Well firstly it's not just that one paper, but that paper presents it well. Secondly that paper presents data from thousands of patients and happens to present it in a graph that I find well done and good at communicating the issues.
You should read it, it is the results of a study of 4,202 patients, so quite statistically sound in size.
https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/415179
you will note that its not alone, and in the literature you'll find reasonably good accord with their findings.
So you have any research which supports more than your concern , which actually shows significantly more bleed events in the range we are talking?
You are implying that it's my recommendation that remaining at INR = 4 is the goal, when the reality is I write that you should target your INR according to your valve type and mention this graph to provide reassurance to those who are concerned about what their INR is (when our of range) that the situation is really n it dangerous and to not great and to steer back to their target.
But you seem unable to grasp that subtle difference.
You ignore my point that for a mitral valver 4 is just outside their normal range....