Selected work

Representative Papers

Foresight-England: Development of a National-Scale Generative AI Model of Electronic Health Records for Medical Event Prediction across the COVID-19 Pandemic

Aug 17, 2026

This study addresses the challenges of generative modeling and epidemic event prediction using national-scale Electronic Health Records (EHRs) by constructing the first national-level EHR foundation model. Employing a 243-million-parameter Transformer decoder, the model performs autoregressive learning on data from 61 million individuals through fine-grained clinical code tokenization, enabling zero-shot inference and joint temporal representation. Although quantitative results are currently unavailable due to data access restrictions, this work releases a comprehensive methodological template and evaluation framework. Consequently, it establishes a reproducible paradigm and innovative benchmark for large-scale generative EHR modeling and healthcare event forecasting, facilitating future research in this critical domain despite present empirical limitations.

0 citationsRead paper

Longitudinal wearable monitoring and polygenic risk for incident major depressive disorder in the All of Us Research Program

Aug 06, 2026

This study addresses the integration of genetic susceptibility and real-world behavioral dynamics to improve risk prediction for major depressive disorder (MDD). By combining polygenic risk scores (PRS), electronic health records, and longitudinal behavioral data from Fitbit wearable devices, the authors employ time-varying Cox regression models to examine the joint and interactive effects of PRS and dynamic behavioral features—such as daily step count and sleep stability—on MDD incidence. This work presents the first real-world implementation of a joint gene–digital-phenotype modeling framework, achieving an increase in model C-index from 0.637 to 0.705. Notably, behavioral factors exhibited stronger associations with MDD risk among individuals with high PRS, offering empirical support for genetically informed, personalized prevention strategies.

0 citationsRead paper

Recent publications

Latest Papers

Foresight-England: Development of a National-Scale Generative AI Model of Electronic Health Records for Medical Event Prediction across the COVID-19 Pandemic

Aug 17, 2026

This study addresses the challenges of generative modeling and epidemic event prediction using national-scale Electronic Health Records (EHRs) by constructing the first national-level EHR foundation model. Employing a 243-million-parameter Transformer decoder, the model performs autoregressive learning on data from 61 million individuals through fine-grained clinical code tokenization, enabling zero-shot inference and joint temporal representation. Although quantitative results are currently unavailable due to data access restrictions, this work releases a comprehensive methodological template and evaluation framework. Consequently, it establishes a reproducible paradigm and innovative benchmark for large-scale generative EHR modeling and healthcare event forecasting, facilitating future research in this critical domain despite present empirical limitations.

0 citationsRead paper

Longitudinal wearable monitoring and polygenic risk for incident major depressive disorder in the All of Us Research Program

Aug 06, 2026

This study addresses the integration of genetic susceptibility and real-world behavioral dynamics to improve risk prediction for major depressive disorder (MDD). By combining polygenic risk scores (PRS), electronic health records, and longitudinal behavioral data from Fitbit wearable devices, the authors employ time-varying Cox regression models to examine the joint and interactive effects of PRS and dynamic behavioral features—such as daily step count and sleep stability—on MDD incidence. This work presents the first real-world implementation of a joint gene–digital-phenotype modeling framework, achieving an increase in model C-index from 0.637 to 0.705. Notably, behavioral factors exhibited stronger associations with MDD risk among individuals with high PRS, offering empirical support for genetically informed, personalized prevention strategies.

0 citationsRead paper