We appreciate the letter by Miskulin et al1 regarding our recent study2 and agree with the importance of assessing and accounting for confounding bias in non-experimental studies of preventive interventions. Miskulin et al raise an important concern about potential residual confounding due to the relatively small proportion and potentially highly selected group of high-dose influenza vaccine (HDV) recipients in the national data set that we analyzed in the years following the 2009 introduction of HDV.