CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Marginal accelerated failure time mixture cure model for clustered survival data Authors:  Duze Fan - Dalian University of Technology (China)
Yingwei Peng - Queen\'s University (Canada)
Yi Niu - Dalian University of Technology (China) [presenting]
Abstract: The semiparametric accelerated failure time (AFT) mixture cure model is an appealing alternative to the semiparametric proportional hazards mixture cure model in analyzing multivariate failure time data with long-term survivors. However, the former received less attention than the latter due to the complexity of the estimation method for the former, and the model was not proposed for clustered survival data. We consider a marginal semiparametric AFT mixture cure model for clustered failure time data with a potential cure fraction. We propose a generalized estimating equations (GEE) approach based on the Expectation-Solution (ES) algorithm to estimate the regression parameters in the model. The correlation structures within clusters are modeled by working correlation matrices in the GEE. We use a bootstrap method to obtain the variances of the estimators. Numerical studies demonstrate that the efficiency gain of the regression coefficient estimators is robust to the misspecification of working matrices, and the efficiency is higher when the working correlation structure is closer to the truth. Finally, we apply the model and the proposed method to analyze the data from a smoking cessation study and a tonsil cancer study for illustration.