I think this is typical lazy journalism, where they take the data at face value only.
I think that the underlying data of cycling slowly has masked the true cause: inexperience...
Although not always true, I think there is a strong positive link between speed and experience. That is to say: faster cyclists are usually more experienced cyclists, with several years of regular cycling on various types of roads under their belts (i.e. how they became fast), knowing how to assert and position themselves so as to be safer. (Although, Zwift may change this!)
Using this relationship, the headline could equally read: "Cycling inexperience raises odds of an accident", which makes sense.
I think that a better experiment would be to take the same cyclist (or even better, a large group of cyclists) and get that same cyclist to ride at different speeds over a period of time to see if there was a notable difference in how they were treated on the roads. This would give a feel for the change in the chance of an accident.
However, what hasn't been mentioned here is the other side of the coin, i.e. the severity of the accident given an accident occurs. I expect this would increase with the speed of the cyclist, e.g. when cars pull out on cyclists at junctions. It is the combination of the chance and the severity of an accident which defines the total risk to the cyclist.
Obviously, there are lots of other factors which are at play too, but these are the ones that seemed prominent to me.