Meta-analysis of data from studies of diagnostic test accuracy is more complicated than most other forms of meta-analysis (principally due to the paired nature of the main outcome measures sensitivity/specificity). As such the early involvement of a statistician is advisable.