Meta-Analysis of Diagnostic Accuracy with mada

The R-package mada is a tool for the meta-analysis of diagnostic accuracy. In contrast to univariate meta-analysis, diagnostic meta-analysis requires bivariate models. An additional challenge is to provide a summary receiver operating characteristic curves that seek to integrate receiver operator characteristic curves of primary studies. The package implements the approach of Reitsma, Glas, Rutjes, Scholten, Bossuyt, and Zwinderman (2005), which in the absence of covariates is equivalent to the HSROC model of Rutter and Gatsonis (2001). More recent models by Doebler, Holling, and Böhning (2012) and Holling, Böhning, and Böhning (2012b) are also available, including meta-regression for the first approach. In addition a range of functions for descriptive statistics and graphics are provided.