生存树方法及其在预后分析中的应用  

The Survival Tree Method and its Application in Prognostic Analysis

作　　者：郎素平[1] 余红梅[2] 王彤[2] 何大卫[2] 

机构地区：[1]南京医科大学公共卫生学院流行病与卫生统计学系,210029 [2]山西医科大学公共卫生学院卫生统计教研室,030001

出　　处：《中国卫生统计》2006年第1期13-15,共3页Chinese Journal of Health Statistics

摘　　要：目的针对临床随访资料,利用生存树方法进行预后分组,进一步完善预后分析的内容。方法从生存树建立的主要过程———树的生长、剪枝以及子树的选择来介绍生存树方法,并对一乳腺增生症复发资料进行分析。结果生存树方法显示:疗效、疗程以及增生程度作为对预后影响最大的因素用来预后分组,得到5个预后子群,其中位生存期分别为30·3月、21·4月、17·3月、11·0月、7·3月。结论应用生存树方法进行预后分析,不但可以得到具有不同临床特征的预后子群(预后分组),同时可以获得预后相关因素。Objective The survival tree method can identify subsets of the patient population with homogeneous prognosis. It is also useful for perfecting the prognostic analysis. Methods The building of survival tree includes tree growing and tree pruning. And the survival tree method was used to do prognostic analysis of 616 Hyperplasia disease of mammary gland patients with complete clinical and pathologic material. Restilts Treatment effect, during of treatment and the degree of hyperplasia were found to be most powerful factors in prognostic subgroups. With the survival tree method, five subgroups of patients were made with median survival times of 30,3, 21.4, 17.3, ll.0 and 7.3 months respectively. Conclusion Compared with Cox regression, the survival tree method may help us not only tofind the prognostic subgroups of homogeneous prognosis, but also to identify the independent prognostic factors.

关 键 词：生存分析 生存树 预后 

分 类 号：R449[医药卫生—诊断学]