Overview

Jack Bowden is a statistician interested in developing statistical tools and methods to support cutting edge biomedical research. In 2019 he was appointed Professor of Biomedical Data Science within the University of Exeter Medical School.

Before moving to Exeter, he led a research programme within the MRC Integrative Epidemiology Unit at the University of Bristol, developing robust approaches for causal inference within observational studies using genetic data.  Prior to that, his work focused on developing methods to preserve the validity and improve the efficiency of clinical trials, within the MRC Biostatistics Unit at the University of Cambridge.

In Exeter, he will primarily support the world class research undertaken into Diabetes. He is keen to apply his expertise to help learn about fundamental causes of this varied condition and its effective treatment, using genetics, observational studies and clinical trials.  He maintains an honorary position at the University of Bristol and hopes to use his existing research connections to further strengthen the collaborative links between these two great institutions.

Qualifications

BSc Hons  Mathematics

MSc. Medical Statistics

PhD  Biostatistics

Back to top

Research

Research interests

Mendelian randomization (MR)

MR uses individuals’ genetic variants as `Instrumental Variables’ (IVs) to turn observational data into a pseudo randomized controlled trial. The technique enables scientists to estimate the causal effect of intervening and modifying a health exposure (such as a person’s diet or body mass index) on downstream disease outcome (such as the risk of developing Type 2 Diabetes).  MR circumvents the problem of confounding that inhibits straightforward causal estimation using observational studies.   Jack has helped to develop a range of statistical methods for MR, as well as methods for sensitivity analysis that help to determine whether individual genetic variants satisfy the IV assumptions, in order for the approach to draw reliable causal conclusions.  

Clinical Trials Methodology

Jack has developed a range of statistical methods to preserve the validity, improve the efficiency and increase the patient benefit of randomized clinical trials. Specific areas of interest include:

• Estimating treatment effects in multi-arm multi-stage trials;

• Accounting for non-adherence and intercurrent events;

• Performing sample size re-estimation;

• Implementing response adaptive randomization;

• Optimizing clinical trial design.

Research projects

Current Projects:

• Adjusting for selection bias and missing data in clinical trials and epidemiology  - in collaboration with Kate Tilling, Matt Tudball and Dea Hazewinkel (Bristol);
• Assessing the causal role of sleep disturbance on Type II diabetes traits – in collaboration with Martin Rutter (Manchester), Deborah Lawlor (Bristol) and Mike Weedon (Exeter);
• Developing Instrumental variable methods to account for intercurrent events in academic and industry-led trials - in collaborations with Jonathan Bartlett (Uni of Bath) and Novartis;
• Adjusting for pleiotropy, weak instrument bias and winner’s curse in Mendelian randomization - in collaboration with Ciarrah Barry (Bath), Chin Yang Shapland and Gibran Hemani,  (Bristol);
• Utilizing gene-environment interactions in Mendelian randomization – in collaboration with Wesley Spiller and George Davey Smith (Bristol);
• Multivariable Mendelian randomization – in collaboration with Eleanor Sanderson (Bristol)

Back to top

Publications

Back to top