So I don't know where how that "34" I quoted earlier is justified, since I'm coming up with 29.
29 fatalities per 554 is 5% of the population, over 50 years of space flight. My hand-wavy calculation to get to "deaths per 100,000 per year is 29 fatalities / 554 people / 50 years * 100000 = 105/100K. (Previously I did it as 34 / 450 / 50 * 100000 = 151)
The numbers I'm computing are the effective deaths per 100,000 per year. That's different than the number you are calculating, which is the number of fatal launches per year, I think.
Let's suppose that the active period of an astronaut is 10 years. The number of astronaut selected in the last 10 years is approximately 70 (counting by hand from http://en.wikipedia.org/wiki/List_of_astronauts_by_year_of_s... ). Approximately only one half of them fly each year (~30 in 2009, ~25 in 2010 : http://en.wikipedia.org/wiki/List_of_human_spaceflights,_200... )
So, there are approximately 30 / 50 dead per year, from a population of 70 active astronauts, that is ~1%, or using a population of 30 flying astronauts we get a 2%.
"only 339 candidates have been selected to date", but a couple of those candidates finished the program but did not go on with a career as an astronaut (one went into nuclear medicine instead).
"Twelve members of the Astronaut Corps were killed during spaceflight, Space Shuttle missions STS-51-L and STS-107. An additional seven were killed in training accidents. Sonny Carter died in a plane crash while traveling on NASA business."
So another way to reckon this is as 19 dead total from a total population of 339. 19/339 = 5.6% for the US program.
Assuming an average of 10 years, (1-p)10=(1-0.056) gives the probability of a person per year, which is 0.57%, or 580/100K/year.
This is suspicious similar to the 1/50 (or 1/100) crash probability that Feynman gets in his report about the Challenger disaster. If 1 in 50 of the ships explodes, you have a 1 in 50 chance to die.
The 60 yeas space flight period cancels out , if we suppose that none of them flied twice the same year. We have and average of 25 space travel seats per year and .55 dead per year.
You get a 7.7% = 7700 per 100K, but didn't take in account that most astronauts flight many times.