After scanning the SWOV paper, I was struck by this
It also appears from the article that the collection of data about light usage by cyclists involved in bicycle accidents during dark hours is missing. Law enforcement just may not include the use of lights in their reports so the data is missing. Without data, the problem can't be studied and if the problem has never been studied, there are few conclusions you can draw.It has never been sufficiently studied whether carrying bicycle lights by cyclists actually benefits road safety; therefore, no clear-cut conclusions can be drawn about this issue. Moreover, carrying bicycle lights is not registered as a standard in the case of crashes involving cyclists. It would be relevant to study the research question whether bicycle lights have an effect on road safety...
I also put the lack of study of lights no bicycles into the "well, duh!" category of science. It's a trivial hypothesis that doesn't really deserve much study. Lights on a bicycle at night have the dual purpose of being able to see where you are going and being seen by other road users. The latter is probably of less importance than the former but if your lights are bright enough to see the road, they are bright enough for other road users to see them. On the other hand, if your lights aren't bright enough to see the road by, they are probably ineffective as a signal to other road users in urban environments.
I don't think that I'm going to volunteer as part of a blind test to see if using lights is a good idea or not. As I said above, it's most a "well, duh!" conclusion.