Risk difference (nlmeans) for the conditional logistic model and weight model

[leesw0212]

I found that the macro, nlmeans, is commonly used to calculate the risk difference along with its 95% confidence intervall.

Is it appropriate to use this macro with the conditional logistic model (PROC LOGISTIC with the STRATA option) and the weighted model (PROC LOGISTIC with the WEIGHT option)?

The conditional model is used because the data is matched, while the weighted model is applied for the Inverse Probability of Treatment Weighting (IPTW) model.

[Ksharp]

Yes. I think so . Check GEE model here:
https://support.sas.com/kb/46/997.html

[leesw0212]

Thank you so much!