Estimating subgroup treatment effects from clinical trials is notably difficult due to the limited sample size and under-representation of patients with multimorbidity. Much larger routinely collected observational data, equipped with bespoke designs and robust analytic approaches, can fill those critical evidence gaps and support more informed clinical decisions, ultimately improving patient outcomes, cutting costs, and boosting healthcare quality.
The successful student will investigate methods for the identification and characterisation of heterogeneity in drug safety and effectiveness in subgroups of patients with varying clinical, socio-demographic, and genetic profiles. We will focus on two classes of medications: antihypertensive and antidiabetic drugs, with the primary considerations that (1) they are frequently used and recorded in available outpatient record/s data, (2) our research team have considerable expertise in phenotyping standardised concepts for the study of these drugs and related health outcomes.
To achieve this, we will use data mapped to the internationally recognised OMOP common data model. OMOP-mapped data from UK's CRPD GOLD and Aurum is hosted in our servers at Oxford University, and additional databases licensed and maintained by collaborators from European Health Data & Evidence Network (EHDEN) and Observational Health Data Sciences and Informatics (OHDSI) networks. Finally, we will use UK Biobank data (also hosted at the University of Oxford) to analyse genetic determinants of the study outcomes.
The study/target population will be categorised into subgroups based on baseline risk predicted using valid models developed using cutting-edge statistical and machine-learning techniques (Objective 1). We will then enrich these clinical-data-based models with genetic predictors, such as monogenic mutations or polygenic risk scores, with data from UK Biobank (Objective 2). Thirdly, we will derive estimates for the prespecified drug-safety-outcome pairs in each subgroup (from the dimension of clinical risk only, genetic risk only, and clinical + genetic risk) based on the causal inference framework, including an active-comparator, new-user cohort design, large-scale propensity score adjustment or matching, and negative control outcome experiments and empirical calibration (Objective 3). Lastly, we will systematically repeat the above processes for the comparative drug-effectiveness-outcome pairs, focusing on comparing first- and second-line drugs for each drug class (Objective 4).
The student will be expected to bring their own ideas to enrich the study.
The PhD student will receive training, support and supervision from academic supervisors in a top world class academic environment at the Centre for Statistics in Medicine, University of Oxford. The student will be part of an established research group, the Pharmaco- and Device epidemiology research group, with access to data, state-of-the-art computational resources and facilities, experience in the proposed research. They will join an exciting group of over >20 staff and 10 PhD and MSc students.
The group will ensure hands-on training in real world data analysis using medical records and genetic data from. The student will work on their unique project within an experienced and collaborative supervisory team. They will also be embedding within our international EHDEN and OHDSI networks to ensure additional analytical guidance, training and support. A student would be supported to attend relevant conferences to enrich their studies and financial support will be made available for travel to conferences.
Alongside with the opportunities listed above, the department offers extra training:
NDORMS hosts the Centre for Statistics in Medicine, a centre committed to improving the standard of medical research methodology through research and training on research and methods development, and home to the UK Equator Centre. This enables and encourages research and education to champion transparent and complete reporting of health research through reporting guidelines and training provision. A core curriculum of lectures will be taken to provide a solid foundation in a broad range of subjects including statistics, epidemiology, and big data analysis. All students will be required to attend a 2-day Statistical and Experimental Design course at NDORMS and the Real World Epidemiology: Oxford Summer School. Students will also be required to attend regular seminars within the Department and have access to a variety of other courses run by the Medical Sciences Division Skills Training Team and the wider University, such as the UK Equator Publication School. Finally, the student(s) will be expected to regularly present data in Departmental seminars, the Pharmco- and device epidemiology group and within our external EHDEN and OHDSI collaborators.
KEYWORDS: Real world data, epidemiology, drug safety, genetics, machine learning
HOW TO APPLY
The Department accepts applications throughout the year but it is recommended that, in the first instance, you contact the relevant supervisor(s) or the Graduate Studies Office (firstname.lastname@example.org) who will be able to advise you of the essential requirements.
Interested applicants should have, or expect to obtain, a first or upper second-class BSc degree or equivalent in a relevant subject and will also need to provide evidence of English language competence (where applicable).
The application guide and form is found online and the DPhil will commence in October 2023.
Applications should be made to one of the following programmes using the specified course code:
D.Phil in Clinical Epidemiology and Medical Statistics (course code: RD_NNRA1).
For further information, please visit http://www.ox.ac.uk/admissions/graduate/applying-to-oxford.