{"id":"W4321605311","doi":"10.1101/2023.02.13.23285873","title":"Unlocking the Power of EHRs: Harnessing Unstructured Data for Machine Learning-based Outcome Predictions","year":2023,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Frailty in Older Adults","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Public Health Ontario; University of Toronto","funders":"Data Science Institute, Columbia University; University of Toronto; Institute for Quantitative Social Science, Harvard University","keywords":"Health records; Unstructured data; Machine learning; Robustness (evolution); Predictive power; Computer science; Artificial intelligence; Clinical decision support system; Context (archaeology); Affect (linguistics); Outcome (game theory); Data science; Big data; Psychology; Data mining; Decision support system; Health care","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"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":[0.008701351,0.0007467371,0.0007343325,0.001680164,0.0003155876,0.002723208,0.0009137113,0.0006340998,0.0007606485],"category_scores_gemma":[0.04474403,0.0002893713,0.0005923273,0.001161897,0.0007188678,0.001508791,0.001924397,0.001710455,0.0004574912],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003871272,"about_ca_system_score_gemma":0.001004957,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003090688,"about_ca_topic_score_gemma":0.002936282,"domain_scores_codex":[0.9946805,0.003824203,0.0003094482,0.0005188849,0.0005516647,0.0001153312],"domain_scores_gemma":[0.9684514,0.02452227,0.002009775,0.003187395,0.001446683,0.0003824907],"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":[0.0018547,0.0009851733,0.2618121,0.0008622443,0.0009641923,0.0006549383,0.001311861,0.2465145,0.006724375,0.008691146,0.01170042,0.4579243],"study_design_scores_gemma":[0.00009393474,0.0003346223,0.02353395,0.0003144379,0.0001196076,0.0001700642,0.0004573585,0.936038,0.004314898,0.02980597,0.004749467,0.00006758315],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.4734412,0.00321237,0.5036217,0.007473459,0.0005099485,0.0003923393,0.005428783,0.002240937,0.003679223],"genre_scores_gemma":[0.8875428,0.0007283852,0.1053718,0.0005610402,0.0003819985,0.0001615231,0.004754158,0.000081541,0.0004168186],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008701351,"threshold_uncertainty_score":0.04601771,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1092547781960495,"score_gpt":0.3569994990570273,"score_spread":0.2477447208609778,"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."}}