Are there any good references on how to perform an IPD meta-analysis on diagnostic test accuracy data with covariates?