结局变量为二值时利用CSTE曲线选择最优治疗方案

CSTE curve for selection of the optimal treatment when outcome is binary

本文对二值结局变量数据,基于因果推断理论,提出根据患者的生物标记物进行最优治疗方案选择的统计方法.这种方法基于CSTE(covariate-specific treatment effect)曲线和CSTE曲线的置信带(SCB).CSTE曲线表示在给定生物标记物(协变量)的条件下,处理组的条件平均处理效应.同时,CSTE曲线及其SCB可以被用于对特定的治疗方案选择适宜的患者.本文利用B-样条方法估计CSTE曲线及其CSB,并推导了其近似大样本性质.本文还通过模拟比较研究了CSTE曲线的置信带的有限样本性质,并阐述了CSTE曲线及其置信带在真实数据中如何选择最优治疗方案.

In this paper,we propose a statistical framework for optimal treatment selection for a subgroup of patients,using their biomarker values based on casual inference for binary outcome data.This new method was based on a concept,called covariate-specific treatment effect（CSTE） curve and CSTE curve＇s simultaneous confidence bands（SCBs）,which could be used to represent the average treatment effect of the treatment for a given value of the covariate（biomarker） and to select an optimal treatment for one particular patient.We then propose B-splines methods for estimating the CSTE curves and constructing simultaneous confidence bands for the CSTE curves.We derive the asymptotic properties of the proposed methods.We also conduct extensive simulation studies to evaluate finite-sample properties of the proposed simultaneous confidence bands.Finally,we illustrate the application of the CSTE curve and its simultaneous confidence bands in optimal treatment selection in a real-world data set.