Network meta-analysis for comparing diagnostic test accuracy

About the Project

Often there is more than one medical test to diagnose a particular condition, for example, different brands of lateral flow test for COVID-19. Different studies often give slightly different estimates of how accurate each test is, and some studies directly compare alternative tests while others just evaluate one test. To make best use of this information, statisticians need to analyse the data from different studies together (known as ‘evidence synthesis’). This produces our ‘best’ estimates of how accurate each test is overall, and how the accuracy of different tests compare with each other. Decision-makers such as the National Institute for Health and Care Excellence (NICE) use results from these analyses to guide which test(s) should be used for patients.

Network meta-analysis (NMA) has revolutionised decision-making when multiple treatments exist for a disease. A huge body of NMA research has been a carried out, and consequently, NMA is routinely used to compare treatments [1-6].

Diagnostic test accuracy (DTA) reviews often compare tests but use methods that have limitations, may be prone to bias, and have not been fully evaluated [7-10]. NMA of DTA studies could transform decision-making regarding tests, by enabling the comparative accuracy of all tests to be estimated, even when not all pairs of tests have been compared against each other within studies.

There is tremendous scope, and urgent need, for methods development and evaluation to compare test accuracy in evidence syntheses.

Aims and Objectives

The general aim is to advance methods for comparing diagnostic test accuracy in evidence synthesis. More specific aims will be decided with the chosen applicant.

Methodology

This methodological project will be tailored together with a suitable candidate based on their current skills and skills/knowledge that they would like to develop. The project can be tailored to be a highly statistical requiring a student with strong quantitative skills or be more qualitative requiring a student with familiarity or interest in reviewing systematic reviews and diagnostic test studies. The student will be supported by experts in this area and have access to training and development funds.

Supervisor:

Dr Sarah Donegan - donegan@liv.ac.uk

Dr Maria Sudell - M.E.Sudell@liv.ac.uk

Prof Catrin Tudur Smith - Cat1@liv.ac.uk

To apply for the position, please email Dr Sarah Donegan (donegan@liv.ac.uk) attaching a covering letter, CV and details of 2 referees.