Letter to NHTSA NCSA on FARS Underreporting of Fire Deaths
October 1, 2002
Joseph S. Carra
National Center for Statistics and Analysis
National Highway Traffic Safety Administration
400 7th Street, S.W.
Washington, D.C. 20590
Dear Mr. Carra:
As we discussed in our meeting earlier this year, CAS has long been concerned about fatal crashes that do not show up in FARS and/or FARS data searches. For example, in the late 1970's and early 1980's, we discovered some Ford Park-to-Reverse fatalities were not in FARS. During our work on General Motors C/K pickup fire crashes, we also noted that FARS searches did not produce all the fatal fire crashes about which we were aware, excluding those burn deaths that occurred more than 30 days after the crash. Recently, we have been tracking fire fatalities in Ford Crown Victorias, and have found fatal crashes that are either missing from FARS or are not coded with fire/explosion as the most harmful event. These crashes are listed below.
|12/11/01||Baldwin Co., AL||10847||Smith|
The Louisiana crash (Attachment A) and two
Arizona crashes (Attachments B, C,
were not returned in a full field dump of fatal crashes involving Ford
Crown Victorias with fire/explosion as most harmful event or fire occurred
in vehicle during crash. We were also unable to locate them with the FARS
online query system, using information from accident reports to search
the FARS database.
The Alabama crash (Attachment F) was returned in the full field dump of preliminary 2001 data. However, both the harmful event and most harmful event are listed as "tree." Based on the autopsy reports prepared by Alabama Department of Forensic Sciences(Attachments G through I), smoke inhalation and thermal injury were the cause of death for all three victims. Although the FARS Analytic Reference Guide, p. V-34, states "one can not assume that the most harmful event for a vehicle was the cause of death or injury for any specific individual within the vehicle," in this case all fatalities in the vehicle were caused by fire. As indicated below, attached are all accident and autopsy reports in our possession that relate to the above listed crashes.
When dealing with a relatively small universe such as vehicle crash fires, omission or misclassification of crashes will significantly affect policymaking based on crash data such as FARS. In this instance, the 6 missing fire deaths would constitute 25% of all MY 1992-2001 Crown Victoria occupant fatalities from 1993-2001 with a most harmful event of fire/explosion. Furthermore, as the 6 missing fire deaths all occurred in the Crown Victoria Police Interceptor model, they would constitute approximately 42% of all MY 1992-2001 Crown Victoria Police Interceptor fatalities during the same time period. Even more troubling is the fact that CAS has no assurance that it has found the universe of Ford Crown Victoria fatal crashes, so the true picture is still unknown.
We urge you to view this as an opportunity to examine the records of these four fatal fire crashes with six burn deaths and determine why they did not appear correctly in FARS. We also urge you to go back and correct FARS so that policy analysts have a more accurate data base on which to make their decisions. Your response to the issues raised in this letter is requested.
Michael Brooks - Legal Staff
Attachments (in .PDF)
A - Louisiana Accident Report (Le)
B - Arizona Accident Report (Cruz)
C - Pima County, AZ Autopsy Report (Cruz)
D - Arizona Accident Report (Fink)
E - News
F - Alabama Accident Report (Davis, Smith, Thibodeaux)
G - Alabama Dept. of Forensic Sciences Autopsy
H - Alabama Dept. of Forensic Sciences Autopsy
I - Alabama Dept. of Forensic Sciences Autopsy
1 For Crown Victoria model years 1993-2002, 1992-2001
FARS data currently lists 18 occupant fatalities where most harmful event
is fire/explosion. The six occupant fatalities described in this letter
would raise the total to 24.
2 For Crown Victoria Police Interceptor model years 1993-2002,
1992-2001 FARS data currently lists 8 occupant fatalities where most harmful
event is fire/explosion. The six occupant fatalities described in this
letter would raise the total to 14.