1 Division of Disease Control, Department of International Health, The Johns Hopkins University, School of Hygiene & Public Health, 615 North Wolfe Street, Baltimore, Maryland 21205
Dr Brunskill's primary point appears to be in the final sentence of his letter. As noted in our discussion, we are well aware of the potential for hidden confounding factors in retrospective observational studies... We attempted to minimize this possibility by obtaining information on an extensive list of factors that might have confounded the relationship between vaccination and mortality and by controlling for these factors through multivariate analysis.1 Despite our rigor, we could have failed to measure all possible confounding variables.