{"id":"W4409030284","doi":"10.1101/2025.03.30.25324911","title":"A Quarter-Century of Synthetic Data in Healthcare: Unveiling Trends with Structural Topic Modeling","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Quarter (Canadian coin); Health care; Data science; Regional science; Computer science; History; Geography; Economics; Economic growth; Archaeology","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.02852654,0.0008277916,0.001038264,0.006103255,0.0006719496,0.003574131,0.001046369,0.001191758,0.001623712],"category_scores_gemma":[0.07660045,0.0004832614,0.001882077,0.007132216,0.001210255,0.003058059,0.002000027,0.001672899,0.0004206715],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001023354,"about_ca_system_score_gemma":0.001180212,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003312212,"about_ca_topic_score_gemma":0.003278248,"domain_scores_codex":[0.9860795,0.01032758,0.0008768895,0.001697311,0.0007581323,0.0002605956],"domain_scores_gemma":[0.8694819,0.115219,0.006833651,0.004400255,0.00337734,0.000687857],"domain_codex":null,"domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001297935,0.0003700246,0.481338,0.005142251,0.002296825,0.001052382,0.007773243,0.1580471,0.003847348,0.05491862,0.02195239,0.2619639],"study_design_scores_gemma":[0.0001181926,0.0003664473,0.0669328,0.0009777298,0.0005973124,0.0005013257,0.003841531,0.7856323,0.00210933,0.1109863,0.02781882,0.0001179825],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4363885,0.01488922,0.5216033,0.0109688,0.0006404358,0.0005311987,0.01123367,0.0007895478,0.002955295],"genre_scores_gemma":[0.9032846,0.002837603,0.08304094,0.0005293008,0.0005874391,0.0005827848,0.008504152,0.0001112157,0.0005220518],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9938967,"threshold_uncertainty_score":0.1508646,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0794988932785872,"score_gpt":0.4060837622459037,"score_spread":0.3265848689673165,"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."}}