生存率置信区间的五种估计方法

项永兵; 高玉堂; 金凡; 杨工; 孙璐

Abstract

项永兵,高玉堂,金凡,杨工,孙璐.生存率置信区间的五种估计方法[J].Chinese journal of Epidemiology,1995,16(5):306-309

生存率置信区间的五种估计方法

Five Methods for the Estimation of Confidence Intervals on Survival Rate

Received:September 22, 1994 Revised:November 10, 1994

DOI：

KeyWord: 生存率置信区间寿命表

English Key Word: Survival rate Confidence intervals Life table estimator

FundProject:

Author Name	Affiliation
Xiang Yong-bing	Department of Epidemiology, Shanghai Cancer Institute, Shanghai 200032
Gao Yu-tang	Department of Epidemiology, Shanghai Cancer Institute, Shanghai 200032
Jin Fan	Department of Epidemiology, Shanghai Cancer Institute, Shanghai 200032
杨工	浙江医科大学肿瘤研究所
孙璐	Department of Epidemiology, Shanghai Cancer Institute, Shanghai 200032

Hits: 3148

Download times: 675

Abstract:

生存率是医学随访研究资料分析中常用的指标,例如适用于小样本资料的Kaplan Meier乘积限估计和大样本资料的寿命表法生存率估计。本文对生存率置信区间的估计方法进行了讨论。主要介绍了五种置信区间的估计方法:经典法(基于Greenwood方差公式)、校正法、反正旋转换法、log(-log)转换法及logit转换法。文中给出了两个实例,并就生存率95%置信区间的估计做了详细介绍,还进一步讨论了它们在小样本和大样本随访资料分析中的特点和应用。

English Abstract:

Descriptive analysis on medical follow-up data often involves estimation of survival rates, such as Kaplan-Meier product limit estimator or life table survival rate for small or large scale data sets, respectively. Five methods for estimating the confidence intervals (CI) of survival rates were described. These methods for estimation of intervals were the classic (based on Greenwood's formula), correctness, arcsine transformation, log (-log) transformation and logit transformation. Two examples were described in detailed for the calculation of 95% confidence intervals.

View Fulltext Html FullText View/Add Comment Download reader