骨髓移植数据的删失分位数回归分析及SAS软件实现

Censored Quantile Regression Analysis of Bone Marrow Transplantation Data and Its Implementations of SAS

针对髓细胞性白血病患者异体骨髓移植数据,基于软件SAS 9.4,对各生存时间及其有影响的协变量进行比例风险回归模型(Cox)、加速失效时间模型(AFT)对此数据进行回归分析,并应用拟合优度检验和逐步回归进行自变量选元,在此选元基础上建立了删失分位数回归分析,对此数据进行了更详细的分析,讨论了不同分位数时各协变量的回归系数变化情况.

For myelocytic leukemia patients with allogenic bone marrow transplantation data,we have applied the proportional hazard regression model（ Cox） and accelerate failure time model（ AFT） into this data including the interested failure time and its covariates based on SAS 9. 4 statistical software. The independent variables were chosen by the goodness of fit test and stepwise regression. On the basis of the variable selection,we establish the censored quantile regression model. This data is analyzed in more detail. We discuss the variation of regression coefficients of different covariates in different quantiles.