Statistical analysis of time to event data with incomplete event status

Supervisors:

Prof Alex McConnachie, School of Health & Wellbeing
Prof Colin Berry, School of Cardiovascular & Metabolic Health
Prof Helen Minnis, School of Health & Wellbeing

Summary:

Event-driven, randomised clinical trials are powered based on the number of events that occur. At the design stage, event rates can only be estimated, leading to uncertainty in required sample size and study duration. Event rates can be monitored to allow prediction of when to stop recruitment and follow-up and ensure sufficient events will accrue.

During a trial, there is a mixture of confirmed study endpoints, and potential endpoints which are awaiting adjudication. This adds complexity to the projection of when best to cease recruitment and follow-up. There are similar issues in event-driven trials with interim analyses to potentially stop the trial if there is sufficient evidence of efficacy or futility.

This PhD aims to develop and validate methods address these challenges using real-world clinical trial data and simulation studies. Training in complex survival analysis methods, particularly in relation to interim analyses, model prediction, and simulation will be provided.