How do I obtain prevalence ratios using the logistic command?

[jennifer Thornton]

I am using the "logistic" command to estimate a binary outcome (yes/no response, cross-sectional survey, approximately 9500 respondents) based on six ordinal predictor variables. I would like to estimate prevalence ratios (both adjusted and unadjusted) instead of odds ratios. (Many papers suggest using prevalence ratios in place of odds ratios.)

However, the output I receive from STATA provides odds ratios, not prevalence ratios, and I have been unable to find a solution to this.

Can anyone please help me with the syntax for this?

Thank you.

[Clyde Schechter]

You can't get prevalence ratios directly out of logistic. But if you have properly used factor variable notation (-help fvvarlist-) in your --logistic- command, you can run -margins, post- afterwards. That will give you adjusted prevalences. Then you can use -nlcom- to get prevalence ratios from there.

[jennifer Thornton]

Thank you very much! I will try that. Much appreciated.

[Bjarte Aagnes]

The following have some information related to your question https://thestatsgeek.com/2015/03/09/...data-in-stata/

But, what is the justification for using logistic regression when the relative risk is the parameter of interest? Why not using a statistical model that directely estimates the risk ratio; either the log-binomial modell or the Poisson regression with robust variance estimator.

glm y x , family(binomial) link(log) vce(robust) eform search difficult  
glm y x , family(poisson)  link(log) vce(robust) eform

the above models can also be a estimated using

binreg  y x , rr
poisson y x , irr vce(robust)

[jennifer Thornton]

Thank you belatedly for this helpful response. I appreciate your time.