{"id":"W4206671781","doi":"10.2196/preprints.25157","title":"Assessment of Natural Language Processing Methods for Ascertaining the Expanded Disability Status Scale Score From the Electronic Health Records of Patients With Multiple Sclerosis: Algorithm Development and Validation Study (Preprint)","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Multiple Sclerosis Research Studies","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Regent Park Community Health Centre; St. Michael's Hospital","funders":"","keywords":"Expanded Disability Status Scale; F1 score; Health records; Recall; Artificial intelligence; Preprint; Standard score; Convolutional neural network; Computer science; Electronic health record; Machine learning; Precision and recall; Scale (ratio); Natural language processing; Data mining; Algorithm; Medicine; Multiple sclerosis; Psychology; Cartography; Health care","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02273833,0.001210471,0.0009448031,0.002572038,0.0004688567,0.001699606,0.001733835,0.001531927,0.0008399836],"category_scores_gemma":[0.04939223,0.0003969001,0.001294371,0.001200929,0.0004232723,0.001413918,0.001151122,0.00108457,0.0003884031],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001753756,"about_ca_system_score_gemma":0.002146656,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01414183,"about_ca_topic_score_gemma":0.007952498,"domain_scores_codex":[0.991109,0.005052913,0.001246264,0.001269988,0.001090754,0.000231027],"domain_scores_gemma":[0.9388781,0.05037825,0.001793116,0.001957261,0.006714153,0.0002791748],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002586423,0.003376509,0.2415562,0.0009666602,0.00176724,0.0003377813,0.0005413335,0.1703913,0.007319964,0.0009856694,0.005188717,0.5649821],"study_design_scores_gemma":[0.0002788775,0.0007685242,0.02373894,0.00008502031,0.0001777322,0.0002017305,0.0001814611,0.9683437,0.005238742,0.000384808,0.0005699726,0.00003037439],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.8273641,0.002636472,0.1610912,0.000748428,0.0001598396,0.00191838,0.002456684,0.002449448,0.001175272],"genre_scores_gemma":[0.6524023,0.0007530208,0.336684,0.0003336121,0.0000610357,0.001710151,0.007271474,0.0001076953,0.0006767287],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.02273833,"threshold_uncertainty_score":0.1202532,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08600782236259184,"score_gpt":0.4157951596072546,"score_spread":0.3297873372446627,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}