Here's my analysis of US national bicycle safety, published in 2010.* Nothing significant has changed since then.

***
HOW DANGEROUS IS CYCLING?
Surprisingly, cycling can be argued to be "safe enough", given only
that one is willing to count the intangible benefits of health through
exercise, generally acknowledged as substantial. Here I shall make no
effort to quantify those health benefits because the argument I'm
putting forward is conclusively made by harder statistics and
unexceptional general morality.

In the representative year of 2008, the last for which comprehesive
data is available, 716 cyclists died on US roads, and 52,000 were
injured.

Source: National Highway Traffic Safety Administration

The most convenient way to grasp the meaning of these statistics is to
compare cycling with motoring, the latter ipso facto by motorists'
average mileage accepted by most Americans as safe enough.

Compared to a motorist a cyclist is:
11 times MORE likely to die PER MILE travelled
2.9 times MORE likely to die PER TRIP taken

By adding information about the relative frequency/length/duration of
journeys of cyclists and motorists, we can further conclude that in
the US:

Compared to a motorist, a cyclist is:
3 to 4 times MORE likely to die PER HOUR riding
3 to 4 times LESS likely to die IN A YEAR's riding

Source:
http://www.google.com/url?sa=D&q=htt...Wt7vubH xju7Q

It is the last number, that the average cyclist is 3 to 4 times less
likely to die in a year's riding than a motorist, and enjoys all the
benefits of healthy exercise, that permits us to ignore the greater
per mile/per trip/per hour danger.

***

Andre Jute
* A complete version is at
https://groups.google.com/forum/#!ms...ch/qOFCNhQ1428 . I used the best national figures available then, referring to 2008, but just about nothing has changed since then.