The Fred Hutch Biostatistics Program hosts seminars featuring presentations by Hutch and outside scientists to share their latest developments and recent research. Each seminar includes an hour-long presentation and discussion during which speakers showcase their work and findings.
This seminar will be held on Zoom due to the COVID-19 pandemic.
Biostatistics Seminar Series:
“Defining and estimating subgroup mediation effects with Semi-Competing Risks Data”
In many medical studies, an ultimate failure event such as death is likely to be mediated by the occurrence and timing of other intermediate clinical events. Both the primary outcome and the mediator are event times which are subject to censoring by loss to follow up, but the mediator may further be subject to censoring by the occurrence of the primary outcome, but not vice versa, such that we observe semi-competing risks data.
Due to this special data structure, conventional definition of direct and indirect effects are not applicable.
We propose a principle stratification framework which allow proper definition of direct and indirect effects in one stratum, but total effects are defined for all strata. In particular, we define three principal strata based on the susceptibility of the intermediate event given different treatments and model the marginal structure of the counterfactuals by proportional hazards models. By treating the unobserved stratum membership as a latent variable, we propose an EM algorithm for computation. We study the asymptotic properties of the estimators by the modern empirical process theory and examine the performance of the estimators in numerical studies.