|
Abstract:
|
We study a class of propensity score weighting approaches to the analysis of observational studies with dichotomous treatment and survival outcome subject to independent right censoring. Unlike the case with continuous or categorical outcomes, there are special challenges in the definition and estimation of the causal estimand with survival data. We will discuss these issues, provide several estimands within the potential outcomes framework, and propose methods for consistent estimation and variance estimation. Our analytical framework is based on a general class of weights on the individual causal effect. A simulation study is conducted to evaluate the finite sample performance of the proposed methodology. We also apply our methodology to a data set from a cancer radiation therapy study.
|
