midas is a user-written command for idiot-proof implementation of some of the contemporary statistical methods for meta-analysis of binary diagnostic test accuracy. Primary data synthesis is performed within the bivariate mixed-effects logistic regression modeling framework. Likelihood-based estimation is by adaptive gaussian quadrature using xtmelogit (Stata release 10) with post-estimation procedures for model diagnostics and empirical Bayes predictions. Average sensitivity and specificity (optionally depicted in SROC space with or without confidence and prediction regions), and their derivative likelihood and odds ratios are calculated from the maximum likelihood estimates. midas facilitates exploratory analysis of heterogeneity, threshold-related variability, methodological quality bias, publication and other precision-related biases. Bayes' nomograms, likelihood-ratio matrices, and probability modifying plots may be derived and used to guide patient-based diagnostic decision making. A dataset of studies evaluating axillary staging performance of positron emission tomography in breast cancer patients is provided for illustration of the omnibus capabilities of midas.