I understand that one can run a meta-analyis on the effect of a dichotomous outcome variable as per Cochrane Handbook Ch. 10.4. Is it reasonable to apply this approach to prevalence data where you are not contrasting two conditions, but simply try to summarize the proportion of risky events occuring in a single diagnostic group? In other words, if we're looking to determine the prevalence of Knee OA in a specific patient population, and we have multiple studies providing estimates of that dichotomous outcome can we combine them using meta-analysis for dichotomous effects?

BeckyNYC