If you are reviewing both randomized and non-randomized studies in the same meta-analysis, report the effect sizes separately for the RCTS and non-randomized studies. This is because the effect of a new treatment is likely to be larger in studies that employ a non-randomized design.

Try to be as up-to-date as possible. Mention the cut-off dates in the analysis so that it becomes clear that studies that were published before and after the time mentioned were not missed but not included in the study as part of the design. If you are including only English language articles in the study, mention the rationale for doing so.

Importantly, if you are faced with multiple reports published from the same study on similar topic, include only one study so that information from the same study population should contribute ONLY ONCE to the analysis.

Try to exclude studies with small sample size to avoid overemphasizing small studies. Similarly, make an early decision about the period of follow up of studies. More of these in the sensitivity analysis.

Decide early on whether you are going to include studies with similar exposures or outcomes in your analysis. Base this decision on your experience and understanding of the research question. Remember, if you are too restrictive, you gain face validity, though you may end up omitting important studies.