LIMITATIONS OF THE LOGISTIC ANALYSIS OF EPIDEMIOLOGIC DATA
LIMITATIONS OF THE LOGISTIC ANALYSIS OF EPIDEMIOLOGIC DATA
复制标题
DOI：
10.1093/oxfordjournals.aje.a112849
复制
发表时间：
1979-01-01
影响因子：
5
通讯作者：
GREENLAND, S
中科院分区：
文献类型：
--
作者：
GREENLAND, S
在过去的几年里，病例对照研究的多元技术得到了广泛的发展和传播。这些努力大多集中在使用逻辑模型或密切相关的模型来推导对数优势比(1-4)的线性模型。在两个研究暴露X和Y的情况下，X=*和Y=y时的患病几率与X=X0和Y=Jo时的患病几率之比(即，疾病几率比)通过方程建模
Within the last few years there has been extensive development and dissemination of multivariate techniques for case-control studies. Most of these efforts have focussed on using logistic or closely related models to induce a linear model for the log odds ratio (1-4). In the case of two study exposures, X and Y, the ratio of the odds of disease whenX=*, and Y= y, to the odds of disease when X= x0 and Y= Jo (ie-, the disease odds ratio) is modeled via the equation