# Meta-regression with Mean age in each study

My dataset from Meta analysis looks like below: For example: Stroke as outcome in 9 studies Treatment 1 Treatment 2 Odds Ratio Study Events Total Events Total Weight M-H, Random, 95% CI 1
2
3

This set of information is available for 5 years & 10 years.

In addition to this I have mean ages of patients per each study at the start of the trial. (please note all trials have different start date) For example Study Patients (n) Mean Age Mean followup year Treat. 1 Treat.2 Age(years) Treat.1 Treat.2 Treat.1 Treat.2 1 200 250 (55-70 ) (61+/-5 ) (62.3+/-5) 8+/-2 9+/-3

Now my questions are: 1. Can I fit regression model using this information? 2. I am particularly interested in finding relation between age & odds ratio at different time intervals (for example at 1 year, 5 years, 6 years). So can I use mean age per each study(given as a baseline measure) as predictor & odds ratio per each study at 5 years or 10 years as dependent variable? 3. I would also like to see comparison between each treatment group. So Instead of using odds ratio, what if I calculate odds of each study per each treatment & use it as dependent variable while keeping mean age per study independent variable

So I would like to know if my approach is correct or not? Any help/suggestion is appreciated.

It is a bit difficult to see what your dataset looks like as I think it has been mangled when you posted it. If I understand you correctly from each study you want to use a study level covariate (mean age) and you want to use multiple outcomes for each study measured at different time points. You presumably have the proportion of people who suffered a stroke as outcome. In that case you can certainly do this as a multi-level meta-regresssion specifying an extra random effect to account for the multiple outcomes. I would do it using the metafor package in R (the function rma.mv is relevant here but you may have other software preferences.