| SLR与HLR模型在评估肠道清洁药物有效性中的比较 |
| |
| 引用本文： | 胡一睿,王杨,李卫. SLR与HLR模型在评估肠道清洁药物有效性中的比较[J]. 疾病控制杂志, 2011, 15(2): 153-155 |
| |
| 作者姓名： | 胡一睿 王杨 李卫 |
| |
| 作者单位： | 1. 中国医学科学院阜外心血管病医院卫生部心血管病防治中心生物统计部,北京,100037;北京师范大学数学科学学院,北京,100875
2. 中国医学科学院阜外心血管病医院卫生部心血管病防治中心生物统计部,北京,100037 |
| |
| 摘 要： | 目的探讨传统Logistic回归(standard logistic regression,SLR)模型和多水平Logistic回归(hierarchical logistic regression,HLR)模型的差异及应用。方法以一个由234例受试者参加的多中心随机对照临床试验为例,分别建立SLR和HLR模型,并对两个模型的结果进行分析和比较。结果经空模型检验,数据存在组内相关(ICC=5.58%)。通过HLR模型与SLR模型给出的参数比较发现,SLR认为正常体重与低体重人群相比,肠道清洁质量变差OR=0.30;95%CI=(0.10,0.93)。而HLR给出的结果则认为正常体重与低体重人群相比,肠道清洁质量无差异OR=0.30;95%CI=(0.09,1.01)。结论多水平模型对于处理聚集效应的数据具有优势,当数据存在分级结构或聚集效应时,HLR考虑了数据聚集效应导致的组间方差变化,可信区间的估计更为保守,减少了第I类错误发生的概率,在流行病学资料的处理中具有一定推广性。
|
| 关 键 词： | 模型 理论 Logistic模型 统计学 非参数 |
The comparison between SLR and HLR model in assessing the effectiveness of bowel cleansing drugs |
| |
| HU Yi-rui,WANG Yang,LI Wei. The comparison between SLR and HLR model in assessing the effectiveness of bowel cleansing drugs[J]. Chinese Journal of Disease Control and Prevention, 2011, 15(2): 153-155 |
| |
| Authors: | HU Yi-rui WANG Yang LI Wei |
| |
| Affiliation: | HU Yi-rui1,2,WANG Yang1,LI Wei1. 1.Department of Biostatistics,Cardiovascular Institute of Fuwai Hospital,National Center for Cardiovascular Diseases,Chinese Academy of Medical Science,Beijing 100037,China,2.School of Mathematical Sciences,Beijing Normal University,Beijing 100875 |
| |
| Abstract: | Objective To illustrate the application of hierarchical logistic regression(HLR) in assessing the risk factors of enema procedure,and compare the results with standard logistic regression(SLR).Methods A multi-center randomized control trial which recruited 234 participants was used as an example.SLR and HLR were constructed respectively.Results The clustering effect was detected(ICC=5.58%).Most estimation was similar between the two models.SLR indicated that compared with low weight,normal weight was a risk... |
| |
| Keywords: | Models theoretical Logistic models Statistics nonparametric |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
