The Fatality Analysis Reporting System as a tool for investigating racial and ethnic determinants of motor vehicle crash fatalities

https://doi.org/10.1016/j.aap.2005.03.006Get rights and content

Abstract

The Fatality Analysis Reporting System (FARS) is a Department of Transportation database in the public domain that contains detailed information about fatalities resulting from motor vehicle crashes on public roadways in the United States since 1975. However, data on race and Hispanic ethnicity were not collected by FARS until 1999. Since then, completeness of reported racial and ethnic information has varied from State to State. To assess utility of FARS for investigating race- and ethnicity-specific risk factors associated with motor vehicle crash mortality, we examined yearly national and State-specific reporting rates of race and Hispanic ethnicity for 168,863 motor vehicle crash fatalities from 1999 to 2002. In 1999, national reporting was 85% for race and 78% for Hispanic ethnicity. Over the 4-year study period, a significant linear increase in annual reporting for both race and Hispanic ethnicity was evident at the national level, as reporting by individual States improved over time. In 2002, national reporting rates reached 90% for race and 88% for Hispanic ethnicity. Our findings indicate that FARS has become a valuable resource for population-based studies of motor vehicle crash mortality disparities that exist among racial and ethnic subpopulations in the United States.

Introduction

Motor vehicle crashes account for over 42,000 fatalities and more than 3 million injuries annually (Vyrostek et al., 2004). Numerous reports show that racial and ethnic minority populations in the U.S. are at greater risk of dying in a motor vehicle crash than their majority counterparts (Braver, 2003, Daniels et al., 2002, Harper et al., 2000; Schiff and Becker, 1996). However, efforts to identify preventable determinants of these disparities have been limited by a lack of racial and ethnic data available from the Fatality Analysis Reporting System (FARS), the nation's most comprehensive population-based archive of information about motor vehicle crash fatalities.

FARS was created in 1975 by the Department of Transportation's National Highway Traffic Safety Administration (NHTSA) as a repository of information about motor vehicle crashes on U.S. public roadways (NHTSA, 2004a). Until 1987, information on race and Hispanic ethnicity was not available through FARS, and until 1999 these data elements were not collected by FARS, nor were files containing them available in the public domain (Judy Hilton, personal communication).

For the period from 1987 to 1998, annual FARS files were supplemented with data on race and Hispanic ethnicity obtained by linking death certificate number, death state, death date, age, sex, person type and county with the multiple cause of death (MCOD) file maintained by the National Center for Health Statistics (NCHS) (Morgan, 1999; NHTSA, 2004a). Linked FARS-MCOD files have been used by NHTSA staff, NHTSA contractors and a small number of researchers in academia and industry for studies of racial and ethnic disparities in crash mortality that have been published in NHTSA technical reports (Voas et al., 2000, Morgan, 1999) and peer-reviewed scientific journals (Braver, 2003, Mayrose and Jehle, 2002, Voas et al., 2002). Per an agreement between NHTSA and NCHS, linked FARS-MCOD files are not generally available in the public domain, as are other files (Judy Hilton, personal communication).

In the absence of death certificate numbers to facilitate direct linkage with the NCHS MCOD file, researchers studying motor vehicle mortality disparities among racial and ethnic minority populations in individual States have probabilistically linked race and ethnicity data from the MCOD file using date of death, State of death, age, gender, and other data elements available from FARS public domain files, often supplemented with data from State-level sources such as trauma registries (Fife, 1989, Campos-Outcalt et al., 1997, Campos-Outcalt et al., 2002, Campos-Outcalt et al., 2003, Harper et al., 2000). Apart from potentially prohibitive costs of such linkage for large study populations, a major limitation of this approach is that only about 75 to 85% of FARS fatalities can be matched to the MCOD (Fife, 1989, Campos-Outcalt et al., 1997, Campos-Outcalt et al., 2002, Campos-Outcalt et al., 2003, Harper et al., 2000). The problem of incomplete linkage is compounded in studies of racial and ethnic disparities in mortality because MCOD linkage rates differ by race and Hispanic ethnicity, as well as by gender within racial and Hispanic ethnic subpopulations (Horm, 1999).

Regardless of how linkage is accomplished, another limitation of linked FARS-MCOD data is the considerable time lag preceding completion of the linkage process and availability of data for use by researchers. In 2005, FARS data had only been linked with the MCOD file through 1998 (Judy Hilton, NHTSA, personal communication). One component of the lag reflects time required for NCHS to receive and process death certificate data from individual States, and subsequently finalize annual MCOD files (NHTSA, 2004b). In addition, some States require detailed paperwork and State review board approvals before authorizing NHTSA to link FARS death certificate numbers with NCHS MCOD data (NHTSA, 2004b). The lengthy delay precludes use of linked FARS-MCOD data for surveillance of current race/ethnicity-specific trends in motor vehicle crash fatalities. Recent studies that have used linked files to investigate racial/ethnic differences in crash mortality at the national level have relied on data up to a decade or more out of date at the time of publication (Braver, 2003, Mayrose and Jehle, 2002, Voas et al., 2002).

Starting with calendar year 1999, information on race and Hispanic ethnicity has been collected directly by FARS (Judy Hilton, NCSA/NHTSA, personal communication) and has been included in files available in the public domain. The timely availability of information on race and Hispanic ethnicity directly from FARS holds promise that the FARS database can serve as a methodologically robust resource for U.S. population-based studies of racial and ethnic disparities in motor vehicle crash fatalities. However, because national data are derived from information voluntarily reported by individual States, the utility of FARS for studies at the national level depends on the completeness with which racial and Hispanic ethnic data are reported by each State. To this end, we investigated both national and State-specific trends in reporting on race and Hispanic ethnicity using FARS data for 168,863 motor vehicle crash fatalities occurring in the 50 States and the District of Columbia between January 1, 1999 and December 31, 2002.

Section snippets

Overview of Fatality Analysis Reporting System

FARS is administered by the National Center for Statistics and Analysis of the National Highway Traffic Safety Administration (NHTSA/NCSA). Under a cooperative agreement with NHTSA/NCSA, an agency in each State government provides information on all motor vehicle crashes on public roadways in that State resulting in a fatality within 30 days of the crash (NHTSA, 2004a). FARS data are collected, coded and transmitted by designated State employees, called FARS Analysts, who have undergone formal

Reporting of race

For all 50 States and the District of Columbia combined, information on race was reported for 85.5% of fatal crashes during the 4-year period from 1999 to 2002 (Table 1). Average annual reporting rates for race showed a significant linear increase from 84.6% in 1999 to 89.9% in 2002 (p < 0.0001). In all four calendar years, reporting rates were over 80% in 36 States and over 90% in 33 States (Table 2). Fourteen States plus the District of Columbia had race reporting rates of under 80% in one or

Discussion

More than half of all US States have reported race and Hispanic ethnicity to FARS annually for over 90% of fatalities since racial and Hispanic ethnic data elements were added to the FARS reporting protocol in 1999. While reporting in many States was poor in 1999, by 2002 only 12 States did not report race and Hispanic ethnicity for at least 90% of fatalities. The only States that appeared to be making little effort to report complete information were Alaska, Hawaii, Idaho, Indiana, New York,

Acknowledgments

This work was supported in part by a grant from State Farm®.

The authors wish to thank Barry Eisemann and Judy Hilton of the National Center for Statistical Analysis, National Highway Traffic Safety Administration and Kenneth D. Hackman of the National Institute for Safety Research, Inc. for their assistance with navigating FARS and the following State Farm personnel for their ongoing support: Lonnie Smith, Manager, Community Alliances and Clayton Adams, Vice President of Community Alliances.

References (27)

  • J. Horm

    Multiple causes of death for the National Health Interview Survey

  • J. Mayrose et al.

    An analysis of race and demographic factors among motor vehicle fatalities

    J. Trauma

    (2002)
  • Morgan, C., 1999. Effectiveness of lap/shoulder belts in the back outboard seating positions, Washington, DC, National...
  • Cited by (39)

    • Crash report data analysis for creating scenario-wise, spatio-temporal attention guidance to support computer vision-based perception of fatal crash risks

      2021, Accident Analysis and Prevention
      Citation Excerpt :

      This study is motivated to develop an approach that uses fatal crash report data to enhance CV's ability to perceive fatal crash risks earlier. Studies of fatal crashes have widely used FARS (Clark and Cushing, 2002; Briggs et al., 2005; Yasmin et al., 2015). Analyses of crash report data provide insights into safety enhancement (e.g., USDOT, n.d.b; USDOT, n.d.c; CUBR, n.d.; FHWA, 2018) and support the safety strategy development and law enforcement (e.g., Abdel-Aty et al., 2011; Harb et al., 2008; Pour-Rouholamin and Zhou, 2016; Sharma et al., 2016).

    • Reducing traffic violations in minority localities: Designing a traffic enforcement program through a public participation process

      2018, Accident Analysis and Prevention
      Citation Excerpt :

      In many countries, non-dominant minorities (e.g., members of ethnic or racial minority groups) tend to be over-involved in delinquent and risky behavior and to be over-represented in road traffic crashes and traffic violations (see, e.g., Marshall, 1997; Department for Transport, 2001; Briggs et al., 2005; Hilton, 2006; Factor et al., 2008; Bui, 2009; Veen et al., 2011; Peterson, 2012; Andersen, 2015; Coughenour et al., 2017; Factor, 2018).

    • How Many People Were Killed by Windblown Dust Events in the United States?

      2023, Bulletin of the American Meteorological Society
    View all citing articles on Scopus
    View full text