LongitudinalEvalue - Sensitivity Analysis for Unmeasured Confounding in Longitudinal Studies

Evaluates the sensitivity of a given association to unmeasured confounding. The package consists of three functions. The first starts a 'shiny' app which assesses how strong a time-invariant confounder needs to be associated with the exposure and outcome to explain away a proposed causal association. The second and third functions simulate a time-dependent confounder over time either using a fit from the qmle() function from the 'yuima' package or directly using the observed effect estimate.