Mendelian randomization with repeated measures of a time-varying exposure: an application of structural mean models

January 2022
Epidemiology

Joy Shi, Sonja A. Swanson, Peter Kraft, Bernard Rosner, Immaculata De Vivo, and Miguel A. Hernán

We describe instrumental variable analysis with a time-varying exposure, with an emphasis on the types of causal estimands that can be targeted, the assumptions required for identification, and the use of g-estimation of structural mean models to estimate these effects.

Instrumental variable estimation for a time-varying treatment and a time-to-event outcome via structural nested cumulative failure time models

November 2021
BMC Medical Research Methodology

Joy Shi, Sonja A. Swanson, Peter Kraft, Bernard Rosner, Immaculata De Vivo, and Miguel A. Hernán

We propose the use of g-estimation of structural nested cumulative failure time models for instrumental variable analyses of time-varying exposures and failure time outcomes.