Hello, I am looking at data on child survival from a survey with a stratified two-stage design. I have been trying out the excellent survey package, with which I have done a descriptive analysis and Cox models. I need to calculate rates (with a child-year denominator) and various rate ratios plus confidence intervals (basically, something like the STATA commands strate and stmh used after svyset and stset) but haven't been able to find a convenient way of doing this in R despite much searching. Is this possible, either with the survey package or any other way? Many thanks for any advice. Paul -- Dr. Paul Cleary, Specialist Registrar in Public Health, Chester Health Protection Unit. [[alternative HTML version deleted]]