Appropriate statistical methodology for the design and analysis of data of research studies underpins much medical research work. Medical statisticians at WMS conduct research in applied statistical methodology as well as work on clinical trials and research projects in collaboration with other members of WMS, NHS trusts, NICE, universities and industry partners.
The Statistics and Epidemiology group is interested in developing innovative design for clinical trials. We focus on the development of oncology trial that investigates more than one treatment and/or cancer type simultaneously (master protocols, basket, and umbrella designs), new estimators for adaptive clinical trials that consist of an interim analysis to perform treatment selection or biomarker-driven subgroup selection and trial design for small population with Bayesian decision theory framework.
We are also interested in outbreak detection of syndromes and surveillance for genome sequenced organisms and in using machine learning, geometric and topological approaches to multidimensional data in cancer therapeutics.
The core research areas include the methodology of network meta-analysis and (individual participant data) meta-analysis. We are also interested in the application of statistical approaches within Health Technology Assessment, specifically advanced methods of survival analysis.
Our primary research interests in the analysis of observational studies include the methodology and application of Self-Controlled Case Series (SCCS), case-crossover, cohort and case control methods. We are interested in the methodological development of the SCCS method.
We also developed and maintain R packages:
- asd, which is a package that runs simulations for adaptive seamless designs with and without early outcomes for treatment selection and subpopulation type designs.
- repolr, which is a package that fits linear models to repeated ordinal scores using GEE methodology.
- SCCS, which is a package for self-controlled case series models.