This theme has a research programme that encompasses both applied health research and the development and evaluation of statistical methods. Current methodological topics include risk modelling of health outcomes, modelling clustered data, analysis of health economic data, meta-analysis, missing data, evidence synthesis and Bayesian methods.
Much of the applied research is carried in collaboration with health researchers within the UCLH/UCL Biomedical Research Centre (CBRC) and the PRIMENT CTU. The NIHR grant for the BRC has enabled the formation of a Biostatistics Group which sits across the UCL Statistics Department and the UCL faculty of Life and Biomedical Sciences. Applied research projects include developing the first risk model for heart valve surgery, developing a severity scoring system for multiple sclerosis, validating risk models for cardiac surgery, measuring variation in general practice outcomes, studying cost effectiveness of decisions on treatments for angina pectoris, and randomised trials in health services research and drug development.
|Rumana Omar* (Theme Lead)||Clustered data; missing data; risk prediction models; trials methodology|
|Gareth Ambler*||Clinical trials; missing data; multivariate outcomes; risk prediction|
|Gianluca Baio*||Bayesian analysis; health economic evaluation; observational data|
|Julie Barber*||Dementia research; design and analysis of randomised trials; health economic evaluation|
|Tom Fearn||Measurement error|
|Valerie Isham||Dynamics of processes on networks; epidemics and other population processes; models for infection transmission dynamics; point processes|
|Giampiero Marra*||Copula modelling; HIV prevalence estimation|
|Aidan O'Keeffe*||Dynamic causal inference; regression discontinuity designs; time-to-event and Markov multi-state modelling|
|Ardo van den Hout*||Change-point models; joint models; longitudinal data analysis; multi-state models; survival analysis|
*These theme members are currently accepting applications for PhD supervision
Heart valve surgery has an associated in-hospital mortality rate of 4% to 8%. This project developed a simple risk model to predict the risk of in-hospital mortality for patients undergoing heart valve surgery, to provide information to patients and clinicians and facilitate institutional comparisons. Data on 16,679 patients, obtained from the Society of Cardiothoracic Surgeons of Great Britain and Ireland, were used to develop the risk model. Data on a further 16,160 patients were used to evaluate its performance.
Other Research Output
Biostatistics Web Lectures
The Biostatistics Group has created web lectures to provide an introduction to quantitative methods for health researchers and postgraduate students, and to provide a refresher or revision course for those researchers who have previously attended a module or short course. Funding for these lectures was provided by the Education Theme of the UCLH/UCL Biomedical Research Centre.