03-30-2016 02:04 PM
I am looking to do a stratified analysis of multiple risk factors against binary outcome using complex survey data. Is there any to generate an output similiar to the CHM output for proc freq? I am looking to identify potential effect modifiers and confounders and would like to generate a Mantel-Hanenszel test for the association adjusted for stratification variables/a Breslow-Day test for homogeneity of the odds ratios.
If not is there an easy straightforward approach to doing a similar type of analysis in proc surveylogistic?