Abstract:
|
The joint modelling of longitudinal and survival data has been a rapidly expanding area of methodological research over the past 20 years, and is now making its way into applied research, particularly in the fields of cancer and HIV, with increasing availability of user-friendly software. In this talk, I will describe the stjm package in Stata to fit a variety of joint longitudinal-survival models within the shared parameter framework, with an emphasis on flexible parametric approaches. I will describe the implementation of fully adaptive Gauss-Hermite quadrature to calculate the joint likelihood, illustrating the superiority over non-adaptive quadrature through simulation. With particular emphasis on the use of restricted cubic splines to model both the baseline hazard function and the longitudinal trajectory over time, I will illustrate the package using real and simulated datasets. Finally, I will describe some recent extensions to stjm, to model multiple longitudinal outcomes, each of which can be continuous, categorical or count, jointly with survival, and the incorporation of delayed entry (left truncation).
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.