Using observational data in model-based economic evaluations of stratified medicine in rheumatology

Lead Research Organisation: University of Manchester
Department Name: School of Health Sciences


Stratified (personalised/precision) medicine is promising from an economic perspective since in principle only those patients likely to benefit will receive a treatment. A recurring methodological challenge in economic evaluation of stratified medicine is a result of gaps in the data and evidence. A particular challenge occurs when evaluating a companion diagnostic test to predict the risk of a rare adverse drug reaction (ADR), which precludes a randomised clinical trial (RCT) on the basis that the required sample size to achieve adequate power makes the design unfeasible. Furthermore, the data required to evaluate how the companion diagnostic test can be moved from the research to clinical environment, taking account of the optimum level of service delivery, cannot be readily identified from a RCT. To date, however, there has been limited attention paid to merging the disciplines of statistics and economics to develop ways of using large scale datasets effectively to evaluate the added value of stratified medicine.

Hypothesis: observational data can provide a robust source of data to populate model-based
economic evaluations of a companion diagnostic to identify patients at risk of an adverse drug reaction from a treatment for patients with rheumatoid arthritis (RA).

(1) structure a model-based economic evaluation of a companion diagnostic test to stratify patient populations with RA at risk of an ADR from anti-rheumatic treatment
(2) populate the model with observational data to identify the incremental costs and benefits of the companion diagnostic
(3) identify the optimal level of service delivery given the constraints within the healthcare system.

Methods to be used during the project are:
(1) systematic reviews of the clinical and economic literature
(2) model-based cost effectiveness analysis
(3) regression-based statistical methods that take account of confounding and selection bias
(4) mathematical optimisation methods.

Methods to use existing datasets to generate economic evidence to move stratified medicine from the research to clinical environment.

This PhD involves the development of new quantitative methods but also offers a student the opportunity to work with practising clinicians and produce outputs with practical relevance to inform the more effective use of the NHS budget. Importantly, this PhD focuses on a core cross-cutting research theme stratified (precision) medicine both through the selection of the clinical area for the methodological research but also via the involvement of three academics with core roles in current research programmes across the University.

The student will further develop core skills in quantitative research methods including: systematic review and meta-analysis; model-based cost effectiveness analysis; statistical methods in causal inference and mathematical optimisation methods. The student will need to become proficient in the use of statistical software packages (such as R and/or Stata).

All three Centres (health economics, biostatistics and musculoskeletal research) run internal and external seminars providing an excellent means of knowledge exchange between peers within economics and biostatistics but also practising clinicians. The student will have opportunities to attend relevant modules from the University of Manchester's MSc in the Economics of Health, MSc in Statistics, MSc in Health Data Science and MSc in Rheumatology.


10 25 50