Hi everyone, I'm quite new to Stata 16 and I've acquired it primarily for its powerful meta-analytical capabilities. However, I am wondering how a meta-regression may be done to explore the correlation between an outcome and a variable that affects both treatment and control arms (e.g. age) An example for illustration follows:

A meta-analysis is done to evaluate the effect of a drug in heart failure. Population A takes the drug, while population B doesn't. Now, the two arms' respective age were labelled AgeMean_A, AgeSD_A, AgeMean_B, and AgeSD_B for the mean and SD of populations A and B respectively. The mortality and survival were labelled as Mortality_A, Survival_A, Mortality_B, Survival_B. Sample sizes of population A and B were labelled as N_A and N_B respectively.

Suppose I want to investigate mortality/survival may be affected by age, how should I do it? I quite like the simplified graphical UI that comes with Stata 16 for meta-analysis, but it seems to only allow me to input the age of each arm separately for meta-regression and regress as separate variables (which obviously makes no sense).

Or is it the case that if I want to do what I do, I need to manually / separately compute the difference between ages of the two groups (or, if I use binary outcomes, odds ratios), and then input the computed difference / OR for metaregression?

Thanks. I hope I made myself clear.

More Jeffrey Shi Kai Chan's questions See All
Similar questions and discussions