Streetsblog recently had a post featuring "a map of every traffic fatality in the U.S. between 2004 and 2013, using data from the National Highway Traffic Safety Administration’s Fatality Analysis Reporting System (FARS)." They entitled it "Every Traffic Fatality in the US Mapped." Sadly, that's not true. It isn't EVERY traffic fatality mapped. It's probably as close as one can get without a lot of sleuthing, but there are limits to FARS data that means that some traffic fatalities aren't included. Between 2005-2013, in the DC area, there have been 34 cyclist fatalities reported in FARS.
First of all, FARS only includes traffic fatalities that involve a motor vehicle. That means that single bike crashes like this or this aren't counted. Nor are cyclists or pedestrians who are killed in bike-ped, bike-train or bike-bike crashes (etc...). In the DC area from 2005-2013 there are another 13 such bike related deaths, 10 of them cyclists.
FARS also doesn't include fatal crashes on trails or in parking lots or on private roads even if they do involve a motor vehicle. I don't know of any such crashes in the DC area, but they do happen and NHTSA tries to track them as part of their non-traffic crashes tracking (though they don't appear to break them out by person type).
The nontraffic motor vehicle crashes are a class of crashes that occur off the public traffic ways. These crashes, subsequently referred to as “nontraffic crashes,” are mostly single-vehicle crashes on private roads, two-vehicle crashes in parking facilities, or collisions with pedestrians on driveways. Then there are also noncrash incidents such as a vehicle falling on a person underneath or unintentional carbon monoxide poisoning. Both nontraffic crashes and noncrash incidents have the potential to kill or injure people. Nevertheless, the information on neither of these was available until 2007, when Congress required National Highway Traffic Safety Administration (NHTSA) to start collecting and maintaining information pertinent to these events. Complying with the directive, the NHTSA designed and implemented a virtual data collection system, NiTS that would provide counts and details of the fatalities and injuries to people involved in nontraffic crashes and noncrash incidents. This Crash•Stats focuses only on nontraffic crashes and presents some salient statistics about occupants and nonoccupants killed and injured in such crashes during 2008–2011.
On an average, 1,621 people were killed each year in nontraffic motor vehicle crashes during the four year period, 2008 to 2011. About 39 percent of these people were nonoccupants such as pedestrians and bicyclists.
Then there are crashes that do involve cars that just, for whatever reason, are not reported. It doesn't seem that the Park Police or other federal police departments report FARS data, and other crashes - like this one - just fail to be reported in FARS for unknown reasons. Between 2005-2013 in the DC area there were 5 such fatal bike crashes.
So FARS data, while being the best data set out there and the one used for the annual report on bicycle fatalities, is drastically under reporting bicycle fatalities. If my very limited analysis of 7 jurisdictions is representative of the entire United States, FARS is only capturing about 70% of fatal bike crashes. Which means that the annual number of about 700 fatal bike crashes that NHTSA reports each year should really be more like 1000.
I suspect car and pedestrian fatalities are also under-reported, though not similarly for cars - since fatal single car crashes are reported.
NYC, by the way, has been required to keep track of and report on bike-bike and bike-ped crashes since 2012. Perhaps FARS should expand to include those, (and make all law enforcement agencies - even the Park Police - file reports on all fatalities).
One last thing to consider, while looking through 2013 data, I found a FARS report of a bicycle fatality on October 3, 2013. It was on Pennsylvania Avenue (SR-4) at Marlboro-Ritchie Road adn I can't find any report of it in the media. (According to the FARS report, the cyclist crossed illegally (jaywalking)). Having a crash reported in FARS, but not in the media, happens from time to time. If there are some fatalities that are only reported in FARS and others that are only reported in the media, it stands to reason that there are some fatal bike crashes - not very many I suppose - that are neither reported in FARS or the media. This is probably most likely in crashes where the victim dies more than 30 days after the crash, but probably happens even when that isn't true.