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.

Nomenclature Ontology for Medical And Disease names (NOMAD): taxonomy of types and origins of disease names

The nomenclature of human disease has developed organically over the past centuries using Greek, Latin, and Arabic terminology and reflects the idiosyncrasies of different eras of medical discovery. Despite evident heterogeneity in naming practices, no systematic framework exists for characterising these conventions across all diseases. In this paper, we describe the Nomenclature Ontology for Medical And Disease names (NOMAD), a meta-taxonomy that classifies disease names according to their naming conventions. We developed a two-level taxonomy comprising 9 top-level categories and 20 subcategories and applied it to 22,548 index entries from the ICD-10-CM 2026 Alphabetical Index in a scalable three-stage machine learning-driven classification pipeline. Classification was multi-label, reflecting the compositional nature of medical nomenclature. We classified 99.1% of terms with a mean of 2.12 labels per entry. Anatomical categories were the most prevalent (63.8% of entries), followed by Descriptive (48.4%) and Pathophysiological (40.2%), while Eponymous and Geographical labels were less common than their cultural prominence might suggest (9.7% and 1.9% respectively). Among all Eponymous diseases, we identified only 57 (2.6%) of diseases named after a female person. We manually reviewed a random sample of n=2,255 entries (10%) for accuracy and calculated a full agreement rate of 70% and partial agreement rate of 26% (macro-averaged Cohen's Kappa score 0.832). Naming convention profiles varied substantially across ICD-10-CM chapters, reflecting specialty-specific epistemological traditions: infectious disease chapters were dominated by etiological labels and showed the highest proportion of geographical region related labels, the circulatory chapter by anatomical and pathophysiological labels, and mental and behavioural disorders showed the highest prevalence of socio-behavioral labels.

Machine learning and AI research for Patient Benefit: 20 Critical Questions on Transparency, Replicability, Ethics and Effectiveness

Machine learning (ML), artificial intelligence (AI) and other modern statistical methods are providing new opportunities to operationalize previously untapped and rapidly growing sources of data for patient benefit. Whilst there is a lot of promising research currently being undertaken, the literature as a whole lacks: transparency; clear reporting to facilitate replicability; exploration for potential ethical concerns; and, clear demonstrations of effectiveness. There are many reasons for why these issues exist, but one of the most important that we provide a preliminary solution for here is the current lack of ML/AI- specific best practice guidance. Although there is no consensus on what best practice looks in this field, we believe that interdisciplinary groups pursuing research and impact projects in the ML/AI for health domain would benefit from answering a series of questions based on the important issues that exist when undertaking work of this nature. Here we present 20 questions that span the entire project life cycle, from inception, data analysis, and model evaluation, to implementation, as a means to facilitate project planning and post-hoc (structured) independent evaluation. By beginning to answer these questions in different settings, we can start to understand what constitutes a good answer, and we expect that the resulting discussion will be central to developing an international consensus framework for transparent, replicable, ethical and effective research in artificial intelligence (AI-TREE) for health.

Application of Clinical Concept Embeddings for Heart Failure Prediction in UK EHR data

Electronic health records (EHR) are increasingly being used for constructing disease risk prediction models. Feature engineering in EHR data however is challenging due to their highly dimensional and heterogeneous nature. Low-dimensional representations of EHR data can potentially mitigate these challenges. In this paper, we use global vectors (GloVe) to learn word embeddings for diagnoses and procedures recorded using 13 million ontology terms across 2.7 million hospitalisations in national UK EHR. We demonstrate the utility of these embeddings by evaluating their performance in identifying patients which are at higher risk of being hospitalised for congestive heart failure. Our findings indicate that embeddings can enable the creation of robust EHR-derived disease risk prediction models and address some the limitations associated with manual clinical feature engineering.

Evaluation of Semantic Web Technologies for Storing Computable Definitions of Electronic Health Records Phenotyping Algorithms

Electronic Health Records are electronic data generated during or as a byproduct of routine patient care. Structured, semi-structured and unstructured EHR offer researchers unprecedented phenotypic breadth and depth and have the potential to accelerate the development of precision medicine approaches at scale. A main EHR use-case is defining phenotyping algorithms that identify disease status, onset and severity. Phenotyping algorithms utilize diagnoses, prescriptions, laboratory tests, symptoms and other elements in order to identify patients with or without a specific trait. No common standardized, structured, computable format exists for storing phenotyping algorithms. The majority of algorithms are stored as human-readable descriptive text documents making their translation to code challenging due to their inherent complexity and hinders their sharing and re-use across the community. In this paper, we evaluate the two key Semantic Web Technologies, the Web Ontology Language and the Resource Description Framework, for enabling computable representations of EHR-driven phenotyping algorithms.

Evaluation of Machine Learning Methods to Predict Coronary Artery Disease Using Metabolomic Data

Metabolomic data can potentially enable accurate, non-invasive and low-cost prediction of coronary artery disease. Regression-based analytical approaches however might fail to fully account for interactions between metabolites, rely on a priori selected input features and thus might suffer from poorer accuracy. Supervised machine learning methods can potentially be used in order to fully exploit the dimensionality and richness of the data. In this paper, we systematically implement and evaluate a set of supervised learning methods (L1 regression, random forest classifier) and compare them to traditional regression-based approaches for disease prediction using metabolomic data.