Grant: Operating Grant

Availability of an accurate diagnostic test has been identified as an urgent need in the global battle against tuberculosis (TB). The last two decades have seen a steady growth in research studies on the accuracy of rapid tests for TB screening in vulnerable populations. Systematic reviews and accompanying meta-analyses have played a key role in synthesizing the results of this diverse literature. They have aided in directing medical research towards more promising technologies, and also in identifying failed technologies that have proved detrimental to public health. The last decade has also seen a growing recognition of the need for more sophisticated statistical modeling to support meta-analyses. This proposal describes development of new tools to improve meta-analyses of tuberculosis diagnostic studies. Building on our earlier work in this area, we propose methods for addressing three principal concerns: i) Quantifying and modeling variation across diagnostic studies, ii)Studying the impact of common biases in diagnostic studies, and iii) Estimation of the incremental value of a new test over and above a standard test. Our methods will be applied toe data from published meta-analyses of TB tests. We anticipate that our research will lead to improved methods for evaluating available evidence and aid in policy making. For all models we will develop user-friendly software to permit easy application. The methods we propose will also find application in areas besides tuberculosis where diagnostic meta-analyses are needed.