Identifying risk factors for infections: the role of meta-analyses

Systematic review and meta-analysis are assuming greater importance in influencing policy makers and clinical opinion worldwide. Many discussions and publications have considered the merits of meta-analysis of epidemiologic data.50 Some observers suggest that meta-analysis of observational studies should be abandoned altogether. 51 In the authors' opinion, statistical combination of observational studies should not be a primary goal of a review. Analysis of heterogeneity between longitudinal studies, however, would provide more insights than mathematical calculation of the summary risk. Meta-analyses of risk factors for infections should strictly follow guidelines for meta-analyses of observational studies. A study protocol should be written in advance, completed literature searches performed, and studies selected in a reproducible and objective fashion. Biologic plausibility must be addressed. The reported findings should be interpreted with caution, taking into account the limitations of various methodologic aspects of risk factors studies. An important role for meta-analyses in this field would be to clarify hypotheses to be formulated for future Identifying Risk Factors for Infections 221 studies and stress limitations strictly related to studies on risk factors for bacterial and viral infections.