2 weeks ago - last edited 2 weeks ago
I am using proc genmod with gamma distribution to compare costs across two different cohorts. The sample has variable follow up and also I am using the Inverse probability treatment weighting (IPTW) method. In order to account for variable follow up and apply the weight for IPTW, I am using both offset for log of follow up time and weighting by the inverse probability treatment weight in the same model. Is it ok to use both? When I run the model, I am not getting errors or warnings and the algorithm is converging. So is it fine to use both in the same model?
I will appreciate suggestions if any.