IR and PR (Prevalence ratio) same for Modified Poisson Regression with Robust Error Variance

[Nusrat Jahan Sathi]

Hello


glm y i.x1 i.x2 i.x3 i., link(log) family(poisson) eform

Using the above command, I have found IRR. My data is from a cross-sectional survey. But I need to estimate PR (Prevalence Ratio) from Poisson regression. Are IRR and PR the same?

[Clyde Schechter]

Yes and no. IRR and PR are similar but distinct concepts. They are calculated the same way, but from different kinds of data.

So it all depends on what your variable y represents. If y represents the number of existing cases of something at a given point of time, your variable y is a prevalence count, and the exponentiated coefficients from a Poisson model are prevalence ratios. If y represents the number of new cases arising at a given point of time, your variable y is an incidence count and the exponentiated coefficients from a Poisson model are incidence rate ratios. StataCorp chose to call its exponentiated Poisson coefficients IRRs, probably because in practice that is the more common application. But when you apply a Poisson model to incidence count data, you get incidence rate ratios.

If your data are incidence counts, then you cannot calculate prevalence rates from it (at least not from it a lone). And if you have prevalence counts, you cannot calculate incidence rate ratios from that data alone.

So, if you have data where y is a prevalence count, your regression gives you what you are asking for. If you don't, you cannot get it anyway without additional data.

[Nusrat Jahan Sathi]

Thank you very much.