| 吕筠,何平平,涂文校,李立明.整群抽样调查数据分析中应正确计算抽样误差[J].Chinese journal of Epidemiology,2008,29(1):78-80 |
| 整群抽样调查数据分析中应正确计算抽样误差 |
| Estimation of sampling error on data from cluster sample survey |
| Received:August 17, 2007 |
| DOI： |
| KeyWord: 整群抽样 单纯随机抽样 抽样误差 复杂调查数据 |
| English Key Word: Colorectal cancer Prognosis Survival analysis Proportional hazard regression model Prtional hazard assumption |
| FundProject:国家自然科学基金资助项目(30371239、30571598) |
|
| Hits: 2839 |
| Download times: 2588 |
| Abstract: |
| 为了澄清整群抽样调查数据分析中正确计算抽样误差的必要性，以在某市15岁及以上人群中开展的一次两阶段整群抽样调查为例，分别采用适用于单纯随机抽样数据的方法和考虑了复杂抽样设计的方法对数据进行分析.结果 显示，忽略对复杂抽样设计的考虑，不恰当的采用适用于单纯随机抽样数据的方法进行数据分析，不仅有可能大大低估样本统计量的抽样误差，在进行假设检验时，甚至会得到错误的结果，故正确分析和报告整群抽样调查数据的抽样误差是非常必要的。 |
| English Abstract: |
| Objective To explore the factors related to the prognosis of colorectal cancer (CRC) and to establish a prognostic model for evaluating the prognosis of the patients with CRC. Methods 370 cases with CRC were selected in the study and clinical/pathological factors were collected and patients were followed. Kaplan-Meier method was used to calculate survival rate. Log-rank test and proportional-hazards regression model (Cox model) were used for univariate and multivariate analysis. Log cumulative hazards function plot was used to test Cox model proportional-hazards assumption (PH assumption).Prognostic index (PI) was calculated based on the results of multivariate analysis. Results (1) One-year, three-year and five-year survival rates were 90.5%，78.396 and 76.5%respectively. (2)Lymphatic metastasis, Duckes classification and therapeutic measure were independent prognostic factors of CRC and all passed PH assumption.(3)Patients with different PI were classified into 3 groups and there were significant differences noticed in survival rates (P < 0. 001).(4) Individual survival rate was evaluated based on the prognostic Cox model and Pl. Conclusion Lymphatic metastasis, Ihickes classification and therapeutic measure were independent prognostic factors of CRC. To test PH assumption of the factors, selection of Cox model was essential. Cox model and PI seemed to be available in predicting the long term survivrate of patients with CRC. |
|
View Fulltext
Html FullText
View/Add Comment Download reader |
| Close |