meta regression using STATA 16

[Tom Salvitti]

Hi to everybody..
I try again to explain my problem, hoping that somebody answer me.
Not being very experienced, I would like to ask you a few things to make sure I am doing well.

I have

In Stata16 we can have:

1) meta-analysis with effect size data already present in the data set.
2) meta-analysis with summary data

I've applied the case 1 (meta-analysis with effect size data already present in the data set)


/*objective of the study: prevalence of ADHD (Attention Deficit/Hyperactivity Disorder) in autistic people (ASD)*/

I have these variables in my dataset:
ADHD_ASD = autistic people with ADHD
ASD= autistic people
ID=Intellectual disability
age=age of autistic people
female= %female with ASD



Since my measures of effect-size is the prevalence, I proceeded as follows:



metaprop ADHD_ASD ASD, random cimethod (exact)

rename _ES prev_ASD_ADHD

rename _se_ES se_prev_ASD_ADHD

meta set prev_ASD_ADHD se_prev_ASD_ADHD

/*text covariate age*/

meta regress age
estat bubbleplot

/*text covariate intellectual disability*/

meta regress ID
estat bubbleplot

/*text covariate %female*/

meta regress female
estat bubbleplot

/*text all covariates , multivariable meta-regression*/

meta regress age ID female
meta bias, egger
meta funnelplot, random


What do you think?


Then, I would ask you onother thing, mostly my dubt. If you use onother software, for example SAS, the value of I squared, R squared, egger test, may be a little different, if for example, you use STATA?


Thanks a million

Best

Tommaso Salvitti

[Tom Salvitti]

I'm pretty sure I've done well, even though I don't think there's anyone who can confirm me.