Confounders Are Always a Problem

Abstract

Readers who were intrigued by the surprisingly general beneficial effects of measles vaccine reported by Holt et al¹ may be interested to know that a precedent exists for such a finding. A series of articles by Black, Fox, Elveback, and co-workers^2-6 deserve greater attention. In three independent and classic studies, they documented that preschool children with a history of measles infection had lower cognitive test scores than measles-susceptible children. This effect occurred despite use of approximately six covariates for adjustment.