Online Program

Multivariate Frailty Model for Recurrent-Event Data with Multiple Types

*Khaled F Bedair, PhD student, Department of Statistics, Virginia Tech 
Yilli Hong, Assistant Professor, Department of Statistics, Virginia Tech 

Keywords: MCEM algorithm, Multiple type recurrent events, Multivariate frailty Models, Random effects, Survival analysis

There has been an increasing interest in analyzing the multitype recurrent event data. This type of data arises in many situations when two or more different event types may occur repeatedly over an observation period. For example, subjects are at risk of different disease types, causes of production stoppages, financial transactions in commerce, and insurance claims filed by holders. The interest in this setting is to characterize the incidence rate of event types, estimate the impact of covariates, and understand the correlation structure among event types. We propose a multivariate frailty proportional intensity model with multivariate distributions for the random effects to model the data. The dependence among event types is taken into account as well as the effect of covariates. Maximum likelihood estimates of the regression coefficients, variance-covariance components, and the baseline intensity function are obtained based on a Monte Carlo Expectation Maximization (MCEM) algorithm. The E-step of the algorithm involves the calculation of the conditional expectations of the random effects by using the Metropolis-Hastings sampling. Louis' formula is applied to obtain the variance of the estimator. Simulation studies are presented to illustrate the performance of the proposed method. An application will be described to a randomized controlled clinical trail for the efficacy of nutritional supplements for the prevention of tow types of skin cancer.