Propensity Score Matching


I'm conducting a retrospective study in which we want to analyze time period of 60 days before occurence of a particular adverse event. To do so, we have 15 patients with AE and we've created a control group of 15 pts without AE using PSM. The question is: can we match the time periods in the created control group with time periods in the cases using post-operative days (POD)?


Omega Contributor
So you already have PS, which didn't use POD in the model, now you have 15 controls matched to 15 cases. If they are already matched how would you match them again. How big is the control pool? Can you create PS for all eligible controls then match the best score with the same POD to the Cases?

Please provide more information! Thanks.