Figure 8 compares the point estimates and confidence interval widths of summary sensitivity and specificity from bivariate random effects methods using a normal approximation (multivariate DerSimonian-Laird inverse variance vs. maximum likelihood estimation). The point estimates from the two methods are almost identical, however confidence interval widths are often different (generally larger with maximum likelihood).

Figure 8Comparison of point estimates and confidence interval widths of summary sensitivity and specificity (logit scale) from bivariate random effects methods using a normal approximation (multivariate DerSimonian-Laird inverse variance vs. MLE)

Note: Scatter plot of estimated logit-transformed sensitivity, specificity and their corresponding confidence interval widths from bivariate random effects meta-analyses using the multivariate DerSimonian-Laird and MLE methods (both with a normal approximation to represent within-study variability).

CI = confidence interval width; mult. DL = multivariate DerSimonian-Laird; MLE = maximum likelihood estimation.

From: Results

Cover of An Empirical Assessment of Bivariate Methods for Meta-Analysis of Test Accuracy
An Empirical Assessment of Bivariate Methods for Meta-Analysis of Test Accuracy [Internet].
Dahabreh IJ, Trikalinos TA, Lau J, et al.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.