Recovering True Risks When Multilevel Exposure and Covariables Are Both Misclassified

In July 1996, Sakai City, Japan, experienced the largest outbreak of Escherichia coli 0157:H7 infections ever reported, involving over 7,000 persons. Michino et al. (1) have convincingly demonstrated through a review of school absentee records, a cohort study of over 47,000 schoolchildren, product traceback, and molecular subtyping that illness was due to consumption of contaminated white radish sprouts served through a centralized lunch program. Multiple other outbreaks of E. coli 0157:H7 infections occurred in Japan during the same summer (2). Investigations of these outbreaks as well as the one in Sakai City highlight some of the problems that face public health officials worldwide and illustrate lessons to be learned for investigating foodborne disease outbreaks.