Is It Necessary to Correct for Measurement Error in Nutritional Epidemiology?

Relationships between diet and chronic disease have become the focus of many analytic studies in nutritional epidemiology over the past several decades. Given often relatively limited variation in dietary intake within study populations, the results of observational studies critically depend on accurate assessment of the dietary exposure and other adjustment covariates, especially in view of the moderate-sized relative risks seen in many nutrition studies. Error in measuring exposure leads to a biased and inefficient estimate of the relationship of the exposure to disease (1). In this issue, Beulens and colleagues (2) present a prospective analysis of the relationship between alcohol intake, which participants self-reported using a food-frequency questionnaire, and cardiovascular events among men with hypertension in the Health Professionals Follow-Up Study. The authors report the findings before and after correction for measurement error, which provides a welcome opportunity to comment on the importance of measurement error in nutritional epidemiology.

Measurement error in covariate assessment of prospective studies is generally assumed to be nondifferential with respect to disease, which means that the relationship between true and reported diet is the same for persons who develop the disease as for those who do not. If this assumption is true, most authors in epidemiologic literature assume that measurement error attenuates the true relative risk—that is, it biases its estimate toward no effect. Consequently, authors may interpret null results as perhaps being due to exposure measurement error. In a similar manner, because of measurement error, the true relative risk in a statistically significant association should, if anything, be larger than that reported. This attenuation of effect always occurs when only the main exposure is measured with error, and this error follows the “classical model”: It is additive; it is independent of the size of the true exposure; and it has a mean of …