meta analysis - where to start.

#1
Hello.

I have also posted this in random effects model - meta - analysis - where to start - Cross Validated (stackexchange.com)

I've been given a task to investigate how to carry out the meta analysis on the results from a series of studies. They all report the 'standard mortality rate' with confidence intervals. An example is below ( I am assured that the studies have been selected in line with good practice)

Meta_analysis.PNG


I have searched, read and watched numerous videos and could do with being put on the right track.

As a first step, I think that i need to assume that I should use the random-effects model and that I need to calculate an 'effect size' for each study. What method would be best to employ to obtain the effect size for each study?

I would like to understand how the calculations work before I move onto using a statistical package.

I may be asking the wrong question here so I apologise in advance.

Any help would be appreciated.

Thankyou,

Andy
 
Last edited:

hlsmith

Less is more. Stay pure. Stay poor.
#2
Yes, you can conduct an MA. There are some statistics to examine if you may need to address random effects. Typically a person's content know may guide this.

MA's can be conducted with almost any outcome format. I haven't used SMRs, but I am sure that can be used.