arXiv subjects

Angela M. Wood

Publications and source records attributed to Angela M. Wood.

4 recordsLinked to original sources

arXiv2026

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

Foresight-England (Foresight-E) is the first national-scale generative foundation model of electronic health records (EHRs), developed as a research pilot strictly for COVID-19 research. We evaluated its ability to model the direct and indirect effects of the pandemic. Trained from scratch entirely within the NHS England Secure Data Environment, Foresight-E is a 243-million-parameter transformer decoder. It was trained and evaluated on de-identified, longitudinal EHRs of approximately 61 million individuals, integrating primary/secondary care, death registrations, and COVID-19 data. Training and validation used a 90% subset (54.9 million) spanning November 2018 to December 2022; the remaining 10% (6.1 million) was held out for evaluation. Foresight-E models patient timelines autoregressively, predicting the next medical event given their prior history. At inference, it operates zero-shot, predicting any concept in its ~40,000-code vocabulary without task-specific training. Our tokenisation scheme retains the clinical granularity of ICD-10, OPCS-4, and SNOMED CT codes, jointly representing absolute and relative timing. We designed an evaluation framework for 30-day COVID-19 hospitalisation and mortality, including subgroup analyses by demographic factors and vaccination status. To assess generalisation to unseen future data and the pandemic's indirect effects, we tested the model on medical events from 2023 (beyond its training period), benchmarking against logistic regression and XGBoost. As detailed in the Project Status section, NHS England has paused access to data for the Foresight-E project, meaning quantitative results are currently unavailable. Instead, we share our strategy for tokenisation, architecture, training, inference, and evaluation as a methodological template and case study in the challenges of building population-scale EHR foundation models.

cs.LG↗

arXiv2023

Sequential Re-estimation Learning of Optimal Individualized Treatment Rules Among Ordinal Treatments with Application to Recommended Intervals Between Blood Donations

Personalized medicine has gained much popularity recently as a way of providing better healthcare by tailoring treatments to suit individuals. Our research, motivated by the UK INTERVAL blood donation trial, focuses on estimating the optimal individualized treatment rule (ITR) in the ordinal treatment-arms setting. Restrictions on minimum lengths between whole blood donations exist to safeguard donor health and quality of blood received. However, the evidence-base for these limits is lacking. Moreover, in England, the blood service is interested in making blood donation both safe and sustainable by integrating multi-marker data from INTERVAL and developing personalized donation strategies. As the three inter-donation interval options in INTERVAL have clear orderings, we propose a sequential re-estimation learning method that effectively incorporates "treatment" orderings when identifying optimal ITRs. Furthermore, we incorporate variable selection into our method for both linear and nonlinear decision rules to handle situations with (noise) covariates irrelevant for decision-making. Simulations demonstrate its superior performance over existing methods that assume multiple nominal treatments by achieving smaller misclassification rates and larger value functions. Application to a much-in-demand donor subgroup shows that the estimated optimal ITR achieves both the highest utilities and largest proportions of donors assigned to the safest inter-donation interval option in INTERVAL.

stat.ME↗

arXiv2023

Patient stratification in multi-arm trials: a two-stage procedure with Bayesian profile regression

Precision medicine is an emerging field that takes into account individual heterogeneity to inform better clinical practice. In clinical trials, the evaluation of treatment effect heterogeneity is an important component, and recently, many statistical methods have been proposed for stratifying patients into different subgroups based on such heterogeneity. However, the majority of existing methods developed for this purpose focus on the case with a dichotomous treatment and are not directly applicable to multi-arm trials. In this paper, we consider the problem of patient stratification in multi-arm trial settings and propose a two-stage procedure within the Bayesian nonparametric framework. Specifically, we first use Bayesian additive regression trees (BART) to predict potential outcomes (treatment responses) under different treatment options for each patient, and then we leverage Bayesian profile regression to cluster patients into subgroups according to their baseline characteristics and predicted potential outcomes. We further embed a variable selection procedure into our proposed framework to identify the patient characteristics that actively "drive" the clustering structure. We conduct simulation studies to examine the performance of our proposed method and demonstrate the method by applying it to a UK-based multi-arm blood donation trial, wherein our method uncovers five clinically meaningful donor subgroups.

stat.ME↗

arXiv2023

Optimal risk-assessment scheduling for primary prevention of cardiovascular disease

In this work, we introduce a personalised and age-specific Net Benefit function, composed of benefits and costs, to recommend optimal timing of risk assessments for cardiovascular disease prevention. We extend the 2-stage landmarking model to estimate patient-specific CVD risk profiles, adjusting for time-varying covariates. We apply our model to data from the Clinical Practice Research Datalink, comprising primary care electronic health records from the UK. We find that people at lower risk could be recommended an optimal risk-assessment interval of 5 years or more. Time-varying risk-factors are required to discriminate between more frequent schedules for higher-risk people.

stat.AP↗