Article type
Year
Abstract
Background: All existing diagnostic test accuracy meta-analysis methods focus on a single index test, and are fundamentally noncomparative in nature.
Objectives: We develop novel methods for the joint meta-analysis of comparative studies of diagnostic accuracy that apply two or more tests in the same participants.
Development of methods: We extend the bivariate meta-analysis method proposed by Reitsma et al. (J Clin Epidemiol 2005; 58(10):982–990) to simultaneously meta-analyze M ≥ 2 index tests. We derive and present formulas for calculating the within-study correlations between the sensitivities and between the specificities of the tests under study using data reported in the studies themselves. The proposed methods respect the natural grouping of data by studies, account for the within-study correlation between the sensitivities and between the specificities of the tests (induced because tests are applied to the same participants), allow for between-study correlations between sensitivities and specificities of the compared tests (such as those induced by threshold effects), and calculate asymptotically correct confidence intervals for summary estimates and for differences between summary estimates.
Application: Published meta-analysis of 11 studies on the screening accuracy of detecting trisomy 21 (Down syndrome) in liveborn infants using two tests: shortened humerus (arm bone), and shortened femur (thigh bone).
Results: In the application, separate and joint meta-analyses yielded very similar estimates: For example, the summary sensitivity for a shortened humerus was 38.4% (95%CI: 29.7–47.9%) with the novel method, and 38.8% (30.2–48.1) when shortened humerus was analyzed on its own. However, when calculating the difference in sensitivities or the difference in specificities of the two tests, the novel method yielded tighter confidence intervals compared to separate analyses.
Conclusions: The joint meta-analysis of multiple tests is feasible. It may be preferable over separate analyses for estimating measures of comparative accuracy of diagnostic tests.
Objectives: We develop novel methods for the joint meta-analysis of comparative studies of diagnostic accuracy that apply two or more tests in the same participants.
Development of methods: We extend the bivariate meta-analysis method proposed by Reitsma et al. (J Clin Epidemiol 2005; 58(10):982–990) to simultaneously meta-analyze M ≥ 2 index tests. We derive and present formulas for calculating the within-study correlations between the sensitivities and between the specificities of the tests under study using data reported in the studies themselves. The proposed methods respect the natural grouping of data by studies, account for the within-study correlation between the sensitivities and between the specificities of the tests (induced because tests are applied to the same participants), allow for between-study correlations between sensitivities and specificities of the compared tests (such as those induced by threshold effects), and calculate asymptotically correct confidence intervals for summary estimates and for differences between summary estimates.
Application: Published meta-analysis of 11 studies on the screening accuracy of detecting trisomy 21 (Down syndrome) in liveborn infants using two tests: shortened humerus (arm bone), and shortened femur (thigh bone).
Results: In the application, separate and joint meta-analyses yielded very similar estimates: For example, the summary sensitivity for a shortened humerus was 38.4% (95%CI: 29.7–47.9%) with the novel method, and 38.8% (30.2–48.1) when shortened humerus was analyzed on its own. However, when calculating the difference in sensitivities or the difference in specificities of the two tests, the novel method yielded tighter confidence intervals compared to separate analyses.
Conclusions: The joint meta-analysis of multiple tests is feasible. It may be preferable over separate analyses for estimating measures of comparative accuracy of diagnostic tests.