Abstract:
|
In observational studies of survival time featuring a binary time-dependent treatment, the hazard ratio is often used to represent the treatment effect. However, investigators are often more interested in the difference between survival functions. We propose flexible methods applicable to big data sets for the purpose of to estimating the causal effect of treatment among the treated with respect to survival probability. The objective is to compare post-treatment survival with the survival function that would have been observed in the absence of treatment. The proposed methods utilize prognostic scores, but are otherwise nonparametric. Essentially, each treated patient is matched to a group of similar qualified not-yet- treated patients. The treatment effect is then estimated through a difference in weighted Nelson-Aalen survival curves, which can be subsequently integrated to obtain the corresponding difference in restricted mean survival time. Large-sample properties are derived, with finite-sample properties evaluated through simulation. The proposed methods are then, applied to estimate the effect on the survival of kidney transplantation among end-stage renal disease patients
|