That would take many more years to come up with given the few fires that the AFCI can be expected to prevent. If you apply the fire data, including the fact that 85% of the fires said to be of electrical origin, we could expect that since the adoption of the 2008 code that 1150 fires would have been prevented, assuming that the AFCI would be 100% effective in preventing all fires that are said to be of electrical origin. I did these numbers based on estimated housing starts prior to the downturn, so there would be even less fires to look at. There are an estimated 51,000 dwelling unit fires said to be of electrical origin each year. That would mean that the AFCIs could have been expected to prevent less than 3% of the total fires said to be of electrical origin, and assuming 100% compliance with the NEC, 100% adoption of the NEC, and 100% effectiveness of the AFCI.
I have seen nothing on the costs but it appears you need one of these devices at every point of connection. Even if they are cheap, it will add a lot of labor hours.