1
$\begingroup$

I'm stuck trying to find a way to compare whether or not a given intervention has a greater effect in one group (Kids < 5) than another group (Kids > 5). The outcome I'm comparing is the incidence of diarrhea in children receiving probiotics compared to controls. It is a meta-analysis involving 11 studies and I have decided to specify several subgroups to compare (eg. studies with reported mean ages < 5 and those with mean ages > 5).

Both groups show a statistically significant reduction in incidence between treatment and control. There is also a statistically significant difference in the incidence between the two treatment groups (as well as control groups). Ie. the incidence of diarrhea is higher in studies reporting mean age < 5 years.

I used MedCalc Meta-analysis: Proportions to find pooled rates then used chi-square test for equality of proportions to compare.

What I want to know and can't figure out is whether or not probiotics are more effective when used in the higher risk group compared to when their used in the lower risk group.

Thanks for any and all help, I'm truly stuck on this one.

$\endgroup$
1
$\begingroup$

The concept for which you are searching is meta-regression. I do not use the software you mention so I cannot tell you how to do it but meta-regression is widely available in standard software like R or Stata.

Having said that you do have an inferential issue here as the regression will be an ecological one. You are not able with what you have to see the effect of age as you would need individual ages for that. What you are studying is the effect of being enrolled in a study with average age less than 5. This is not the same as being under 5 yourself. Whether such an analysis would answer your scientific question is a matter for you after further examination of information about the age distributions in the studies.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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