论文部分内容阅读
本文要点如果比数比被解释为相对危险性的话,它将会夸大了所有的效应:小于1的比数比较相对危险性小,大于1的比数比较相对危险性大。当初始危险性增加且比数比偏离1时,夸大程度也随之增加。不过,只有在大效应作用于高初始危险性的人群时,比数比和相对危险性之间才会有严重的差别,所以,将比数比解释为相对危险性的定性判断不一定会有严重错误。在表明危险性减少(比数比小于1)的研究中,比数比低估相对危险性的程度从不会超过初始危险性的水平。
This article points out that if the odds ratio is interpreted as relative risk, it will exaggerate all effects: ratios less than 1 are less dangerous than relatives, and odds greater than 1 are relatively dangerous. When the initial risk increases and the odds ratio deviates from 1, the degree of exaggeration also increases. However, only when the large effect affects people with high initial risk, is there a serious difference between the odds ratio and the relative risk, so the qualitative judgment that the odds ratio is interpreted as relative risk may not necessarily Serious error. In studies showing a reduced risk (odds ratio less than 1), the odds ratio underestimates the relative risk to never exceed the level of initial risk.