Stephen J.LurieMD, PhD, Senior EditorIndividualAuthorJody W.ZylkeMD, Contributing EditorIndividualAuthor
In Reply: We performed univariate and multivariable
analyses of the case-control data, which included all 50 case subjects and
148 control subjects; however, we also performed a subanalysis in which the
4 dead case subjects and their associated 12 control subjects were excluded.
Results from the initial univariate and multivariable analyses remained unchanged
in the subanalysis. We agree with Dr Mulla that to reduce differential (nonrandom)
misclassification bias, researchers conducting case-control studies should
consider the use of control proxies, particularly when a substantial number
of case subjects are either mentally incapacitated or have died; however,
in this case using matched student control subjects, not control proxies,
did not result in any substantial differential misclassification bias.
Bruce MG, Rosenstein NE, Shutt KA. Meningococcal Disease in College Students—Reply. JAMA. 2001;286(21):2668. doi:10.1001/jama.286.21.2665