Challenges in meta-analyses with observational studies.

In the presence of observational study designs, prior to synthesis, investigators should carefully consider whether all studies at hand are able to answer the same clinical question. The potential for a quantitative synthesis should be guided through examination of the amount of clinical and methodological heterogeneity and assessment of possible biases.