Join the University of Bristol to make rigorous causal inference from NHS data faster and more accessible. On this NIHR-funded programme, "Efficient estimation of target trials in electronic health records data", you will develop computationally efficient methods and open-source software for target trial emulation — a powerful approach for estimating the effects of health care interventions from observational data. Current software can take days to run on large datasets; our aim is to make these analyses at least ten times faster and put the tools in the hands of researchers worldwide. You will work with Dr Tom Palmer, Dr Paul Madley-Dowd, Dr Venexia Walker and Professor Jonathan Sterne in Bristol, and Dr Will Hulme at the University of Oxford, joining Bristol's Electronic Health Records team in the Department of Population Health Sciences — a leading UK centre for data science and causal inference research.
Hybrid working is available - it would be preferable if the postholder could attend the EHR team meetings in person on a Tuesday and be in the office 2 days per week as a minimum.
You will develop efficient methodology for sequential target trial emulation — including subsampling strategies, scalable bootstrap inference for large datasets, variable-length time intervals and vectorised algorithms — and implement it in well-engineered, open-source R and Python packages (including SEQTaRget and pySEQTarget), with clear interfaces, documentation and diagnostics. You will benchmark existing R, Python and SAS implementations, run Monte Carlo simulation studies to evaluate statistical and computational performance, and apply the methods to comparative-effectiveness analyses using linked NHS electronic health records, including within the OpenSAFELY platform. You will work openly and reproducibly using Git and GitHub, collaborate with colleagues across Bristol and the UK, and disseminate your work through publications, conferences and software releases. There is scope to contribute to research funding bids and, if you wish, to teaching and supervision.
You have a strong quantitative background in medical statistics, epidemiology, data science or a related discipline, and you enjoy programming. You have experience of statistical software such as R, value open and reproducible research, and want to build high-quality tools that other researchers will use. You can develop original solutions and work independently while contributing to a team. Experience of Python, R and Python package development, Git/GitHub, SQL, causal inference methods, or NHS data sources (such as Hospital Episode Statistics) would be an advantage.
The post can be appointed at Research Associate or Senior Research Associate level. For the Senior role, we also seek a relevant PhD (or equivalent experience) and a track record of peer-reviewed publications; the Research Associate role suits those earlier in their research career with an MSc (or equivalent) and working towards a PhD in a relevant quantitative subject. If you don't meet every desirable criterion but are excited by the work, we encourage you to apply.
Contract type: Open-ended with funding for 3 years from start date (ideally, between 01/10/2026-30/09/2029)
Work pattern: Full time/ 1 FTE
Grade: I/J (Pathway 2)
Salary: £39,906 - £44,746 (Grade I) or £43,482 - £50,253 (Grade J)
School/Unit: Bristol Medical School
Shift pattern: 35 hours per week
This advert will close at 23:59 UK time on 19/07/2026
For informal queries please contact: Dr Tom Palmer (Senior Lecturer in Biostatistics Applied to Genetics): 0117 4559644, [email protected]
We recently launched our strategy to 2030 tying together our mission, vision and values.
