生存分析模型在肺结核患者治愈率分析中的应用  

The application of the survival analysis model in the study of pulmonary tuberculosis cure rate

作　　者：顾海雁[1] 葛祖恂[2] 陈峰[3] 朱和新[2] 荀鹏程[2] 黄连章 唐宇明[2] 吴美兰[2] 

机构地区：[1]上海市徐汇区疾病预防控制中心,上海200031 [2]南通医学院,江苏南通226001 [3]南京医科大学,江苏南京210029 [4]盐城市疾病预防控制中心,江苏盐城224002

出　　处：《江苏预防医学》2005年第2期1-4,共4页Jiangsu Journal of Preventive Medicine

摘　　要：目的:将生存分析方法应用于肺结核病治愈率的动态描述及其影响因素分析。方法:以江苏省盐城市6个县登记治疗并实施全程督导管理的172例肺结核初治菌阳病人的3年随访资料为例,采用生存分析估计治愈率,用非参数的Cox比例风险回归模型对肺结核治愈率进行单因素和多因素分析,并对有统计学意义的影响因素尝试用Weibull分布的参数方法拟合治愈率曲线。结果:生存曲线给出不同时间点的治愈率,随着治疗时间的持续,肺结核病人的治愈率由低到高,最终接近100%;Cox回归模型显示肺结核病治愈率受多种因素影响,如吸烟、不同治疗方案及不同区县等。结论:生存分析方法适用于肺结核病等慢性传染病疗效指标的动态描述及其影响因素的分析。Objective: Survival analysis was applied to describe the cure rate of pulmonary tuberculosis dynamically and screening the related influential factors. Methods: Based on a 3-year follow-up study data of 172 first-treated pulmonary tuberculosis patients of positive bacillus with DOTS in 6 counties of Yancheng City, cure rate was analyzed with single and multiple Cox proportional hazards regression model, respectively, and was attempted to fit a curve of Weibull distribution according to the influential factors which was significant. Results: The curve of cure rate was fitted with Weibull functions, which showed that the cure rate increased gradually till 100% as the cure time went on. Moreover the cure rate of tuberculosis was affected by various factors of patients, the Cox regression model showed that smoking, the remedy of tuberculosis and the living county of patients influenced the cure effect of tuberculosis. Conclusion: The method of survival analysis was shown to be suitable for the dynamic description of the cure effect of chronic infectious disease such as tuberculosis and for the influential factors analysis.

关 键 词：肺结核 治愈率 生存分析 COX回归模型 

分 类 号：R195.1[医药卫生—卫生统计学] R521[医药卫生—卫生事业管理]