biostudies-literatureYildiz YNational Institute for Health Research (NIHR)1715506https://www.ebi.ac.uk/biostudies/studies/S-EPMC12929515biostudies-literatureUnknown8<h4>Introduction</h4>Transformer-based models have shown strong potential for clinical prediction using electronic health record data, yet their performance can vary depending on modelling decisions and data characteristics.<h4>Methods</h4>In this study, we trained a BEHRT model on hospital-based UK Biobank data and evaluated its performance across four clinical prediction tasks, including next-visit diagnosis and longer-term diagnosis prediction up to five years. We exhaustively assessed the impact of model size, medical terminology (CALIBER vs ICD-10), and data split strategies.<h4>Results</h4>The large model consistently outperformed the smaller one in long-term prediction tasks (AUROC = 0.874 vs 0.858 at 5 years), while differences were marginal in 6-months prediction tasks. PerformancFrontiers in digital healthExtending BEHRT to UK Biobank: assessing transformer model performance in clinical prediction.PMC12929515NIHR301413Jenkins DAYildiz YNenadic GJani MfalseExtending BEHRT to UK Biobank: assessing transformer model performance in clinical prediction.<h4>Introduction</h4>Transformer-based models have shown strong potential for clinical prediction using electronic health record data, yet their performance can vary depending on modelling decisions and data characteristics.<h4>Methods</h4>In this study, we trained a BEHRT model on hospital-based UK Biobank data and evaluated its performance across four clinical prediction tasks, including next-visit diagnosis and longer-term diagnosis prediction up to five years. We exhaustively assessed the impact of model size, medical terminology (CALIBER vs ICD-10), and data split strategies.<h4>Results</h4>The large model consistently outperformed the smaller one in long-term prediction tasks (AUROC = 0.874 vs 0.858 at 5 years), while differences were marginal in 6-months prediction tasks. Performanc2026-01-01T00:00:00Z20262026-07-16T22:08:08.881Z2026-07-11T03:11:59.357ZS-EPMC129295154174367510.3389/fdgth.2026.1715506