Synthesizing evidence of diagnostic studies comes with more complications than synthesizing evidence of clinical trials. This is because the performance of diagnostic tests is evaluated for participants who have the target condition as well as for participants who do not have the target condition, usually in terms of sensitivity and specificity, respectively. There is a natural trade-off between sensitivity and specificity as lowering the threshold, at which participants will be diagnosed as positive, will increase sensitivity but at the same time reduce specificity. Thus, appropriate methods for diagnostic meta-analysis deal with pairs of sensitivity and specificity to preserve the bivariate nature of diagnostic accuracy. In the present chapter, we present a number of approaches to diagnostic meta-analysis and focus on the most commonly applied methods that are able to incorporate systematic variation between studies in addition to differences in the applied thresholds.
I want to thank Prof. Philipp Doebler and Prof. Gerta Rücker for their very helpful comments on this chapter.
- 1.Macaskill P, Gatsonis C, Deeks J, Harbord R, Takwoingi Y. Cochrane handbook for systematic reviews of diagnostic test accuracy. London: The Cochrane Collaboration; 2010.Google Scholar
- 7.Van Houwelingen HC, Zwinderman KH, Stijnen T. A bivariate approach to meta-analysis. Stat Med. 1993;12:2273–84.Google Scholar
- 8.Van Houwelingen HC, Arends LR, Stijnen T. Advanced methods in meta-analysis: multivariate approach and meta-regression. Stat Med. 2002;21:589–624.Google Scholar
- 11.Sweeting MJ, Sutton AJ, Lambert PC. What to add to nothing? Use and avoidance of continuity corrections in meta-analysis of sparse data. Stat Med. 2004;23:1351–75.Google Scholar