The size of the military and the relative frequency of the various types of casualties vary over time, the meaningful numbers are casualties in each category per person or man-year. The only chart I could find quickly does not cover post 2005 casualties, but it's easy to extraplate out to the present as we know the casualty numbers.
One thing that really jumps out when you look at the numbers that way is the success of the US Military in holding down accidental deaths in the last 5 years, something that requires a very serious ongoing commitment to improving operational methods and then training and leading troops to take advantage of them under wartime conditions.
Now suppose you compared them to a similar population of the most dangerous mid-sized cities in the United States, it would be interesting to see how the combat deaths in the field compare to the civilian deaths due to homicide and accidents?
Many of the fatalities are accidental deaths - vehicle rollovers, falling coke machines, training-related deaths, etc. The NCO who died at Ft. Hood during a LandNav course this summer is a classic example of deaths that occur, as is the Colonel I worked for who died of a heart attack during his morning PT run.
TRADOC (training and doctrine command) was making a full-court press effort at reducing accidental and training deaths the last few years of my service - '90-'96. That should serve to explain part of the drop in numbers in the list above during those years.
It may be interesting to note that many of the OEF/OIF deaths are accidental rather than being combat related.