Dear all,
In response to the comments about overlapping CIs and dose-response trend, Prof. Li sent us an article for reference with his explanations: “Another paper by Chang YH regarding the dose-response relation between alcohol drinking and 30-day mortality among driver victims after motor vehicle crash. Figure 2 show very significant linear trend (p＜0.0001), but the CIs for three point relative risk estimated are all overlapped. This example also demonstrates that it is not appropriate to use overlapping or not-overlapping of CIs to declare dose-response trend." If you have any questions, please feel free to start a discussion in this week’s thread.
Thank you and have a nice weekend.
Nicole.