We disagree with Thompson and colleagues’ assertion that our trends study1 is less robust compared with cohort studies. The strength of our approach is that, unlike cohort studies, we analyzed all deaths in the total elderly US population and are free of the biases to which cohort studies are subject. The comments by Thompson et al are surprising given that they used a similar study design and also found an increase in influenza-related mortality in recent years, which they suggested could be partially explained by aging of the population.2 Nevertheless, the critique by Thompson et al centers on our modeling assumptions, most of which were addressed in our article, including the duration of epidemic periods. We have provided age-adjusted excess mortality estimates, which addressed the paradoxical observation that mortality rates increased concomitant with a quadrupling in elderly vaccination coverage.1 If Thompson et al would also adjust their excess mortality estimates, they could test the robustness of our finding that increasing influenza vaccination coverage was associated with less mortality reduction than expected.
Simonsen L, Reichert TA, Viboud C, Blackwelder WC, Taylor RJ, Miller MA. Influenza Vaccination Among the Elderly in the United States—Reply. Arch Intern Med. 2005;165(17):2039-2040. doi:10.1001/archinte.165.17.2039