Prof.  Sanjoy  Sinha
Carleton University,  Canada

Title: Joint Modeling of Longitudinal and Time-to-event Data

Abstract:

In clinical studies, often subjects are measured repeatedly over a given period of time. Longitudinal measurements from a subject are naturally correlated. Mixed models are widely used to describe the dependence among longitudinal outcomes. In addition to the longitudinal data, we often collect time-to-event data (e.g., recurrence time of a tumor) from the subjects. When multiple outcomes are observed from a subject with a clear dependence among the outcomes, a natural way of analyzing the outcomes and their associations would be the use of a joint model. I will discuss a likelihood method for jointly analyzing the longitudinal and time-to-event data. The method is developed to deal with left-censored covariates often observed in clinical studies due to the limit of detection. Empirical properties of the proposed estimators will be discussed using results from a Monte Carlo study. An application will be provided using a large clinical dataset of pneumonia patients obtained from the Genetic and Inflammatory Markers of Sepsis (GenIMS) study.

Biography:

Sanjoy Sinha received his Ph.D. degree in Statistics from Dalhousie University, Halifax, Canada. He is a Full Professor of Statistics in the School of Mathematics and Statistics at Carleton University, Ottawa, Canada. His research interests are in survival analysis, longitudinal data analysis, missing data analysis, mixed models, and robust inference. He is actively involved in graduate students supervision at Carleton University, and supervised 5 PhD students and 12 M.Sc. students over the past twelve years. Currently, he is supervising 4 PhD students. He is also actively involved in collaborative research with many clinical practitioners and epidemiologists at Health Canada and professors at different universities in Canada and USA. His research papers were published in various statistics journals, which include Journal of the American Statistical Association, Journal of Multivariate Analysis, and The Canadian Journal of Statistics.