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001252801,"about_ca_topic_score_gemma":0.0002595016,"domain_scores_codex":[0.9961511,0.0005176076,0.001805059,0.0002821385,0.001042914,0.0002011063],"domain_scores_gemma":[0.9970922,0.0007723064,0.000734464,0.0005200387,0.0005771123,0.0003038858],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0004629845,0.0007316229,0.8951862,0.005580852,0.00002631017,0.000005765178,0.01345875,0.00001812375,0.00001544393,0.01477926,0.01433082,0.05540388],"study_design_scores_gemma":[0.003078236,0.000286363,0.9532402,0.00328571,0.00004095878,0.00000128066,0.01263135,0.005717394,0.00002100687,0.001124057,0.02042073,0.0001527454],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9418654,0.0006925642,0.02034076,0.02420816,0.002790533,0.002316368,0.001551373,0.000116197,0.006118578],"genre_scores_gemma":[0.9953417,0.0004782687,0.0007630064,0.0005177607,0.0003070895,0.000037952,0.002517298,0.00000726584,0.00002969747],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.058054,"threshold_uncertainty_score":0.9944711,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2036378305","doi":"10.1097/00005650-200502000-00012","title":"Comparison of Coding of Heart Failure and Comorbidities in Administrative and Clinical Data for Use in Outcomes Research","year":2005,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":361,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Université de Montréal; Montreal Heart Institute; Sunnybrook Health Science Centre; Heart and Stroke Foundation; University of Toronto; Institute for Clinical Evaluative Sciences","funders":"","keywords":"Heart failure; Coding (social sciences); Medicine; MEDLINE; Intensive care medicine; Emergency medicine; Internal medicine; Statistics; Mathematics; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.835869500357248,"gpt":0.7076753544341328,"spread":0.1281941459231153,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.005342675,0.00006419997,0.0005040606,0.0001466692,0.0001256625,0.000003918844,0.0001550887,0.0002870071,0.00008125373],"category_scores_gemma":[0.009339699,0.000050081,0.00001618624,0.0001186105,0.0003299943,0.0001442152,0.000208206,0.001032637,0.000001410072],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000463217,"about_ca_system_score_gemma":0.00049749,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004860708,"about_ca_topic_score_gemma":0.0165909,"domain_scores_codex":[0.9971903,0.0006623389,0.001183511,0.0001677797,0.0005189027,0.0002771576],"domain_scores_gemma":[0.9921731,0.007062017,0.0001454571,0.0002210406,0.0001805025,0.0002178501],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00007797728,0.00005051884,0.97473,0.001243665,0.000003091854,5.668369e-7,0.008103047,1.814994e-7,4.879884e-7,0.000598729,0.01048503,0.004706729],"study_design_scores_gemma":[0.003166743,0.0004898831,0.8695498,0.001910395,0.00000802597,6.361852e-7,0.08440755,0.009658325,0.00001087722,0.00006515338,0.03064734,0.0000852368],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.987998,0.0003278339,0.00009801691,0.01048872,0.0001034632,0.0006580562,0.0001581663,0.00000740494,0.0001603473],"genre_scores_gemma":[0.9970414,0.0001173986,0.002128438,0.000447425,0.00008337466,0.00002622397,0.0001317707,0.000004606694,0.00001937077],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1051801,"threshold_uncertainty_score":0.9990051,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2018983979","doi":"10.1186/1472-6963-8-12","title":"Risk adjustment performance of Charlson and Elixhauser comorbidities in ICD-9 and ICD-10 administrative databases","year":2008,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":345,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"South Health Campus; Provincial Health Services Authority; University of Calgary; Alberta Health Services","funders":"Canadian Institutes of Health Research; Health Canada; Ministry of Health, British Columbia; Fondation pour la Recherche Médicale","keywords":"Medicine; ICD-10; Health administration; Health informatics; Nursing research; Quality of Life Research; Database; Emergency medicine; Public health; Medical emergency; Nursing","retraction":null,"screen_n_in":null,"score":{"opus":0.4268680489924949,"gpt":0.5499640113109536,"spread":0.1230959623184588,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.005199576,0.0001473294,0.0004267736,0.0003589968,0.001406087,0.000006225835,0.0001642758,0.0001367887,0.0004106053],"category_scores_gemma":[0.0001647579,0.0001226509,0.0000144466,0.0003846042,0.0002603059,0.0003489933,0.0002264983,0.001162551,0.00009612233],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001765624,"about_ca_system_score_gemma":0.001779067,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007578716,"about_ca_topic_score_gemma":0.004958704,"domain_scores_codex":[0.995704,0.001299581,0.001030428,0.00029012,0.0007917053,0.0008841408],"domain_scores_gemma":[0.9968353,0.001646001,0.0003752984,0.0003136145,0.0003244781,0.0005053225],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0005814073,0.00009569011,0.9301374,0.02243058,0.000006243707,0.00000173188,0.04125906,0.000003545542,0.000004655627,0.0004231382,0.001461962,0.003594592],"study_design_scores_gemma":[0.001235159,0.0007705639,0.9625187,0.001907912,0.000002332671,0.000004203072,0.01815242,0.005412908,0.00002619857,0.00002353453,0.009848759,0.00009731536],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9935021,0.002876312,0.00001307839,0.0005449053,0.0001128416,0.001298455,0.0001477801,0.00002869337,0.001475812],"genre_scores_gemma":[0.9676799,0.02953464,0.0005743279,0.0009267433,0.0001678775,0.0002187921,0.0001020632,0.00001399748,0.0007817106],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0323813,"threshold_uncertainty_score":0.999894,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2807046356","doi":"10.1186/s12913-018-3148-0","title":"Identifying diabetes cases from administrative data: a population-based validation study","year":2018,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":279,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"St. Michael's Hospital; Health Sciences Centre; Sunnybrook Health Science Centre; Eli Lilly (Canada); Institute for Clinical Evaluative Sciences; Public Health Ontario; Women's College Hospital; University Health Network; University of Toronto","funders":"Canadian Institutes of Health Research; Connaught Fund; University of Toronto; Banting and Best Diabetes Centre, University of Toronto; Institute for Clinical Evaluative Sciences","keywords":"Medicine; Health informatics; Health administration; Diabetes mellitus; Nursing research; Diagnosis code; Medical record; Health care; Medical prescription; Family medicine; Public health; Population; MEDLINE; Health services research; Nursing; Environmental health; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.7183962041045853,"gpt":0.6513766667406515,"spread":0.06701953736393373,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.01220454,0.0001823236,0.0004052675,0.0004111898,0.004489538,0.0001001852,0.0008175456,0.000204176,0.001137425],"category_scores_gemma":[0.0009659084,0.0001598468,0.00002556233,0.0009070306,0.00008114751,0.0006990107,0.0004188087,0.00111109,0.001347607],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005076556,"about_ca_system_score_gemma":0.003309957,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.03137476,"about_ca_topic_score_gemma":0.02971227,"domain_scores_codex":[0.9901062,0.004871374,0.001478297,0.0006349062,0.001665597,0.001243634],"domain_scores_gemma":[0.9908422,0.005699197,0.0005603651,0.001258952,0.0009706827,0.0006686628],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001505257,0.0001859643,0.9767916,0.00290593,0.00001395098,0.000001706167,0.01699383,0.000002545503,0.000004555831,0.00006780503,0.0008981649,0.00198344],"study_design_scores_gemma":[0.001292778,0.0009705953,0.923399,0.001469795,0.00001024909,1.450974e-7,0.04415919,0.02488659,0.00001984827,0.0003273908,0.003306766,0.0001576501],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9931533,0.0001470053,0.0006704528,0.0009704964,0.0007603738,0.003241566,0.0003320041,0.0001698747,0.0005549798],"genre_scores_gemma":[0.9903743,0.00001869052,0.001874194,0.002734031,0.001556493,0.0003756504,0.002904603,0.00002783455,0.0001341743],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.05339258,"threshold_uncertainty_score":0.9997756,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2413622533","doi":"10.1093/ije/dyw075","title":"Maelstrom Research guidelines for rigorous retrospective data harmonization","year":2016,"lang":"en","type":"article","venue":"International Journal of Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":272,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Ontario Institute for Cancer Research; McMaster University; McGill University Health Centre","funders":"Economic and Social Research Council; Medical Research Council; European Commission; Government of Ontario; Wellcome Trust; Partenariat Canadien Contre Le Cancer; University of Bristol; Canadian Institutes of Health Research; Genome Canada","keywords":"Harmonization; Standardization; Computer science; Interdependence; Data science; Data quality; Management science; Inefficiency; Process management; Service (business); Business; Political science; Engineering; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.9102577888579144,"gpt":0.7132176123109125,"spread":0.1970401765470019,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.03033094,0.00008121627,0.0003626726,0.00031389,0.0002750891,0.000004185789,0.0009659494,0.0002206451,0.000596109],"category_scores_gemma":[0.1792491,0.00004712503,0.00005947803,0.00008897006,0.0001303517,0.0003906416,0.0002222756,0.0005747786,0.0001500678],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004885821,"about_ca_system_score_gemma":0.0006249119,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006635353,"about_ca_topic_score_gemma":0.00002889563,"domain_scores_codex":[0.99532,0.001420464,0.002143316,0.0001914918,0.0004908575,0.0004339329],"domain_scores_gemma":[0.9756521,0.01252297,0.001421884,0.0003261691,0.00988039,0.0001964914],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004942121,0.00002684677,0.0770784,0.00003666873,0.00009269299,0.000004238141,0.0002424816,0.000008147736,0.0001670299,0.04130189,0.812304,0.06824335],"study_design_scores_gemma":[0.002368683,0.0003840306,0.04159765,0.000830777,0.00001592441,0.00004069824,0.0003140806,0.002649163,0.00001941739,0.0977874,0.8538994,0.00009276628],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01370608,0.0004455184,0.6469598,0.3282716,0.007878886,0.0004711904,0.0003243943,0.00002359356,0.001918891],"genre_scores_gemma":[0.7678973,0.006158649,0.1502759,0.04076356,0.02788717,0.0001113597,0.0004163007,0.00006128704,0.006428486],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7541912,"threshold_uncertainty_score":0.9984784,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2044184365","doi":"10.1097/mlr.0b013e3181484347","title":"Cross-National Comparative Performance of Three Versions of the ICD-10 Charlson Index","year":2007,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":258,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Calgary","funders":"","keywords":"Charlson comorbidity index; Medicine; Logistic regression; Statistics; Comorbidity; Receiver operating characteristic; Statistic; Diagnosis code; Index (typography); Population; Demography; Internal medicine; Mathematics; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.2327585030962916,"gpt":0.5044259325204773,"spread":0.2716674294241858,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001873646,0.0000769883,0.0002174833,0.00006713831,0.0004968541,0.000001339174,0.0002754956,0.0002708541,0.003429961],"category_scores_gemma":[0.001067668,0.00004856921,0.00005572158,0.0002474853,0.0002738161,0.00007377125,0.0001134348,0.0007500746,0.00008990034],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001453713,"about_ca_system_score_gemma":0.00106398,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009645991,"about_ca_topic_score_gemma":0.0002886298,"domain_scores_codex":[0.9973146,0.00009609874,0.0007310872,0.00009262194,0.001481089,0.0002844615],"domain_scores_gemma":[0.9978597,0.0006514288,0.0003333153,0.000163794,0.0007745309,0.0002172499],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002341723,0.0000433904,0.9598707,0.001576917,0.00001306141,4.777094e-7,0.01949416,0.00002832921,0.000009828017,0.002246591,0.01060407,0.005878292],"study_design_scores_gemma":[0.00103897,0.0001114289,0.9634071,0.000691379,0.000006878529,6.393108e-7,0.003021246,0.003311584,0.0001682459,0.00004917045,0.02813106,0.00006231782],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9562225,0.0001067953,0.0004715931,0.001031587,0.000687068,0.0003300565,0.00003478504,0.00002003502,0.04109558],"genre_scores_gemma":[0.9981896,0.00001264383,0.00004390291,0.001075551,0.0002196213,0.0000128631,0.00003158905,0.000004099148,0.0004100755],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04196715,"threshold_uncertainty_score":0.997481,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2168376827","doi":"10.1370/afm.1644","title":"Validating the 8 CPCSSN Case Definitions for Chronic Disease Surveillance in a Primary Care Database of Electronic Health Records","year":2014,"lang":"en","type":"article","venue":"The Annals of Family Medicine","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":257,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Dalhousie University; University of Alberta; University of Calgary; University of British Columbia; Queen's University","funders":"","keywords":"Medicine; Parkinsonism; Dementia; Epilepsy; Diabetes mellitus; COPD; Public health; Disease; Database; Family medicine; Psychiatry; Internal medicine; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.5731373414421918,"gpt":0.5304380823746074,"spread":0.04269925906758432,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008538462,0.0001257634,0.0004458902,0.0001079642,0.0006121965,0.000001462652,0.000248832,0.00005729438,0.00002969274],"category_scores_gemma":[0.002499919,0.00006941079,0.00005390129,0.0003137746,0.0001908968,0.00006711434,0.00006229428,0.0005651012,0.000004916713],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001756115,"about_ca_system_score_gemma":0.001725313,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002317324,"about_ca_topic_score_gemma":0.001367063,"domain_scores_codex":[0.9967275,0.001001766,0.001177539,0.0001505597,0.0003378031,0.0006048168],"domain_scores_gemma":[0.9948027,0.003335071,0.0007573817,0.0005749117,0.0003422235,0.0001877341],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.004241789,0.0004852297,0.03864949,0.1201612,0.0001939069,0.00002326023,0.1150657,0.000974734,0.0005014875,0.1223361,0.2786057,0.3187614],"study_design_scores_gemma":[0.02996318,0.01297631,0.417908,0.05578006,0.0003212085,0.00006753403,0.1075743,0.05574073,0.0001105209,0.02860722,0.2895448,0.001406188],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8422477,0.02006479,0.008243288,0.1179313,0.0007976545,0.004208551,0.0007077791,0.00007312911,0.005725791],"genre_scores_gemma":[0.9777439,0.004242511,0.0001002798,0.01674361,0.0004528413,0.0002329186,0.0004366976,0.00001452696,0.00003268724],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3792585,"threshold_uncertainty_score":0.4708582,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2005159922","doi":"10.1016/j.jclinepi.2011.08.002","title":"Administrative database research has unique characteristics that can risk biased results","year":2011,"lang":"en","type":"review","venue":"Journal of Clinical Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":235,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Institute for Clinical Evaluative Sciences; University of Toronto; Ottawa Hospital; University of Ottawa","funders":"","keywords":"Observational study; Database; Computer science; MEDLINE; Health care; Medical prescription; Data science; Medicine; Data mining; Nursing; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.978490436321102,"gpt":0.764329112763572,"spread":0.21416132355753,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","research_integrity"],"consensus_categories":["metaresearch","research_integrity"],"category_scores_codex":[0.3706848,0.0005047735,0.0119937,0.0006653999,0.001100745,0.00001180597,0.001274529,0.003538831,0.0006433549],"category_scores_gemma":[0.675068,0.0003041395,0.001465123,0.0004193241,0.001096483,0.0001509043,0.0003461154,0.02712127,0.0006542752],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000355711,"about_ca_system_score_gemma":0.01779042,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004242925,"about_ca_topic_score_gemma":0.0001804128,"domain_scores_codex":[0.8311744,0.1394637,0.02678469,0.0005556643,0.0006775964,0.001343938],"domain_scores_gemma":[0.5430184,0.4051441,0.04590332,0.001404456,0.002249218,0.002280554],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001058859,0.0001826942,0.003839236,0.01434188,0.0002596944,0.00008840746,0.0006511527,8.420653e-8,2.128538e-9,0.002971765,0.2358226,0.7407836],"study_design_scores_gemma":[0.001105939,0.001555861,0.002977665,0.03037403,0.0003704326,0.00004687908,0.0001802109,0.00003129755,1.120157e-8,0.001596237,0.9615468,0.0002145854],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0001543408,0.9594496,0.002580039,0.01343605,0.01084158,0.002844157,0.00348707,0.00005399598,0.007153201],"genre_scores_gemma":[0.0001406102,0.9861513,0.003441185,0.003291104,0.005488936,0.00006221623,0.0005905774,0.00004769896,0.0007863054],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.740569,"threshold_uncertainty_score":0.9999411,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1966516148","doi":"10.1111/j.1528-1167.2009.02201.x","title":"How accurate is ICD coding for epilepsy?","year":2009,"lang":"en","type":"article","venue":"Epilepsia","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":227,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"Canadian Institutes of Health Research","keywords":"Epilepsy; Medicine; ICD-10; Medical record; Pediatrics; Internal medicine; Psychiatry","retraction":null,"screen_n_in":null,"score":{"opus":0.2505814439391201,"gpt":0.4659326877107203,"spread":0.2153512437716002,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001185652,0.0001553846,0.0003026287,0.0001038258,0.001133619,0.00003172748,0.0001865168,0.0002740687,0.0008674445],"category_scores_gemma":[0.000904446,0.0001309228,0.00008684692,0.0001697684,0.0000259187,0.000317828,0.00002409619,0.0005751171,0.0004984412],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001239941,"about_ca_system_score_gemma":0.0002200391,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009972809,"about_ca_topic_score_gemma":0.000002957116,"domain_scores_codex":[0.9980656,0.0001317738,0.0005882421,0.0002169833,0.0002622524,0.0007351744],"domain_scores_gemma":[0.9983042,0.0005810801,0.0003285641,0.0002896235,0.0001882873,0.0003082446],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001180979,0.00002771255,0.003025419,0.0004672164,0.000009254427,0.00000132642,0.003890873,0.000001805915,0.0001085775,0.03894253,0.9114863,0.04192092],"study_design_scores_gemma":[0.001692285,0.0003175599,0.02770732,0.0004065355,0.00002048925,0.000001550536,0.001214754,0.007879675,0.0001106405,0.00749955,0.9528943,0.0002553989],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.07605797,0.0004520926,0.3202405,0.4133674,0.007242972,0.006585263,0.0002645873,0.001313118,0.1744761],"genre_scores_gemma":[0.9349098,0.0001722958,0.003654344,0.04516672,0.001391418,0.0001919365,0.00009005249,0.0000199483,0.01440352],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8588518,"threshold_uncertainty_score":0.9497909,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2036668944","doi":"10.1016/j.jclinepi.2014.02.019","title":"Validation of international algorithms to identify adults with inflammatory bowel disease in health administrative data from Ontario, Canada","year":2014,"lang":"en","type":"article","venue":"Journal of Clinical Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":221,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Western University; McMaster University; Mount Sinai Hospital; University of Ottawa; University of Toronto; Children's Hospital of Eastern Ontario; Institute for Clinical Evaluative Sciences; Ottawa Hospital; Canada Research Chairs","funders":"Canadian Institutes of Health Research; Ontario Ministry of Health and Long-Term Care; Canadian Child Health Clinician Scientist Program; Institute for Clinical Evaluative Sciences; American College of Gastroenterology","keywords":"Medicine; Inflammatory bowel disease; Algorithm; Incidence (geometry); Ulcerative colitis; Disease; Crohn's disease; Cohort; Health care; Internal medicine; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.5059050188247863,"gpt":0.6000459891503751,"spread":0.09414097032558877,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.02931539,0.0001166142,0.001161501,0.0001185343,0.000094779,0.000002282932,0.0005636775,0.000187267,0.0004098555],"category_scores_gemma":[0.06383207,0.00008395237,0.00006237917,0.00007486752,0.00007366927,0.0002216069,0.0001296522,0.001527322,0.00001343498],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004055837,"about_ca_system_score_gemma":0.0106275,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.5413673,"about_ca_topic_score_gemma":0.8129163,"domain_scores_codex":[0.9868208,0.005202078,0.006937145,0.0002379648,0.000458053,0.000343933],"domain_scores_gemma":[0.9692108,0.02260651,0.006209773,0.0004380502,0.0004990716,0.001035771],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001781025,0.00006196305,0.9715067,0.0001815129,0.00003669331,0.00000668374,0.0008423589,0.0004457592,1.6599e-7,0.0002963423,0.02189143,0.002949411],"study_design_scores_gemma":[0.001851408,0.0004774276,0.9574386,0.001514206,0.00001159218,0.000001076748,0.0003697321,0.006412035,3.813926e-7,0.0007977901,0.03105788,0.00006789337],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9420113,0.00004575258,0.01529686,0.03906675,0.002844319,0.0003584413,0.0001296519,0.000005444945,0.0002415073],"genre_scores_gemma":[0.9765614,0.0001481527,0.007152515,0.01472038,0.001061505,0.000009164302,0.0002602009,0.000008112279,0.00007851356],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.271549,"threshold_uncertainty_score":0.9995241,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2171602507","doi":"10.1111/j.1475-6773.2011.01290.x","title":"International Comparability of Patient Safety Indicators in 15 OECD Member Countries: A Methodological Approach of Adjustment by Secondary Diagnoses","year":2011,"lang":"en","type":"article","venue":"Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":205,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Division of Chemistry; Universiteit van Amsterdam; European Commission; Commonwealth Fund","keywords":"Comparability; Medicine; Patient safety; Proxy (statistics); International comparisons; Medical diagnosis; Demography; Health care; Data quality; Data collection; Statistics; Operations management","retraction":null,"screen_n_in":null,"score":{"opus":0.4991142863975995,"gpt":0.5598804464066812,"spread":0.06076616000908175,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.01446141,0.0001431298,0.0006605632,0.0006408637,0.0003414973,0.000003056316,0.0004892012,0.0002956718,0.004554309],"category_scores_gemma":[0.0003061398,0.0001116174,0.0000495979,0.0005878577,0.0002734736,0.0001655934,0.0003355063,0.001522475,0.000065374],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005529309,"about_ca_system_score_gemma":0.001325124,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008727355,"about_ca_topic_score_gemma":0.0004480654,"domain_scores_codex":[0.9907899,0.004487495,0.002219359,0.0003277409,0.001360301,0.0008151879],"domain_scores_gemma":[0.9946101,0.003109132,0.001009304,0.0003834539,0.0004573292,0.0004306735],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.003420715,0.002332966,0.6132293,0.04115413,0.00007287813,0.000001927779,0.2744252,0.000017568,0.000009728192,0.008560946,0.0116755,0.04509912],"study_design_scores_gemma":[0.00347332,0.001566897,0.746997,0.001838934,0.000007436038,0.000001574521,0.07776845,0.002489493,0.0002324745,0.001063268,0.164306,0.0002551051],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9356086,0.001802202,0.0003688598,0.001730841,0.0005900408,0.003434865,0.0004231814,0.0000486789,0.0559927],"genre_scores_gemma":[0.9898604,0.002575272,0.004288265,0.002494532,0.00006479823,0.0004219962,0.0002341242,0.00001265187,0.00004800074],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1966567,"threshold_uncertainty_score":0.9978736,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2125287424","doi":"10.1212/wnl.0b013e3182684707","title":"Recommendations for optimal ICD codes to study neurologic conditions","year":2012,"lang":"en","type":"review","venue":"Neurology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":202,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Canadian Institutes of Health Research; Alberta Innovates; Public Health Agency; Public Health Agency of Canada","keywords":"Medicine; Internal medicine; Diagnosis code; Parkinsonism; Dementia; Population; MEDLINE; Disease","retraction":null,"screen_n_in":null,"score":{"opus":0.6120930994399504,"gpt":0.5846588603050891,"spread":0.02743423913486132,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.001723149,0.0003513162,0.001415381,0.0003956768,0.001473413,0.00000879201,0.0003916156,0.0007836858,0.001567977],"category_scores_gemma":[0.0007810114,0.0002839905,0.0001919115,0.0003047096,0.00005029942,0.00008476373,0.0002040429,0.001986783,0.002325897],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005066137,"about_ca_system_score_gemma":0.0004458576,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003515737,"about_ca_topic_score_gemma":0.00005011172,"domain_scores_codex":[0.9946069,0.0022031,0.001602904,0.0004293823,0.0001693433,0.000988418],"domain_scores_gemma":[0.994243,0.003904353,0.0007133413,0.0005030434,0.0001636718,0.0004725823],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000206346,0.0006829568,0.0002914668,0.01903405,0.0001378072,0.000008604934,0.002785449,0.00001041101,4.818217e-8,0.003640816,0.1180231,0.8551789],"study_design_scores_gemma":[0.0004322439,0.002030942,0.0001481329,0.0002195782,0.0004981816,0.00001238133,0.0000519516,0.00002521864,4.189322e-9,0.00005853281,0.9963387,0.0001841809],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0006309541,0.9077896,0.004636662,0.03111406,0.01577311,0.03335784,0.002036043,0.0008835345,0.003778132],"genre_scores_gemma":[0.0004577691,0.9418337,0.0008089339,0.03675609,0.001977362,0.01577744,0.001669221,0.000104155,0.000615266],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.8783155,"threshold_uncertainty_score":0.9999612,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2970976623","doi":"10.1097/mlr.0000000000000905","title":"Validation of the Combined Comorbidity Index of Charlson and Elixhauser to Predict 30-Day Mortality Across ICD-9 and ICD-10","year":2018,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":197,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Institut National d'Excellence en Santé et en Services Sociaux; Université Laval; Institut National de Santé Publique du Québec","funders":"","keywords":"Medicine; Logistic regression; Charlson comorbidity index; ICD-10; Confidence interval; Comorbidity; Index (typography); Cohort; Population; Cohort study; Statistics; Demography; Emergency medicine; Internal medicine; Environmental health; Psychiatry; Mathematics; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.1313195582417165,"gpt":0.465341768028919,"spread":0.3340222097872025,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002487438,0.0001044031,0.0003012601,0.0000336466,0.0004648943,0.000004705349,0.0001792807,0.0003241784,0.001081993],"category_scores_gemma":[0.002634666,0.00006849262,0.00002800077,0.0001691324,0.0003704562,0.00006923838,0.0002653398,0.0004800922,0.00002181034],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005274342,"about_ca_system_score_gemma":0.0003153133,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004976483,"about_ca_topic_score_gemma":0.0003730534,"domain_scores_codex":[0.9975816,0.0003695198,0.0007095467,0.0001641697,0.0008688383,0.00030631],"domain_scores_gemma":[0.9982198,0.0003419963,0.0002816755,0.0003019034,0.0004459091,0.0004086731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00007293475,0.00003310622,0.9395317,0.002162943,0.00001527143,4.350125e-7,0.03824397,4.708623e-7,0.00004152377,0.0001972118,0.009874873,0.00982554],"study_design_scores_gemma":[0.001146177,0.0002738432,0.9863364,0.0006853506,0.00001889445,3.637147e-7,0.002986989,0.0006503808,0.0005271333,0.00009021311,0.007209864,0.00007439788],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9942743,0.00007084108,0.0003502163,0.002672485,0.0006328548,0.000639928,0.0001057135,0.00002993057,0.001223709],"genre_scores_gemma":[0.9976993,0.00004135433,0.00001101626,0.00179553,0.0002694637,0.00003505492,0.00003504621,0.000007724391,0.000105468],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04680468,"threshold_uncertainty_score":0.9998311,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W203757376","doi":"","title":"Evaluating causes of death and morbidity in Iran, global burden of diseases, injuries, and risk factors study 2010.","year":2014,"lang":"en","type":"article","venue":"PubMed","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":194,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Life expectancy; Medicine; Years of potential life lost; Disease; Cause of death; Burden of disease; Pediatrics; Demography; Injury prevention; Poison control; Environmental health; Population; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.3093862088744037,"gpt":0.4648304943166247,"spread":0.155444285442221,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002510753,0.00008337107,0.0002928083,0.00006162941,0.0001738909,0.000003458037,0.00006060158,0.00008478104,0.00001267295],"category_scores_gemma":[0.003818673,0.00006484412,0.00001268568,0.0001140394,0.00005506461,0.00009442395,0.00008327594,0.0002248946,7.703954e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006321612,"about_ca_system_score_gemma":0.00007538309,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01603529,"about_ca_topic_score_gemma":0.000875393,"domain_scores_codex":[0.9980074,0.0006762152,0.000626431,0.0001257371,0.0002861965,0.0002780861],"domain_scores_gemma":[0.998542,0.0006492903,0.0003838262,0.0001464596,0.00007029404,0.0002081015],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00005299725,0.00003883758,0.9774607,0.0004260356,0.000009485636,6.749993e-8,0.003285319,0.000008963366,2.545467e-7,0.00008899693,0.0001255353,0.01850281],"study_design_scores_gemma":[0.00107675,0.0001056077,0.9951308,0.00005786843,0.00004525184,5.461092e-8,0.00210739,0.0009468295,0.000001457584,0.0003728289,0.0001031881,0.00005198325],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9978319,0.0001782306,0.00002654636,0.00009437089,0.0001916041,0.00125735,0.0000899322,0.00001789535,0.0003121518],"genre_scores_gemma":[0.9994695,0.0001159249,0.0000432102,0.00003521337,0.00006388194,0.00022877,0.000008934448,0.000004034478,0.00003055006],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01845083,"threshold_uncertainty_score":0.990517,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1981830415","doi":"10.1016/j.thromres.2005.06.015","title":"Accuracy of coding for possible warfarin complications in hospital discharge abstracts","year":2005,"lang":"en","type":"article","venue":"Thrombosis Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":193,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa; Health Canada","funders":"","keywords":"Medicine; Diagnosis code; Gold standard (test); Positive predicative value; Medical diagnosis; Medical record; ICD-10; Internal medicine; Predictive value of tests; Predictive value; Population; Radiology; Psychiatry","retraction":null,"screen_n_in":null,"score":{"opus":0.5391367357924423,"gpt":0.5961498095873871,"spread":0.0570130737949448,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004539221,0.00008355314,0.000251256,0.0003637354,0.0007996982,0.00001145825,0.0002944449,0.0001701536,0.0004636101],"category_scores_gemma":[0.00234669,0.00007337093,0.00004334791,0.000472233,0.0001097278,0.0002984776,0.0001172786,0.0009285978,0.0004419367],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002844441,"about_ca_system_score_gemma":0.0004480986,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006645946,"about_ca_topic_score_gemma":0.0003156103,"domain_scores_codex":[0.9973294,0.0002639354,0.0009286773,0.0001827442,0.0005249656,0.0007702875],"domain_scores_gemma":[0.9959507,0.002824319,0.0002264478,0.0003219778,0.0004754886,0.0002010561],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"observational","study_design_scores_codex":[0.0003435505,0.003422782,0.1646444,0.01006144,0.00009282764,9.303305e-7,0.0956404,0.0003640304,0.01062018,0.2600869,0.3250272,0.1296954],"study_design_scores_gemma":[0.002426659,0.0003880684,0.9070057,0.001264759,0.000007028188,2.870255e-7,0.005299027,0.01121038,0.001185761,0.003094222,0.06788949,0.0002285993],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9191901,0.00009491551,0.001207345,0.03814873,0.0002045582,0.003066039,0.00008518958,0.0000663925,0.03793672],"genre_scores_gemma":[0.9964669,0.0002439167,0.001844374,0.0002588754,0.0001945665,0.0005362856,0.00007548788,0.00001414719,0.0003654376],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7423614,"threshold_uncertainty_score":0.6150713,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1993370148","doi":"10.1016/j.jclinepi.2011.01.001","title":"Administrative database research infrequently used validated diagnostic or procedural codes","year":2011,"lang":"en","type":"review","venue":"Journal of Clinical Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":176,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital; Institute for Clinical Evaluative Sciences","funders":"","keywords":"Diagnosis code; Medicine; Confidence interval; Medical diagnosis; Code (set theory); MEDLINE; Current Procedural Terminology; Statistics; Data mining; Computer science; Internal medicine; Surgery; Mathematics; Pathology; Population; Environmental health; Programming language; Set (abstract data type)","retraction":null,"screen_n_in":null,"score":{"opus":0.970072185892032,"gpt":0.7783446361477604,"spread":0.1917275497442716,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","research_integrity","insufficient_payload"],"category_scores_codex":[0.2361836,0.0005096289,0.0121994,0.0006287952,0.0006692589,0.000009658605,0.001316268,0.002921479,0.0029699],"category_scores_gemma":[0.7962933,0.0002682489,0.001242981,0.0006395396,0.0008805885,0.0003008501,0.0003335288,0.01757989,0.001543173],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003193962,"about_ca_system_score_gemma":0.02020565,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009309265,"about_ca_topic_score_gemma":0.00009716341,"domain_scores_codex":[0.8794377,0.0909698,0.02684036,0.0005195562,0.0007584637,0.001474139],"domain_scores_gemma":[0.3586034,0.6105151,0.02519116,0.0009753736,0.002608476,0.002106482],"domain_codex":null,"domain_gemma":"methods","domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001082222,0.0004429865,0.005376616,0.06577522,0.0005525716,0.0002364853,0.0008242571,3.741878e-7,1.713066e-8,0.01077904,0.4449058,0.4700244],"study_design_scores_gemma":[0.001036656,0.002203903,0.0005293027,0.04692689,0.0004009902,0.0001263449,0.0002438445,0.00001646469,2.534824e-8,0.002082781,0.9462131,0.0002196795],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0001415334,0.9816145,0.001289673,0.004390487,0.006206604,0.002899852,0.0002734094,0.00004854648,0.003135397],"genre_scores_gemma":[0.0000628055,0.9867932,0.003340689,0.00414266,0.004174953,0.0001756542,0.0002503409,0.00004931039,0.001010419],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.5601096,"threshold_uncertainty_score":0.999977,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2318106356","doi":"10.1097/mlr.0b013e3181ef9d3e","title":"The Development, Evolution, and Modifications of ICD-10","year":2010,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":171,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Canadian Institutes of Health Research; Alberta Health Services","funders":"Canadian Institutes of Health Research","keywords":"Comparability; ICD-10; Context (archaeology); Medicine; Stakeholder; Political science; Geography; Public relations; Psychiatry","retraction":null,"screen_n_in":null,"score":{"opus":0.1255139691831797,"gpt":0.4509560587752253,"spread":0.3254420895920456,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001259403,0.00005222153,0.00009696164,0.00003334095,0.001243126,0.000003213964,0.0001508998,0.0002130555,0.000927862],"category_scores_gemma":[0.003209049,0.00003187698,0.0000132497,0.00009884358,0.0001724595,0.00003602896,0.00005975991,0.000729246,0.0001350107],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003971047,"about_ca_system_score_gemma":0.001200426,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004799793,"about_ca_topic_score_gemma":0.0009941069,"domain_scores_codex":[0.9986303,0.00009878221,0.0005129161,0.0000798031,0.0004516446,0.0002265393],"domain_scores_gemma":[0.998444,0.000653229,0.0001313801,0.0001783059,0.00030487,0.0002881979],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00008940238,0.0000657566,0.1545093,0.003049319,0.00003200865,0.000001225011,0.09742006,7.699544e-7,0.0001555623,0.2201254,0.1469733,0.3775779],"study_design_scores_gemma":[0.0005001036,0.00002320197,0.1572994,0.0001446686,0.000006486442,0.000001136899,0.008688209,0.0005248213,0.00002150774,0.0005972006,0.8321218,0.00007141707],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8520093,0.001196192,0.008284086,0.02577625,0.002688117,0.001029904,0.00002225603,0.0001752347,0.1088186],"genre_scores_gemma":[0.9976228,0.00003508919,0.0005021698,0.0004752812,0.0001878646,0.0000822496,0.00002206506,0.000004521645,0.001067986],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6851485,"threshold_uncertainty_score":0.9999854,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2046869249","doi":"10.1016/j.diabres.2010.03.007","title":"Validating ICD coding algorithms for diabetes mellitus from administrative data","year":2010,"lang":"en","type":"article","venue":"Diabetes Research and Clinical Practice","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":169,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia; University of Calgary","funders":"Canadian Institutes of Health Research; Alberta Innovates","keywords":"Medicine; Diabetes mellitus; Coding (social sciences); Algorithm; Computer science; Statistics; Endocrinology; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.7586158407304545,"gpt":0.6784496236929183,"spread":0.08016621703753624,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.03151885,0.0001626307,0.0004446774,0.00009182676,0.002061064,0.0001263336,0.000646689,0.0005103543,0.0005332944],"category_scores_gemma":[0.1756029,0.0001335736,0.00004899064,0.0002397338,0.0004468648,0.001148904,0.0006357827,0.004749292,0.0002517522],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004034972,"about_ca_system_score_gemma":0.0007960666,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001389099,"about_ca_topic_score_gemma":0.00008867327,"domain_scores_codex":[0.9938936,0.001977691,0.001460217,0.0006432541,0.0008032325,0.001222044],"domain_scores_gemma":[0.871309,0.1251944,0.0005374822,0.0008991371,0.001151138,0.0009088307],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0007392784,0.001026882,0.1937646,0.001536848,0.0004638215,0.000007877889,0.002457755,1.82829e-7,0.0008206915,0.01045202,0.1852995,0.6034306],"study_design_scores_gemma":[0.002441814,0.001335184,0.01498167,0.0004227439,0.0000943825,2.400486e-7,0.005480725,0.03923388,0.0001822089,0.008200403,0.9273139,0.00031284],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9070839,0.0009013488,0.002084858,0.05080721,0.003898611,0.003805661,0.001662756,0.0002252772,0.02953032],"genre_scores_gemma":[0.9055397,0.002680541,0.07018765,0.01101637,0.006707605,0.0005896949,0.002340493,0.00006755228,0.0008703775],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7420144,"threshold_uncertainty_score":0.9992381,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1823858897","doi":"10.1186/1472-6963-6-77","title":"Identifying priorities in methodological research using ICD-9-CM and ICD-10 administrative data: report from an international consortium","year":2006,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":167,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Health Sciences Centre; University of Toronto; Canadian Institute for Health Information; Sunnybrook Health Science Centre; University of Manitoba; Institute for Clinical Evaluative Sciences; University of Calgary; Statistics Canada; Centre for Advancing Health Outcomes; Manitoba Health; University of British Columbia","funders":"Institute of Health Services and Policy Research; Canadian Institutes of Health Research; Health Canada","keywords":"Health informatics; Health services research; Public health; Nursing research; Health administration; Medicine; Documentation; Data quality; Population; Environmental health; Nursing; Business; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.8777235728205727,"gpt":0.7188583613722298,"spread":0.1588652114483429,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0576641,0.0001920949,0.0005635337,0.0008221185,0.002540704,0.0001707421,0.00102592,0.0004544307,0.0009974276],"category_scores_gemma":[0.002217558,0.0001715944,0.00002439888,0.0008386681,0.0004018788,0.001277508,0.001238534,0.003174552,0.0001366588],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008600192,"about_ca_system_score_gemma":0.004778435,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.06394432,"about_ca_topic_score_gemma":0.02648875,"domain_scores_codex":[0.9801426,0.01266185,0.002192993,0.0009352072,0.002386035,0.001681305],"domain_scores_gemma":[0.9876099,0.008913172,0.0005097003,0.001063249,0.001209223,0.0006947636],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001233283,0.0003166733,0.954626,0.007911721,0.00002516946,0.0002384842,0.02240278,0.0000232828,0.0001637674,0.005546454,0.002702612,0.00480978],"study_design_scores_gemma":[0.001391812,0.0002620559,0.8724997,0.002072579,0.000004256291,0.00002894268,0.0531367,0.03182264,0.00001853149,0.006818668,0.03171922,0.0002248739],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9858532,0.0009247351,0.001682187,0.002714304,0.000679787,0.001936649,0.0002396823,0.00009353979,0.005875892],"genre_scores_gemma":[0.9481887,0.0007251445,0.04240887,0.001395699,0.002466823,0.0002803382,0.002265242,0.00004550782,0.002223626],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.08212627,"threshold_uncertainty_score":0.9999158,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2741572302","doi":"10.1007/978-1-4939-6640-0","title":"Statistical Analysis with Measurement Error or Misclassification","year":2017,"lang":"en","type":"book","venue":"Springer series in statistics","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":146,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Statistics; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.3889613763643948,"gpt":0.4725584168177415,"spread":0.08359704045334676,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.002011987,0.0003744473,0.0009347445,0.0004229119,0.0009190284,0.00004676999,0.0003752818,0.0005694568,0.002323133],"category_scores_gemma":[0.002109586,0.0002796402,0.00004249721,0.0001632601,0.0003251254,0.0001544501,0.0001095104,0.001674343,0.0003506162],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001499126,"about_ca_system_score_gemma":0.004759775,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001904602,"about_ca_topic_score_gemma":0.01173279,"domain_scores_codex":[0.9958662,0.0002720218,0.001403637,0.0004401302,0.001332393,0.0006855601],"domain_scores_gemma":[0.9960968,0.0006614137,0.001277714,0.0009195851,0.0007305915,0.0003139168],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.002714411,0.0001410368,0.01832457,0.01363447,0.001254141,0.0002715993,0.007543877,0.00009180424,0.000001117778,0.330399,0.5979652,0.02765869],"study_design_scores_gemma":[0.001027942,0.0003430396,0.05732705,0.001863989,0.000985748,0.000002128093,0.0004922139,0.001354609,4.687845e-7,0.004302524,0.9317169,0.0005834241],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"other","genre_scores_codex":[0.00003782821,0.0001459893,0.5324335,0.001031147,0.001522425,0.002017495,0.003014063,0.0001686776,0.4596289],"genre_scores_gemma":[0.003573359,0.0006057338,0.09309883,0.0006663475,0.0006834298,0.0005266382,0.002553353,0.0001157494,0.8981766],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.4393346,"threshold_uncertainty_score":0.9999655,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1557384345","doi":"10.1186/1472-6963-6-161","title":"ICD-10 coding algorithms for defining comorbidities of acute myocardial infarction","year":2006,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":145,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Provincial Health Services Authority; University of Calgary","funders":"Canadian Institutes of Health Research; Ministry of Health, British Columbia; Fondation pour la Recherche Médicale","keywords":"Medicine; Comorbidity; ICD-10; Myocardial infarction; Algorithm; Diagnosis code; Heart failure; Health informatics; Internal medicine; Emergency medicine; Intensive care medicine; Public health; Population; Pathology; Psychiatry","retraction":null,"screen_n_in":null,"score":{"opus":0.2625953014050725,"gpt":0.5358839472520113,"spread":0.2732886458469387,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.009681874,0.0001586322,0.000504576,0.0005047243,0.002581919,0.0000248123,0.0003023884,0.0003119909,0.0002939493],"category_scores_gemma":[0.0001283074,0.0001458278,0.00009182026,0.0005974818,0.0001104085,0.0002761535,0.000156138,0.0009707327,0.0002420293],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004701322,"about_ca_system_score_gemma":0.002265659,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008024312,"about_ca_topic_score_gemma":0.0009469118,"domain_scores_codex":[0.9944448,0.001183544,0.001525599,0.0002885071,0.001136359,0.001421205],"domain_scores_gemma":[0.9949021,0.002731667,0.0005557514,0.0003435183,0.001108166,0.0003587417],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.005094165,0.0006647306,0.3250856,0.1966283,0.0003491896,0.000004928492,0.0821787,0.000553225,0.0004206061,0.1662045,0.1658813,0.05693478],"study_design_scores_gemma":[0.008567887,0.002430475,0.3311345,0.006469938,0.00006249201,0.000007458976,0.04168426,0.1394291,0.0001634423,0.007144354,0.4622289,0.0006772664],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8129522,0.004213361,0.06283534,0.01395075,0.008014778,0.02061989,0.003112539,0.00121618,0.07308501],"genre_scores_gemma":[0.9605963,0.001326333,0.02163451,0.004916048,0.0033557,0.001958208,0.002413914,0.00009904055,0.003699963],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2963476,"threshold_uncertainty_score":0.9987166,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2765979443","doi":"10.1097/mlr.0000000000000824","title":"Adaptation and Validation of the Combined Comorbidity Score for ICD-10-CM","year":2017,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":138,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Medicine; ICD-10; Comorbidity; Cohort; Logistic regression; Statistic; Confidence interval; Diagnosis code; Internal medicine; Emergency medicine; Statistics; Population; Psychiatry; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.3560323928742396,"gpt":0.4814119887053111,"spread":0.1253795958310715,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.001092763,0.00006062217,0.0001565016,0.00002066669,0.0015019,0.000008218508,0.0002226319,0.0002087723,0.0003725998],"category_scores_gemma":[0.005886836,0.00003878652,0.00003365047,0.00002750965,0.0001431691,0.00009006549,0.0001009777,0.0003050825,0.00001274772],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004833776,"about_ca_system_score_gemma":0.0004142321,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002095342,"about_ca_topic_score_gemma":0.0002267219,"domain_scores_codex":[0.9987227,0.0001515266,0.000390886,0.00009182253,0.0004657599,0.000177278],"domain_scores_gemma":[0.9984266,0.0004229429,0.0003790363,0.0003020705,0.0003011897,0.0001681167],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0007631393,0.000108753,0.4538666,0.01722414,0.00004248893,0.00000180231,0.07258403,0.00001410504,0.00004539723,0.0275224,0.139841,0.2879862],"study_design_scores_gemma":[0.00917161,0.0006157642,0.7818001,0.00342748,0.0001017115,0.000001716016,0.01327508,0.02186305,0.0005701074,0.004068266,0.1648039,0.0003011659],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.966336,0.0001306998,0.003607156,0.02034177,0.001801896,0.001429626,0.0000671474,0.00004275465,0.006242922],"genre_scores_gemma":[0.9981126,0.00003849567,0.0001366919,0.0009890535,0.000218615,0.00008425553,0.000079875,0.000005610303,0.0003348088],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3279335,"threshold_uncertainty_score":0.999798,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2133188589","doi":"10.1007/s00198-008-0689-9","title":"Special report on the 2007 adult and pediatric Position Development Conferences of the International Society for Clinical Densitometry","year":2008,"lang":"en","type":"article","venue":"Osteoporosis International","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":130,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Shriners Hospitals for Children - Canada; University of British Columbia","funders":"","keywords":"Medicine; Standardization; Medical physics; Medical education; Relevance (law); Position paper; Family medicine; Pathology; Political science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.2354663556628102,"gpt":0.4576065724831413,"spread":0.2221402168203311,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00159101,0.00009180982,0.0001461943,0.00005728917,0.0007416053,0.00001028186,0.0002883142,0.0001557065,0.0004327043],"category_scores_gemma":[0.001696534,0.00005452011,0.0001285118,0.00008361898,0.000148583,0.00008111349,0.0001172241,0.0003936466,0.0000314123],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001429528,"about_ca_system_score_gemma":0.0008311209,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000209871,"about_ca_topic_score_gemma":0.00001967211,"domain_scores_codex":[0.9979027,0.00007586483,0.001006935,0.0001718302,0.000677933,0.0001647336],"domain_scores_gemma":[0.9972533,0.001104337,0.0006911231,0.0001385242,0.0007443943,0.00006832011],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001072665,0.00007164514,0.51851,0.00005647297,0.00005429836,9.364636e-7,0.002176377,0.000001089027,0.000004410522,0.006141382,0.4711056,0.001770436],"study_design_scores_gemma":[0.001159241,0.00007961545,0.7957237,0.0001607549,0.00002459177,0.0000246602,0.001299006,0.0005123157,0.00009397047,0.0003302275,0.2004809,0.000111022],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.945779,0.00001424008,0.0003839863,0.01645438,0.01029922,0.0005971272,0.00005193469,0.00001995214,0.02640011],"genre_scores_gemma":[0.9748635,0.0005526114,0.002126029,0.00892623,0.009862809,0.0001410111,0.0001580357,0.000009832268,0.003359971],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2772136,"threshold_uncertainty_score":0.5703903,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2503745853","doi":"10.1136/bmjopen-2015-009952","title":"Systematic review of validated case definitions for diabetes in ICD-9-coded and ICD-10-coded data in adult populations","year":2016,"lang":"en","type":"review","venue":"BMJ Open","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":129,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Alberta Children's Hospital; Hotchkiss Brain Institute; University of Calgary","funders":"Cumming School of Medicine, University of Calgary; Canadian Institutes of Health Research; Alberta Innovates; University of Calgary","keywords":"Medicine; Diabetes mellitus; MEDLINE; Population; Diagnosis code; Predictive value; Epidemiology; Internal medicine; Family medicine; Environmental health","retraction":null,"screen_n_in":null,"score":{"opus":0.7108628528866195,"gpt":0.6097166158974214,"spread":0.1011462369891981,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01033811,0.0003227967,0.003554288,0.0003853015,0.0003379999,0.00002521122,0.0008660103,0.0004649749,0.0002976261],"category_scores_gemma":[0.01753754,0.0002171529,0.0001092686,0.0005606792,0.00004284589,0.0004301385,0.0007537853,0.0005404826,0.0001623089],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000213842,"about_ca_system_score_gemma":0.001547401,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004764912,"about_ca_topic_score_gemma":0.001670014,"domain_scores_codex":[0.9914417,0.002267068,0.005049994,0.0004182307,0.0002981391,0.0005248156],"domain_scores_gemma":[0.9919671,0.003322659,0.002611299,0.00155743,0.0003317333,0.0002097598],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.000007439021,0.00003126108,0.00003705984,0.9851791,0.00002772303,0.000007168741,0.0001454695,3.017943e-8,9.754146e-9,0.002011338,0.007011173,0.005542218],"study_design_scores_gemma":[0.0009128578,0.00003398219,0.000007132142,0.9570051,0.0005174915,0.00002293548,0.0001101366,0.0002550439,1.918323e-8,0.0002869359,0.04060515,0.0002431433],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00001065022,0.9212714,0.0001214109,0.001212032,0.0002617187,0.07118206,0.004335587,0.0000331595,0.001572028],"genre_scores_gemma":[0.000008460729,0.9663018,0.00157477,0.001118012,0.00006557356,0.02503417,0.005508557,0.00004135818,0.0003472743],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.04614789,"threshold_uncertainty_score":0.9907382,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2328431583","doi":"10.9778/cmajo.20140128","title":"Representativeness of patients and providers in the Canadian Primary Care Sentinel Surveillance Network: a cross-sectional study","year":2016,"lang":"en","type":"article","venue":"CMAJ Open","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":129,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":true},"ca_institutions":"Queen's University; University of Calgary","funders":"","keywords":"Medicine; Representativeness heuristic; Census; Family medicine; Primary care; Cross-sectional study; Population; Sample (material); Demography; Environmental health; Psychology","retraction":null,"screen_n_in":null,"score":{"opus":0.1962442400131248,"gpt":0.4711928459567593,"spread":0.2749486059436345,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002170847,0.00005892571,0.0001462647,0.00004404057,0.000614274,0.00002744767,0.0002284622,0.00006989429,0.00006802829],"category_scores_gemma":[0.0003102178,0.00003301564,0.00001003333,0.0001301477,0.00006071363,0.0002107045,0.0001318207,0.0001848221,0.00001220336],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002067722,"about_ca_system_score_gemma":0.0008329288,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.1763852,"about_ca_topic_score_gemma":0.5052322,"domain_scores_codex":[0.9982637,0.0006342338,0.0004575643,0.0001393752,0.0002285969,0.0002765663],"domain_scores_gemma":[0.998864,0.0004790877,0.0001837871,0.0001658431,0.0002050487,0.0001022533],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00006994454,0.0000226798,0.9883801,0.0001791948,0.000003412853,6.310109e-7,0.008869968,0.000002871332,1.405483e-7,0.00006905173,0.001310794,0.001091164],"study_design_scores_gemma":[0.001612208,0.00005097817,0.9938647,0.00009583969,0.000001102725,1.136428e-7,0.002161486,0.000007551916,1.242223e-7,0.00006271605,0.002100412,0.00004274432],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9865069,0.000020806,0.00001051151,0.0007694101,0.000223547,0.002201273,0.00003665405,0.000005220176,0.01022567],"genre_scores_gemma":[0.9983186,0.000006337986,0.00001668436,0.00110681,0.00006339906,0.0002188328,0.00004417159,0.000004644664,0.0002205941],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3288469,"threshold_uncertainty_score":0.8290993,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2073780661","doi":"10.1186/1472-6963-14-70","title":"Coding of obesity in administrative hospital discharge abstract data: accuracy and impact for future research studies","year":2014,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":128,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Libin Cardiovascular Institute of Alberta; University of Alberta; University of Calgary","funders":"Alberta Heritage Foundation for Medical Research; Fondation pour la Recherche Médicale; Alberta Innovates; Alberta Innovates - Health Solutions","keywords":"Nursing research; Health administration; Health informatics; Medicine; Public health; Coding (social sciences); Quality of Life Research; Health services research; Obesity; Hospital discharge; Nursing; Intensive care medicine; Internal medicine; Sociology; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.6267415592290458,"gpt":0.6724859504531115,"spread":0.04574439122406571,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts"],"consensus_categories":[],"category_scores_codex":[0.04419873,0.0001757335,0.0006536973,0.0005442225,0.002165821,0.00003336232,0.0007089445,0.0002805303,0.00005799339],"category_scores_gemma":[0.003690228,0.0001309814,0.00003617434,0.0007851322,0.0002363443,0.0006816673,0.0006972149,0.002238577,0.00003930735],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004638894,"about_ca_system_score_gemma":0.00302068,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002831758,"about_ca_topic_score_gemma":0.006392391,"domain_scores_codex":[0.9923044,0.002808556,0.001387962,0.0005242013,0.00128894,0.001685985],"domain_scores_gemma":[0.9805689,0.01591431,0.0004851517,0.0008177444,0.001526451,0.0006874512],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001301913,0.0005267026,0.6457556,0.1384945,0.00005634431,0.000001574575,0.1728247,0.000003431506,0.00002532138,0.01123061,0.007884351,0.02189489],"study_design_scores_gemma":[0.001850541,0.001539105,0.8084129,0.003977173,0.000003005473,5.47775e-7,0.1633603,0.006420311,0.000009331164,0.001598771,0.01267626,0.0001518002],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9861755,0.002384651,0.0001156765,0.005530819,0.00040755,0.004367634,0.0003643832,0.00003398452,0.0006198347],"genre_scores_gemma":[0.9928079,0.004539885,0.0008275725,0.0003420662,0.0008041341,0.0003438182,0.0002363555,0.00002049882,0.00007778286],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1626572,"threshold_uncertainty_score":0.9991332,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2768917383","doi":"10.1186/s12913-017-2697-y","title":"Barriers to data quality resulting from the process of coding health information to administrative data: a qualitative study","year":2017,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":124,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Health Sciences Centre; University of Calgary","funders":"O'Brien Institute for Public Health, University of Calgary; Canadian Institutes of Health Research; Alberta Innovates","keywords":"Data quality; Health informatics; Health administration; Nursing research; Health services research; Data collection; Quality (philosophy); Coding (social sciences); Information quality; Data science; Medicine; Computer science; Knowledge management; Public health; Information system; Business; Nursing; Marketing; Statistics; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.7821812141022199,"gpt":0.7238918634367201,"spread":0.05828935066549978,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","open_science"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.08977824,0.0002263555,0.0007546121,0.0002891465,0.01075595,0.0002038207,0.005386893,0.0001576273,0.0001452828],"category_scores_gemma":[0.01816348,0.0001649346,0.00002327785,0.0007555418,0.0001459208,0.00241267,0.003526596,0.001762091,0.0003050691],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005056095,"about_ca_system_score_gemma":0.01221277,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.103001,"about_ca_topic_score_gemma":0.06834656,"domain_scores_codex":[0.9796536,0.01174231,0.003419671,0.0007642279,0.002836392,0.001583796],"domain_scores_gemma":[0.9798763,0.007947517,0.002563195,0.005548592,0.001801087,0.002263339],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.0008699228,0.00006424313,0.03487903,0.005769485,0.00002561683,2.359483e-7,0.9435176,0.000006494143,7.929213e-7,0.0003646656,0.00746972,0.007032245],"study_design_scores_gemma":[0.001068198,0.000647935,0.1621353,0.002070039,0.000003672651,8.576485e-8,0.8200419,0.003895365,0.000001474378,0.0001172546,0.009893475,0.0001253399],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8893311,0.0001425515,0.004371568,0.07588966,0.0008735507,0.01740672,0.009462138,0.0001323935,0.002390324],"genre_scores_gemma":[0.9808752,0.00005131488,0.002194515,0.0141118,0.0004158107,0.0004053577,0.001871564,0.00001862042,0.00005579674],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1272562,"threshold_uncertainty_score":0.9999945,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1589504844","doi":"10.1002/0470842555.ch20","title":"Health Databases in Saskatchewan","year":2000,"lang":"en","type":"other","venue":"Pharmacoepidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":124,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Saskatchewan Health","funders":"","keywords":"Strengths and weaknesses; Database; Computer science; Data science; Psychology","retraction":null,"screen_n_in":null,"score":{"opus":0.4054909531439705,"gpt":0.5730101256759584,"spread":0.1675191725319879,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","research_integrity","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.003752942,0.0003047654,0.001273214,0.0005298242,0.0002137449,9.692332e-7,0.0002623583,0.0004006518,0.0412615],"category_scores_gemma":[0.0005225195,0.0002642229,0.00006321107,0.0002105947,0.00009017565,0.00004519228,0.00008121258,0.002368387,0.003565661],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001984337,"about_ca_system_score_gemma":0.001390936,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.05166915,"about_ca_topic_score_gemma":0.04031486,"domain_scores_codex":[0.992714,0.003806404,0.001669819,0.0004148615,0.0001650349,0.001229915],"domain_scores_gemma":[0.9964652,0.001734718,0.0009387546,0.0003936029,0.00001111374,0.0004565936],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00003657988,0.0000264284,0.009229084,0.001410796,0.00001285889,0.000004207649,0.0007302542,0.000002595921,3.333979e-7,0.0002055547,0.9607931,0.02754824],"study_design_scores_gemma":[0.000926886,0.0000250748,0.001065574,0.001252671,0.000007835882,0.000002002911,0.0003063935,0.0002475114,1.510988e-7,0.00008048279,0.9958966,0.0001888607],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.00007213658,0.006437196,0.001023299,0.02096709,0.003082865,0.00181481,0.0007153124,0.0006593741,0.9652279],"genre_scores_gemma":[0.001432488,0.02400625,0.002570638,0.17938,0.00432883,0.0006477374,0.003894155,0.0005126763,0.7832272],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.1820007,"threshold_uncertainty_score":0.999981,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1980271665","doi":"10.1111/j.1475-6773.2012.01440.x","title":"Case Definitions for Acute Myocardial Infarction in Administrative Databases and Their Impact on In‐Hospital Mortality Rates","year":2012,"lang":"en","type":"article","venue":"Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":112,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Alberta Health Services; University of Calgary","funders":"Canadian Institutes of Health Research; Alberta Innovates; Alberta Innovates - Health Solutions","keywords":"Medicine; Myocardial infarction; Comparability; Diagnosis code; ICD-10; Hospital discharge; Emergency medicine; MEDLINE; Database; Medical emergency; Intensive care medicine; Internal medicine; Population; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.5993403575484629,"gpt":0.6351433606023774,"spread":0.03580300305391448,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006683906,0.0001638396,0.0003607071,0.0004361541,0.001218528,0.00001853493,0.00009188015,0.0001614026,0.00008294761],"category_scores_gemma":[0.0002264326,0.0001174917,0.0000457195,0.0005223568,0.00007639567,0.0005475579,0.00009013484,0.001261139,0.00005437765],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005298493,"about_ca_system_score_gemma":0.001119937,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0141934,"about_ca_topic_score_gemma":0.008546084,"domain_scores_codex":[0.9960846,0.001247875,0.0007845206,0.0002348158,0.0003420252,0.001306134],"domain_scores_gemma":[0.9960887,0.00254581,0.0002373168,0.0002601108,0.0002213317,0.0006467484],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001323066,0.0009326835,0.7943236,0.007976343,0.000110299,0.00003670621,0.1752746,0.000009583884,0.000007794716,0.01114213,0.00293573,0.005927402],"study_design_scores_gemma":[0.002010371,0.001602058,0.9314082,0.001162956,0.000008014645,0.00003142796,0.05694634,0.00272964,0.0000100959,0.0008315723,0.003072264,0.0001870154],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9913399,0.0003078851,0.00003930938,0.003718326,0.0003105992,0.002565645,0.0008562062,0.00003166925,0.0008305027],"genre_scores_gemma":[0.9948991,0.001176705,0.0001414142,0.001850428,0.0003433752,0.001041669,0.0005206166,0.00001474635,0.00001192646],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1370846,"threshold_uncertainty_score":0.9923711,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2052667634","doi":"10.1097/mlr.0b013e3182329778","title":"Administrative Hospitalization Database Validation of Cardiac Procedure Codes","year":2011,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":112,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Health Sciences Centre; Institute for Clinical Evaluative Sciences; University of Toronto; University Health Network; Toronto General Hospital; Sunnybrook Health Science Centre; Ontario Stroke Network; Canadian Institute for Health Information","funders":"Canadian Institutes of Health Research","keywords":"Medicine; Conventional PCI; Percutaneous coronary intervention; Cardiac catheterization; Emergency medicine; Cardiac surgery; Coronary artery disease; Database; Surgery; Medical emergency; Internal medicine; Myocardial infarction","retraction":null,"screen_n_in":null,"score":{"opus":0.2707885593169428,"gpt":0.4702386477621798,"spread":0.1994500884452369,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0008966117,0.00009830695,0.0002405093,0.00006414711,0.0002696568,0.000001937469,0.0001437414,0.0002830575,0.002451673],"category_scores_gemma":[0.002325767,0.00007788692,0.00004097579,0.0001819496,0.00009090154,0.0001519593,0.00005285281,0.0004294528,0.0001017322],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007487314,"about_ca_system_score_gemma":0.001090566,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001920408,"about_ca_topic_score_gemma":0.0000525699,"domain_scores_codex":[0.9979091,0.0002989302,0.0006431273,0.0001480758,0.0007417653,0.0002590369],"domain_scores_gemma":[0.9985511,0.0001828497,0.0002804156,0.0001996913,0.000451704,0.0003342784],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003324866,0.0002142513,0.5263983,0.0158397,0.00006982705,0.0000100787,0.3514405,9.882176e-7,0.00009028025,0.02898256,0.05236677,0.0242542],"study_design_scores_gemma":[0.005934741,0.003387307,0.4528632,0.01282604,0.000458508,0.000007303326,0.3558771,0.002252807,0.0303711,0.002600142,0.1317527,0.001669085],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8912141,0.0006976754,0.004826398,0.00112489,0.001909881,0.00170809,0.000358547,0.0002232338,0.09793714],"genre_scores_gemma":[0.9974215,0.0001262092,0.0005241021,0.0005734501,0.0002337186,0.0001165281,0.0008652089,0.000010278,0.0001289984],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1062074,"threshold_uncertainty_score":0.9984602,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2003471247","doi":"10.1186/s12875-015-0223-z","title":"From patient care to research: a validation study examining the factors contributing to data quality in a primary care electronic medical record database","year":2015,"lang":"en","type":"article","venue":"BMC Family Practice","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":109,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Calgary; University of Manitoba","funders":"","keywords":"Medicine; Medical record; Audit; COPD; Comorbidity; Diagnosis code; MEDLINE; Data quality; Coding (social sciences); Family medicine; Database; Psychiatry; Population; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.8174282791767788,"gpt":0.6145045267607676,"spread":0.2029237524160112,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.03176273,0.0001771117,0.0003610936,0.0002026559,0.001069065,0.00005528674,0.0008857041,0.0002102302,0.00004553656],"category_scores_gemma":[0.1242606,0.0001305152,0.00001555527,0.0007984552,0.00002987545,0.0008631642,0.001870183,0.002646524,0.0002756983],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002154463,"about_ca_system_score_gemma":0.006672024,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.05024121,"about_ca_topic_score_gemma":0.01619688,"domain_scores_codex":[0.9824132,0.01173054,0.001407953,0.0006143174,0.00261855,0.001215399],"domain_scores_gemma":[0.9774373,0.01851663,0.0004622668,0.00150268,0.001270006,0.0008111651],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.002574869,0.0003286545,0.3969501,0.0005691033,0.00003065303,0.00001779472,0.5292618,0.00002202034,0.00001992397,0.0000868752,0.01901918,0.051119],"study_design_scores_gemma":[0.001505309,0.0008964973,0.1149323,0.0005270645,0.00002459376,3.443433e-7,0.7861418,0.0002464433,0.000002110865,0.000009932612,0.09556175,0.0001518842],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9852653,0.0003200415,0.003880365,0.002308236,0.000810776,0.003904883,0.0003914815,0.00007529472,0.003043645],"genre_scores_gemma":[0.9894938,0.00002873262,0.001432974,0.005483564,0.0003631236,0.000394109,0.002763728,0.00002165392,0.00001835492],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2820178,"threshold_uncertainty_score":0.9996544,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2645528799","doi":"10.1093/ije/dyw248","title":"Data Resource Profile: National electronic medical record data from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN)","year":2017,"lang":"en","type":"article","venue":"International Journal of Epidemiology","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":97,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Queen's University; University of Alberta; University of Calgary","funders":"","keywords":"Primary care; Medicine; Electronic medical record; Medical record; Resource (disambiguation); Medical emergency; Family medicine; Computer science; Internal medicine; Computer network","retraction":null,"screen_n_in":null,"score":{"opus":0.4583917073867911,"gpt":0.535116700476173,"spread":0.07672499308938191,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","open_science","research_integrity","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.02430296,0.0001410927,0.0004677307,0.00008671752,0.001563309,0.00003303299,0.006631441,0.0003947794,0.001279005],"category_scores_gemma":[0.06270233,0.00009522196,0.0000561565,0.00004824398,0.0002471638,0.000463208,0.001270237,0.002386202,0.0001664205],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008921097,"about_ca_system_score_gemma":0.009470568,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.03609387,"about_ca_topic_score_gemma":0.3118232,"domain_scores_codex":[0.9939445,0.002109405,0.001859899,0.0003146225,0.001040889,0.0007307163],"domain_scores_gemma":[0.9852726,0.009128826,0.002570756,0.001340264,0.00120492,0.0004825792],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001274651,0.000009335354,0.322321,0.00002145703,0.0001118482,0.00001333662,0.0001434848,0.00001864386,1.394556e-7,0.002500119,0.6620983,0.01263482],"study_design_scores_gemma":[0.0005734189,0.00002211413,0.2978632,0.0003063959,0.000009889673,0.00002600083,0.00009823914,0.007021216,1.937337e-8,0.002437385,0.6915755,0.00006665642],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.03532541,0.005561314,0.006262418,0.8451145,0.01986998,0.0009205844,0.005805169,0.00006119952,0.08107939],"genre_scores_gemma":[0.7528601,0.00339955,0.004078424,0.1765921,0.03803719,0.00002077161,0.02424672,0.00004654382,0.0007185868],"genre_candidate":"commentary","genre_consensus":null,"teacher_disagreement_score":0.7175347,"threshold_uncertainty_score":0.9999154,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2046795159","doi":"10.1111/j.1475-6773.2005.00419.x","title":"Does Date Stamping ICD‐9‐CM Codes Increase the Value of Clinical Information in Administrative Data?","year":2005,"lang":"en","type":"article","venue":"Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":82,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Agency for Healthcare Research and Quality; Dartmouth College","keywords":"Medicine; Diagnosis code; Comorbidity; Emergency medicine; Heart failure; Myocardial infarction; ICD-10; Population; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.600592101594237,"gpt":0.6699631960682774,"spread":0.06937109447404033,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.0436551,0.000137223,0.0005014673,0.0003239618,0.001331264,0.00003254922,0.001087432,0.0002699167,0.0002996902],"category_scores_gemma":[0.0017146,0.00007382359,0.00003182478,0.0007311006,0.0002050704,0.001406293,0.0006821501,0.002643497,0.0004174804],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001892175,"about_ca_system_score_gemma":0.004492612,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01976614,"about_ca_topic_score_gemma":0.01863454,"domain_scores_codex":[0.9886082,0.005416215,0.003251055,0.0002758908,0.001372152,0.001076445],"domain_scores_gemma":[0.9909819,0.0054689,0.001253487,0.001094954,0.0006542981,0.0005464549],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003486829,0.0007562581,0.2604975,0.05008357,0.00008043535,0.000005768959,0.2567786,0.0002917334,0.000005970067,0.02989398,0.03219726,0.3659222],"study_design_scores_gemma":[0.00225963,0.0004201613,0.2136578,0.003341569,0.000005657831,0.000001090851,0.06633066,0.1438092,0.0000104732,0.0004813326,0.5695199,0.0001626253],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8958966,0.0004287806,0.0004170273,0.08291479,0.0008780329,0.00482511,0.0007284702,0.0001107032,0.01380054],"genre_scores_gemma":[0.9798985,0.002466704,0.001399681,0.01444706,0.0006705302,0.0001560284,0.0007861876,0.00001169549,0.0001636373],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5373226,"threshold_uncertainty_score":0.9999689,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2012122949","doi":"10.1186/1472-6963-11-194","title":"Improved accuracy of co-morbidity coding over time after the introduction of ICD-10 administrative data","year":2011,"lang":"en","type":"article","venue":"BMC Health Services Research","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":81,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Medicine; Chart; Kappa; Coding (social sciences); Health administration; Health informatics; Nursing research; ICD-10; Emergency medicine; Pediatrics; Statistics; Public health; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.5146041854955904,"gpt":0.583941607519064,"spread":0.0693374220234736,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0149131,0.0001417967,0.0004196685,0.0002127541,0.0009213002,0.00001017128,0.0009553456,0.0002216149,0.008634484],"category_scores_gemma":[0.0009803405,0.00009677462,0.00003847092,0.0005623171,0.0002476279,0.0005068478,0.0005198866,0.00130434,0.0004782003],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001817553,"about_ca_system_score_gemma":0.003108093,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005043841,"about_ca_topic_score_gemma":0.001245717,"domain_scores_codex":[0.9936428,0.002622606,0.001510931,0.0003640082,0.0009876981,0.0008719641],"domain_scores_gemma":[0.9939929,0.002407867,0.0009443207,0.001412066,0.0008776595,0.0003651798],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.02279752,0.001651502,0.2227142,0.1639886,0.0002356834,0.000004758905,0.3667186,0.000003335018,0.001981771,0.005390375,0.1871044,0.0274092],"study_design_scores_gemma":[0.002920583,0.002158754,0.7913706,0.002608857,0.00003193031,0.00000425638,0.03547262,0.03928279,0.001050453,0.0005488525,0.1241741,0.0003762188],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9716759,0.0006710588,0.0007832056,0.004821569,0.0008454136,0.006297065,0.0009740641,0.0001150753,0.01381659],"genre_scores_gemma":[0.994889,0.00028749,0.0009707694,0.001081867,0.0009033854,0.0001965783,0.0004129934,0.00002112475,0.0012368],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5686563,"threshold_uncertainty_score":0.9922718,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2749759883","doi":"10.9778/cmajo.20170036","title":"Coder perspectives on physician-related barriers to producing high-quality administrative data: a qualitative study","year":2017,"lang":"en","type":"article","venue":"CMAJ Open","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":80,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"about_ca":true},"ca_institutions":"University of Calgary","funders":"","keywords":"Qualitative research; Quality (philosophy); Business; Qualitative property; Data quality; Nursing; Public relations; Psychology; Medicine; Political science; Computer science; Sociology; Marketing","retraction":null,"screen_n_in":null,"score":{"opus":0.5926865241881064,"gpt":0.6389193990915385,"spread":0.04623287490343209,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.008219472,0.0002052094,0.0005254538,0.00006942692,0.00483239,0.0001595583,0.001637071,0.0001305931,0.0007737655],"category_scores_gemma":[0.01235723,0.000164195,0.00002571186,0.0001299657,0.0001327497,0.001031686,0.001110868,0.0009137637,0.001126274],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002715967,"about_ca_system_score_gemma":0.001280445,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007396118,"about_ca_topic_score_gemma":0.001460211,"domain_scores_codex":[0.9953331,0.002118341,0.0008578814,0.0007025993,0.0004889853,0.0004991171],"domain_scores_gemma":[0.9952773,0.0008544928,0.0008254469,0.002094947,0.0003546389,0.0005932167],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.0004126418,0.000203975,0.002444422,0.000115992,0.00008847177,0.000003216861,0.9568246,0.000003548061,0.00001215778,0.01499606,0.02331224,0.001582694],"study_design_scores_gemma":[0.001949533,0.0007292638,0.105606,0.0004531096,0.00002553111,8.330682e-8,0.8859609,0.00004219629,0.00001795944,0.001014468,0.003959509,0.0002413914],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8935106,0.00001278711,0.0001898616,0.02112786,0.0009365565,0.006931359,0.0002923378,0.00009037339,0.07690827],"genre_scores_gemma":[0.990966,0.000004305471,0.0004004173,0.003561192,0.0002693526,0.0004865838,0.00009086796,0.00002021702,0.004201064],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1031616,"threshold_uncertainty_score":0.9996515,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4385322996","doi":"10.1016/s2468-2667(23)00123-8","title":"The burden and trend of diseases and their risk factors in Australia, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019","year":2023,"lang":"en","type":"article","venue":"The Lancet Public Health","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":75,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Cilag; Leibniz-Gemeinschaft; National Health and Medical Research Council; Medical Research Council; Max-Planck-Institut für demografische Forschung; Western Sydney University; H. Lundbeck A/S; Society for Mental Health Research; Servier; Deakin University; Cabrini Foundation; Fondation de la recherche en santé du Nouveau-Brunswick; University of Sydney; World Health Organization; Wellcome Trust; Government of Western Australia; University of Pennsylvania; UNICEF; Arthritis Australia; Queensland Health; International Association for the Study of Pain; Pfizer; Biogen; Department of Health and Aged Care, Australian Government; Curtin University of Technology; Sanofi; Amgen; Bill and Melinda Gates Foundation; Australian Government; International Labour Organization; Institute for Health Metrics and Evaluation; LivaNova; National Heart Foundation of Australia; U.S. Department of Veterans Affairs","keywords":"Burden of disease; Medicine; Disease; Disease burden; MEDLINE; Environmental health; Political science; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.2693069298686623,"gpt":0.4695835153047814,"spread":0.2002765854361191,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00876315,0.00014777,0.0007933021,0.0001420918,0.001022697,0.00003019079,0.0003733211,0.00005938969,0.00001021614],"category_scores_gemma":[0.002317223,0.00005943977,0.00008136183,0.001079244,0.0001568155,0.00009092935,0.000130676,0.0003082062,0.000003179433],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001097086,"about_ca_system_score_gemma":0.0004754904,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01913145,"about_ca_topic_score_gemma":0.002132009,"domain_scores_codex":[0.9957572,0.001962577,0.001116619,0.0001736427,0.000350853,0.0006391156],"domain_scores_gemma":[0.9920549,0.006046137,0.0008592166,0.0006036157,0.00008509168,0.0003509827],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002360206,0.00006077794,0.8918701,0.01608308,0.0005376589,2.287548e-7,0.07348767,0.0003564505,1.179852e-7,0.002424363,0.01241459,0.002528937],"study_design_scores_gemma":[0.001070015,0.0001672862,0.9179095,0.0005984986,0.0002221655,9.666282e-8,0.06031026,0.01687694,2.040818e-8,0.0008517086,0.001921505,0.00007197923],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9447187,0.001442876,0.0001500181,0.04839964,0.000199298,0.003453716,0.001570523,0.0000449091,0.00002034094],"genre_scores_gemma":[0.9973978,0.001516434,0.000004218091,0.0002058602,0.0002582613,0.000301462,0.00006966441,0.000007196616,0.0002390845],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.05267914,"threshold_uncertainty_score":0.9874002,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3198282417","doi":"10.2196/23230","title":"Automatic ICD-10 Coding and Training System: Deep Neural Network Based on Supervised Learning","year":2021,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":73,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"Ministry of Science and Technology, Taiwan","keywords":"Computer science; Coding (social sciences); Artificial intelligence; Deep learning; Encoder; Artificial neural network; Medical diagnosis; Natural language processing; Medical classification; Machine learning; Language model; Autoencoder; ICD-10; Speech recognition; Medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.1282203436536169,"gpt":0.4106309465578995,"spread":0.2824106029042825,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.003182166,0.000244234,0.0005738199,0.0001321642,0.001447255,0.00005183149,0.000196532,0.0005516251,0.003516628],"category_scores_gemma":[0.002387139,0.0002014525,0.000073035,0.0004227782,0.0001003697,0.0002826063,0.0001305894,0.002169149,0.0004454574],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001944113,"about_ca_system_score_gemma":0.0008793746,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006576281,"about_ca_topic_score_gemma":0.00001068744,"domain_scores_codex":[0.9950857,0.0006505788,0.001852922,0.0001470431,0.001304696,0.0009591118],"domain_scores_gemma":[0.9957448,0.002302511,0.0004842668,0.0002830215,0.0002023706,0.0009830124],"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":[0.000251901,0.0001713193,0.01365824,0.05453917,0.0001086413,0.0001989832,0.1871388,0.007743313,0.000004915772,0.01530793,0.04428297,0.6765939],"study_design_scores_gemma":[0.001687275,0.0001346458,0.001306717,0.00326352,0.00001999789,0.00002097573,0.03012036,0.9512517,8.58133e-7,0.00001983981,0.01198454,0.0001895978],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8079873,0.000233419,0.03980399,0.00928612,0.003475177,0.002486408,0.00001417962,0.00220641,0.134507],"genre_scores_gemma":[0.9781799,0.0000318153,0.004118003,0.01609817,0.0007776066,0.0002118475,0.0001743905,0.00003076993,0.0003775011],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9435084,"threshold_uncertainty_score":0.9998527,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W1982351272","doi":"10.1097/mlr.0b013e3181649439","title":"Comparison and Validity of Procedures Coded With ICD-9-CM and ICD-10-CA/CCI","year":2008,"lang":"en","type":"article","venue":"Medical Care","topic":"Medical Coding and Health Information","field":"Health Professions","cited_by":69,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Calgary; Alberta Health Services","funders":"Health Canada","keywords":"ICD-10; Kappa; Medicine; Chart; Coding (social sciences); Predictive value; Population; Hospital discharge; Statistics; Mathematics; Internal medicine; Environmental health; Psychiatry","retraction":null,"screen_n_in":null,"score":{"opus":0.2327204175428824,"gpt":0.4709423912395587,"spread":0.2382219736966764,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0006690212,0.0001202552,0.000396629,0.00005830449,0.0005818934,0.000002819003,0.00009531071,0.0002892925,0.001190349],"category_scores_gemma":[0.001628432,0.00008134396,0.00001715946,0.000112975,0.0002576472,0.00006782222,0.0000761439,0.0005853758,0.00003089052],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005546819,"about_ca_system_score_gemma":0.0008457325,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000187456,"about_ca_topic_score_gemma":0.0005407071,"domain_scores_codex":[0.997923,0.0002232484,0.0005854775,0.0001731962,0.0007650417,0.0003300871],"domain_scores_gemma":[0.998323,0.0004984373,0.0002083165,0.0001522737,0.000242418,0.0005756092],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002912033,0.00006556542,0.8349922,0.008799821,0.00001588623,0.00002082478,0.0448826,0.000001408117,0.000005468381,0.000368744,0.1028666,0.007689626],"study_design_scores_gemma":[0.007511269,0.00187584,0.7888298,0.004080424,0.00009384881,0.00009899132,0.02601319,0.004075698,0.0001407757,0.0001651127,0.166593,0.0005220291],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.987892,0.0007152811,0.0004723975,0.003357761,0.0001543558,0.0004962305,0.00001742549,0.0000678025,0.006826776],"genre_scores_gemma":[0.996972,0.000335998,0.000348504,0.001634672,0.0001743587,0.0000458131,0.00005841728,0.000009717961,0.0004205386],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0637264,"threshold_uncertainty_score":0.9997227,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}