Causal mediation analysis with a failure time outcome in the presence of exposure measurement error

Causal mediation analysis is widely used in health science research to evaluate the extent to which an intermediate variable explains an observed exposure-outcome relationship. However, the validity of analysis can be compromised when the exposure is measured with error, which is common in health science studies. This article investigates the impact of exposure measurement error on assessing mediation with a failure time outcome, where a Cox proportional hazard model is considered for the outcome. When the outcome is rare with no exposure-mediator interaction, we show that the unadjusted estimators of the natural indirect and direct effects can be biased into either direction, but the unadjusted estimator of the mediation proportion is approximately unbiased as long as measurement error is not large or the mediator-exposure association is not strong. We propose ordinary regression calibration and risk set regression calibration approaches to correct the exposure measurement error-induced bias in estimating mediation effects and to allow for an exposure-mediator interaction in the Cox outcome model. The proposed approaches require a validation study to characterize the measurement error process between the true exposure and its error-prone counterpart. We apply the proposed approaches to the Health Professionals Follow-up study to evaluate extent to which body mass index mediates the effect of vigorous physical activity on the risk of cardiovascular diseases, and assess the finite-sample properties of the proposed estimators via simulations.