Joint modelling of latent trajectories for dynamic prediction of competing outcomes in patients with liver disease

Liver disease is a common cause of illness and death that is increasing in western countries such as the UK and Denmark, particularly for people under 65 years old. However, the population of liver disease patients consists of people who differ with respect to both their cause of disease, and additional factors such as co-existing conditions and their general medical history. Consequently, disease progression may vary substantially amongst patients: some patients may die early from their liver disease, others might be more at risk of death from causes unrelated to liver disease, whereas for other patients their liver disease does not impact on their overall survival.

 Modelling links between patient data on past and current medical history (such as blood test measurements collected over time) and liver disease progression helps clinicians to identify variables associated with different disease patterns and, ultimately, will allow predictions to be made for key patient outcomes such as survival/mortality.

 This studentship will combine novel methods in dynamic prediction, longitudinal data analysis, survival and competing risk models, and latent class modelling within a probabilistic framework. The goal is to develop methods to significantly change how we care for patients with liver disease by identifying whether patients would benefit from specialist care, which specialty, and when or when not to intervene within the reality of a resource-stretched healthcare system.

This studentship will be based in the Nottingham NIHR Biomedical Research Centre and will include access to large longitudinal population-based cohorts with liver disease, including test results, patient-level variables, and outcomes during follow-up. The collaboration between the University of Nottingham (UK) and Aarhus University (Denmark) places the student in world-leading centres for routine health data analysis, with extensive experience in implementing novel clinical epidemiological methods to study disease occurrence, aetiology, outcomes, and real-world impact. The student will be well supported by a supervisory team that consists of Professors in hepatology and epidemiology, as well as experts in computational statistics and model fitting using routine health care data.

The student will learn techniques in:

Epidemiology – Training to design studies that can draw meaningful conclusions from real-world, large-scale routine healthcare datasets.

Statistics – Developing cutting-edge expertise in longitudinal data analysis, survival modelling and latent class models.

Computational statistics – Training in programming and implementing statistical methodology.

Hepatology – Knowledge of chronic liver disease, its causes, outcomes, and treatments.

Start date: October 2024