My role involves a mixture of application and methodology that typifies the work of the Biostatistics Unit. I work with investigators at Royal Papworth Hospital, helping with the statistical aspects of clinical trials. I also spend time developing new methodology, with the aim of improving clinical trial design.

I recently (Nov 2020) submitted a PhD thesis, supervised by Adrian Mander and Michael Grayling. The aims of the work are to investigate improvements to single- and two-arm phase II clinical trials with binary outcomes, as well as multi-outcome multi-stage trials with continuous outcomes. The binary outcome work involves using stochastic curtailment to end trials when the final decision is otherwise not certain, while still preserving the specified type-I and type-II error-rates. The multi-outcome work involves generalising the number of outcomes that need to show promise in order to declare a trial a success. Normally, this is either one outcome or all outcomes, but our work allows any number of outcomes to be used.

I've created the following R packages for anyone who's interested in using these methods: SCsinglearm for stochastic curtailment in single-arm trials; twoarmSC for stochastic curtailment in two-arm trials and moms for multi-outcome multi-stage trials with continuous outcomes.