I would like to conduct a meta-analysis, and later a meta-regression, of estimated prevalence rates in 40 observational studies. I'd like to know: a) what the true prevalence rate is likely to be b) whether any of five variations between studies and settings influence the estimated prevalence rate.

Typically, texts on meta-analysis assume that the studies under analysis are clinical trials or similar experimental designs. Discussions are almost always couched in terms of effect sizes. However, I am looking to analyse a proportion of the population, not the effect of a treatment. As a newcomer to meta-analysis, there appears to be little methodological guidance on conducting meta-analysis of observational studies of prevalence rates.

Can I conduct the analyses described above? Can anyone recommend a text for study that covers observational studies of prevalence rates?


1 Answer 1


A day in the library resulted in the discovery of:

Alexander J. Sutton, Keith R. Abrams, David R Jones, Trevor A. Sheldon and Fujian Song (2000). Methods for Meta-Analysis in Medical Research. Wiley: London.

That text treats synthesis observation studies in Chapters 2 and 16.

A somewhat outdated exposition can be found in:

Greenland, S. (1987). QUANTITATIVE METHODS IN THE REVIEW OF EPIDEMIOLOGIC LITERATURE. Epidemiologic Reviews, 9(1), 1–30.

Methods for Meta-Analysis in Medical Research


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.