Correcting for the bias caused by exposure measurement error in epidemiological studies.

An important goal of many epidemiological studies is to estimate the magnitude of association between an exposure and an outcome. Exposure measurement error causes bias in such estimates of association and can be substantial. In this article, we describe the problem of exposure measurement error and its effects.