epiR软件包在肿瘤流行病学分层分析中应用

Application of epiR package in stratified analysis in the field of cancer epidemiology

目的分层分析是将数据按照某个(些)需要控制的变量进行分层,然后再估计暴露与结局之间关联强度的一种资料分析方法,是肿瘤流行病学研究中最常用的控制混杂的方法之一。本研究旨在应用R软件的epiR软件包实现肿瘤流行病学资料的分层分析,为识别和控制混杂因素提供新的统计学工具。方法结合2009年的一项关于乳腺癌的病例对照研究数据,以体质指数(body mass index,BMI)为分层变量,分析血清中抵抗素含量和乳腺癌发病的关系,计算采用epiR软件包实现。结果未经BMI调整时,血清中抵抗素含量与乳腺癌发生的关联强度ORc=3.431(95%CI:1.590~7.406),P=0.001;经BMI分层调整后,血清中抵抗素含量与乳腺癌发生的关联强度ORmh=3.809(95%CI:1.703~8.518),P=0.001。结论调整BMI混杂因素的影响之后,血清中抵抗素含量与乳腺癌发生的关联性依然存在,表明血清中抵抗素含量可能是乳腺癌的危险因素之一。epiR软件包程序书写简单,结果输出丰富,能方便地完成分层分析,可以为肿瘤流行病学研究人员开展分层分析提供参考。

OBJECTIVE Stratified analysis is a method of data analysis that stratifies population into distinct categories based on levels of a parameter,and then estimates the level-specific associations between exposure and outcome.It is one of the most popular methods to control confounding in the field of cancer epidemiology.This study aimed to introduce a R package named as epiR,which could conduct stratified analysis more easily for freshman.METHODS Data from a case-control study of breast cancer was download as examples,which aimed to explore the association between serum resistin content and the risk of breast cancer.Body mass index(BMI)was considered as the stratification variable,and the stratified analysis was finished with epiR.RESULTS The crude odds ratio between serum resistin content and breast cancer was 3.431(95%CI:1.590-7.406,P=0.001).This value turned to be 3.809(95%CI:1.703-8.518,P=0.001)after stratified with BMI.CONCLUSIONS After adjusting for the influence of BMI confounding factors,the correlation between serum resistin content and breast cancer persisted,indicating that serum resistin content may be one of the risk factors for breast cancer.The epiR package is easy to handle with various useful output results simultaneously.It could provide a new choice for the stratified analysis for numerous cancer epidemiology researchers.