The overarching theme of this research is to develop methods which improve the decision-making on when and when not to apply test research in practice.
Previous research conducted by Health Informatics (HI) has demonstrated that primary studies and meta-analyses which evaluate a test's accuracy have the potential to produce estimates which are implausible for some settings. As a result, the group has developed a more selective approach to meta-analysis, "a tailored meta-analysis model", when estimating the accuracy in a particular setting. Furthermore, HI has begun to develop quantitative methods that test the statistical validity of diagnostic meta-analysis results. These latter methods will affect test evaluation reviews in general and the tailored approach will supplement them.
However, the methodologies behind tailored meta-analysis and statistical validity require further development and evaluation and this forms the focus of the current research. An important part of the methodological research will include the exploration and development of optimisation algorithms that are able to handle constraints imposed by routine data collected from the clinical setting of interest. Ultimately the output of this research will lay the foundation for a decision support tool that helps inform clinicians on the expected performance of diagnostic tests used in practice.