Statistics Colloquium
September 30 (Fri) 11 a.m.
Alavi Commons Room, 6625 Everett Tower

Matching Methods for Obtaining Survival Functions to Estimate the Effect of a Time-Dependent Treatment

Yun Li
Biostatistics, University of Michigan

In observational studies of survival time featuring a binary time-dependent treatment,the hazard ratio (an instantaneous measure) is often used to represent the treatment effect. However, investigators are often more interested in the difference in survival functions, which is not straightforward to obtain. We propose semi-parametric methods to estimate the causal effect on the post-treatment survival among the treated. The objective is to compare post-treatment survival among patients who receive treatment, with the survival function that would have been observed in the absence of treatment. We use various matching techniques to select comparable patients for each treated patient. After the adjustment covariate distributions are balanced through matching, non-parametric methods can be employed to estimate the survival functions. We conduct matching by comparing a treated patient with his/her potential matches by the closeness of their risk scores that either base on the death hazard or time-dependent treatment propensity or both. The censoring distribution is accounted for by inverse probability of censoring weighting. The proposed methods are applied to examine the effect of transplantation on the post-transplant survival curves among end-stage renal disease patients.

All statistics students are expected to attend.


Past colloquiums


Department of Statistics
3304 Everett Tower
Western Michigan University
Kalamazoo MI 49008-5152 USA
(269) 387-1420 | (269) 387-1419 Fax