(insert bike credentials here)
For me one of the largest (if not THE largest) factor* in bike safety is route choice, and the main factor in that for me is density and speed of car traffic. As comfortable as I feel "vehicular" cycling (e.g. cutting across lanes to take a true left turn, taking the lane to avoid the "dooring zone" along parked cars), those are the moments when I am most exposed and at risk. Likewise the safest intersections are those that do not have cars passing through!
As an example, a colleague of mine commutes from near my house to our office, and we take very different routes - him along a major artery with (mostly) protected bike lanes, me along back streets with sharrows or no bike markings at all. Even in a protected line, he has to contend with fast and constant traffic alongside him, frequent intersections where he has no protection and is even harder to see (since he's popping out from behind a line of parked cars); the intersections are constantly in use and drivers turn and accelerate faster to make smaller windows in the faster traffic. My commute along back streets is leisurely; I can ride in the middle of the street without fear of getting doored or getting overtaken at speed unexpectedly, and intersections are calm. Of course the major artery is more direct (and so perhaps slightly faster), but it's also a major designated bike route and for that reason sees a lot of bike traffic even though I think it's less safe.
Of course in denser neighborhoods and cities it may be difficult to find "back streets" with minimal car traffic and which still get to where you're going. Still, I think that should be a goal of new bike infrastructure development - I'm happy to act like a car, make turns from the correct lane, stop at red lights, signal, etc., and I don't need a protected bike lane either - a normal road with light, slow car traffic is fine. A set of these which connects the major neighborhoods in a community (e.g. running parallel to arteries but not quite in the central areas) makes for a very nice bike experience (c.f. Berkeley's bike boulevards which is naturally my inspiration here!).
So, when this article mentions the requirement of protected bike lanes on all streets, while I think that's great and I love seeing bicycling given more consideration in urban planning, I also think that it seems a little coarse. Some roads don't need to have any bike traffic at all, some roads can be wonderful for biking without any modification at all, and some roads ought to be made better for biking but in more nuanced ways (primarily traffic easing - roundabouts, narrower roadways, removing through-access, etc - that discourages car traffic without impeding bike traffic). Of course if you want to build independent bike trails (or non-grade lanes, or other more significant changes) that's wonderful too!
Other important factors IMO:
Cyclist density - everyone should bike! Probably preaching to the choir here but there's safety in numbers, drivers get used to seeing cyclists and how they behave, and it makes it easier to advocate for better infrastructure. Also relevant on a micro-level for route choice
* Cycling experience/behavior - including "vehicular" maneuvers, general awareness of dangerous/risky situations and driver behavior
* Lights