CAJ | 학술논문

以福建省汉族人群原发性肺癌环境危险因素研究为实例,通过构建广义线性模型,利用λ值、偏差度及Pearson x2拟合优度检验等判断方法选择合适模型,进行多因素及因素间交互作用分析,发现自变量之间存在相乘交互作用,进而选用logistic回归模型分析,最终得到吸烟者有7个、非吸烟者有9个单独作用的肺癌危险因素.其中吸烟者中重度吸烟与肺部罹患过炎症性疾病呈正相乘作用,非吸烟者中被动吸烟与少吃新鲜水果呈正相乘作用.运用广义线性模型筛选合适模型并进行交互作用分析,利于全面合理分析流行病学数据.
이복건성한족인군원발성폐암배경위험인소연구위실례,통과구건엄의선성모형,이용λ치、편차도급Pearson x2의합우도검험등판단방법선택합괄모형,진행다인소급인소간교호작용분석,발현자변량지간존재상승교호작용,진이선용logistic회귀모형분석,최종득도흡연자유7개、비흡연자유9개단독작용적폐암위험인소.기중흡연자중중도흡연여폐부리환과염증성질병정정상승작용,비흡연자중피동흡연여소흘신선수과정정상승작용.운용엄의선성모형사선합괄모형병진행교호작용분석,리우전면합리분석류행병학수거.
The purpose of this study was to use the data on lung cancer in Han Chinese in Fujian province to explore the value of a generalized,linear model and to investigate the impact related to environment factors on lung cancer as well as the independent and interaction effects on the development of lung cancer.SAS 9.2 was used to build a generalized linear model to evaluate the influence factors and interaction of lung cancer on both smokers and non-smokers.Results showed that the relationship of the factors was multiplied.Under the logistic regression analysis,seven risk factors and nine risk factors were noticed in smokers or in non-smokers,respectively.Heavy smokers and lung diseases appeared a positive multiplying effect on smokers while passive smoking and fresh fruits showed positive multiplying effects on non-smokers.The generalized linear models could filter suitable models thus facilitating further research on the interaction between the two.It seemed easy to carry on the comprehensive and rational analysis on related epidemiological data.