Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.


Multimodal AI for risk prediction and disentangling cardiovascular multimorbidity


Cardiovascular diseases (CVD) are the number one cause of mortality globally. Additionally, multimorbidity (i.e., the presence of two or more chronic conditions) is on the rise, with recent findings estimating that over a sixth of the UK population will be multimorbid by 2035. Given the critical issues of multimorbidity of cardiovascular (and metabolic) disorders, it is imperative to comprehensively explore facets of risk and protection concerning CV multimorbidity.

In parallel, the development and application of artificial intelligence (AI) models and growing access to multimodal electronic health records (EHR) and other clinical data (e.g., genetic data types, images, etc) have opened the door to unprecedented opportunities for (high-dimensional) data-driven clinical research. On one hand, AI models like Transformers capable of handling various modalities have pushed the bar in terms of predictive performance in numerous population health investigations. Additionally, the data-driven approach underpinning AI has not only boosted predictive power but also paved the way for model-driven knowledge discovery ultimately augmenting our understanding of medicine and incrementally iterating towards “precision medicine”.

This DPhil project is an exciting opportunity within the Deep Medicine group to continue to progress multidimensional research in the space of cardiovascular multimorbidity. The opportunity offers the DPhil student to review the existing space of research and data availabilities and then define a set of hypotheses to concretely explore prediction of risk and the exploration of risk factors in those with multimorbidity (e.g., patients with diabetes, rheumatoid arthritis, and autoimmune conditions) using multimodal AI modelling methods. Utilizing multiple modalities including EHR and image data, the project will require developing multimodal AI models that can flexibly adapt to data modality availability and richness. While generation of hypotheses concerning risk and protection within these high-risk subgroups via interpretability exploration of “black box” AI models will be necessary, the project will importantly close the loop with rigorous evaluation within a formal hypothesis testing (causal inference) framework capturing the full spectrum of evidence-based population health research. This investigation will involve bridging the fields of epidemiology, multimodal AI, and cardiovascular medicine to achieve the ultimate goal of delivering high-impact methodological and clinical research. 

In the space of AI and CVD studies specifically, the Deep Medicine research group at NDWRH led by Professor Kazem Rahimi has been a pioneering force for exploring AI-driven methods for studying CVD risk and etiology utilizing rich UK EHR. From risk prediction to breaking open “black box” AI models to conducting causal inference, the Deep Medicine group has been actively leading the space of multimodal AI for CVD epidemiological research. 

Candidates should have a strong background in mathematics, computer science, statistics, or related engineering fields. Additional experience/training in biomedical sciences and epidemiology is not required but would be preferred.


In terms of skills, the DPhil student with competencies in AI, statistics, or related fields will receive further advanced training in AI modelling for EHR and other data modalities available for study within Deep Medicine, traditional epidemiological methods (e.g., cohort study design, conventional statistical models, etc), and minimally parallel computing framework approaches for efficient exploration of large-scale multimodal datasets.

Additionally, the DPhil student will receive training in conducting comprehensive literature reviews, writing academic papers in peer-reviewed journals and conferences, and will work with a strong inter-disciplinary team of AI scientists, epidemiologists, cardiologists, and statisticians to conduct high-quality doctoral studies

The university also provides a wide-range of training opportunities.

Funding information

Unfunded, however, the Deep Medicine group will assist the DPhil candidate in securing funding.


To apply for this research degree, please click here.