Thanks for bringing this article [https://bmjopen.bmj.com/content/9/6/e028655] to our attention, Steve! It's a useful study.
I do think inclusion of Science and Nature odd, as they are not primarily medical journal; I also think the baseline Journal Impact Factor of 15, while useful to control the data set in a useful way for this study, eliminates the vast majority of journals of interest to most. I work primarily with the vision community; the top-ranked 'ophthalmology' journal in the 2018 JIF rankings, recently released, is just 11.768, and is the only one over a JIF of 10. The reality is that the potential users and uses of many journals and even, as with vision, entire fields of research, is too small to generate enough citations to reach anything close to an average of 15 within 2 years of publications.
HIFA profile: Pamela Sieving is a special volunteer at the National Eye Institute/National Institutes of Health, and an independent consultant in biomedical information access; she works primarily in the vision community to increase access to information needed to preserve and restore vision. pamsieving AT gmail.com