{"id":"W2098161667","doi":"10.2196/jmir.8.4.e28","title":"Evaluating Common De-Identification Heuristics for Personal Health Information","year":2006,"lang":"en","type":"article","venue":"Journal of Medical Internet Research","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":72,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University; Agricultural Research Institute of Ontario","funders":"Ontario Centres of Excellence","keywords":"Observational study; Heuristics; Health Insurance Portability and Accountability Act; Identification (biology); Protected health information; Agency (philosophy); Personally identifiable information; Confidentiality; Internet privacy; Accountability; Data science; Computer science; Medicine; Nursing; Computer security; HRHIS; Public health; Health policy; Political science; Sociology","routes":{"ca_aff":true,"ca_fund":true,"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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.09196275,0.001456926,0.001309696,0.0047265,0.001986263,0.003589434,0.002588163,0.002472621,0.001209637],"category_scores_gemma":[0.2784725,0.0008058482,0.001787835,0.003135822,0.002805183,0.004539701,0.003200635,0.002255109,0.0001388273],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007620784,"about_ca_system_score_gemma":0.005480855,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01726138,"about_ca_topic_score_gemma":0.01362033,"domain_scores_codex":[0.9335848,0.04399648,0.005300188,0.006685022,0.008817239,0.001616245],"domain_scores_gemma":[0.3706336,0.5774354,0.02081823,0.01893284,0.01013111,0.00204881],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","study_design_scores_codex":[0.01027359,0.002156207,0.145086,0.002339726,0.002530697,0.0003580522,0.003829806,0.5722616,0.001670838,0.03127383,0.005623457,0.2225961],"study_design_scores_gemma":[0.0005097949,0.002178813,0.01918882,0.000372751,0.0004355395,0.0003162254,0.001250007,0.950521,0.002978968,0.02023577,0.001894848,0.0001174573],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8617275,0.002034201,0.1268484,0.000945229,0.0001357965,0.001998461,0.001221471,0.0007433465,0.004345531],"genre_scores_gemma":[0.8749155,0.0002273025,0.1224249,0.0002341029,0.00002835349,0.0005294319,0.001137147,0.00004454122,0.0004587066],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9080372,"threshold_uncertainty_score":0.4863513,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.6632204600192323,"score_gpt":0.7130576617179335,"score_spread":0.04983720169870121,"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."}}