{"id":"W4410957636","doi":"10.1016/j.landig.2025.01.013","title":"Importance of sample size on the quality and utility of AI-based prediction models for healthcare","year":2025,"lang":"en","type":"review","venue":"The Lancet Digital Health","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":50,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"European Regional Development Fund; Medical Research Council; Vlaamse regering; South Asian Health Foundation; Vifor Pharma; KU Leuven; National Institute for Health and Care Research; Birmingham Biomedical Research Centre; Engineering and Physical Sciences Research Council; CSL Behring; UK Research and Innovation; Department of Health and Social Care; University Hospitals Birmingham NHS Foundation Trust; Cancer Research UK; Wellcome Trust; Fonds Wetenschappelijk Onderzoek","keywords":"Sample (material); Sample size determination; Health care; Computer science; Quality (philosophy); Agency (philosophy); Artificial intelligence; Outcome (game theory); Machine learning; Data science; Data mining; Risk analysis (engineering); Medicine; Statistics; Mathematics","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002030386,0.0002075898,0.001618424,0.00005234468,0.0001508105,0.00001520546,0.0001772377,0.0001402914,0.000005995699],"category_scores_gemma":[0.002870039,0.0001141215,0.0002373211,0.0002816344,0.0001735123,0.00006023474,0.00002481306,0.0004137342,5.188174e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000115654,"about_ca_system_score_gemma":0.003027937,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001173812,"about_ca_topic_score_gemma":0.0001898533,"domain_scores_codex":[0.9974591,0.0002600628,0.001380896,0.0003136527,0.0002690052,0.0003173393],"domain_scores_gemma":[0.9884778,0.009476243,0.0008228628,0.0008184156,0.0002965209,0.0001081661],"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":[0.0002580764,0.0001155561,0.001948359,0.08736352,0.00003199635,5.229034e-8,0.0002276063,0.000002471329,8.155411e-9,0.005252172,0.00231064,0.9024895],"study_design_scores_gemma":[0.0007624492,0.005796402,0.007646642,0.1593926,0.0007889497,0.00001591396,0.002136867,0.004988088,0.00005514827,0.3696001,0.4479633,0.0008535238],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.001033192,0.9404429,0.001389541,0.04439348,0.0003315263,0.005435123,0.006552598,0.00004531568,0.0003763368],"genre_scores_gemma":[0.08248061,0.9110117,0.0001585511,0.005235251,0.0003449035,0.0002945568,0.0004020244,0.00002207759,0.00005033355],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.901636,"threshold_uncertainty_score":0.537143,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5278624674459516,"score_gpt":0.5476276044938809,"score_spread":0.01976513704792937,"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."}}