Ex-urban sprawl as a factor in traffic fatalities and EMS response times in the southeastern United States.

AuthorLambert, Thomas E.

Many different writers (Atkinson and Oleson 1996; Barnett 1995; Burchell and Lisotkin 1995; Burchell et al., 1998; Carruthers and Ulfarsson 2002; Ciscel 2001; Ewing 1997; and Glaeser and Kahn 2003) have examined the direct and indirect costs of unplanned growth or sprawl. However, an area only recently examined is the impact of sprawl on traffic fatalities (Ewing, Schieber and Zegeer, 2003; Lucy 2003; and Lucy 2000). Besides a case study of the Chicago area, which found emergency medical services (EMS) delays due to sprawl (American Farmland Trust 1998), another issue not examined on a larger scale is the degree to which sprawl might be contributing to delays in EMS. In this research note, we develop models similar to the ones used by Reid Ewing, Richard Schieber and Charles Zegeer (2003), Stefan Felder and Henrik Brinkmann (2002) and Theodore Keeler (1994) in order to assess the impact that the built environment has on EMS response times and the rate of traffic fatalities in the southeastern United States.

Reid Ewing (1997) reviews 17 studies concerning sprawl and identifies four characteristics defining it: low-density, strip development, scattered development, and leapfrog development. Peter Gordon and Harry Richardson (1997), in their criticism of planners who promote "compact cities," suggest that sprawl is low density, dispersed, decentralized, polycentric (many centers), and suburban. The universal mobility of the auto has allowed job and home to be miles apart. Americans are driving more every year in large part because of the increasingly spread out nature of our metro areas (U.S. Dept. of Transportation (DOT), National Transportation Statistics 1999). The 2001 National Household Travel Survey reports that although Americans were making fewer trips by motor vehicle, average time per trip had gone up including the commute to work (U.S. Department of Transportation 2004). Edward Glaeser and Matthew Kahn (2003) contend that sprawl is a result of a society that has centered itself on the automobile.

As development continues outward, jobs, housing and services grow farther apart. In the past few decades, development patterns that require an automobile trip for every errand force many to drive more every year to accomplish the same things. The long journey to work or for shopping is now accepted as commonplace. Due to families having the luxury of several automobiles, many of these trips (over 81%) are one-person occupied (U.S. DOT 1999). Ewing, Schieber, and Zegeer (2003) and Keeler (1994) show that higher population density is associated with lower traffic fatalities on a per capita basis. Ewing, Schieber and Zegeer create a "sprawl index" demonstrating that more sprawled metro counties (i.e., those having low general population density, large/long block sizes, and census tracts with population densities below 2,500) have higher traffic fatality rates than their less sprawled counterparts. Also, the more sprawled an area becomes the more difficult for police, fire and EMS to reach many new households and new developments, even those along existing roadways. The alternative is to build new facilities closer to the new developments, which raises the costs of public service provision.


As a measurement of the consequences of sprawl, William Lucy (2003) constructs an index measuring the likelihood of someone becoming a traffic or homicide fatality statistic in different parts of a metro area. He finds higher traffic fatality and homicide rates in ex-urban areas than those in central cities or the inner suburbs of fifteen metro areas. (1) Similarly, Reid Ewing, Rolf Pendall, and Don Chen (2002) find that traffic fatalities are much higher in what they have ranked as the top ten most sprawling metro areas versus the ten least sprawling metro areas in the United States: fifteen average annual traffic deaths versus nine average annual traffic deaths per 100,000 residents.

For this paper, we first looked at fatal traffic crashes and then average EMS run times (from time of notification to arrival of an EMS unit) to an accident site for the year 2002 in the metro areas of eight states that make up the United States Environmental Protection Agency Region 4: Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, and Tennessee. Fatal traffic crash and corresponding EMS time data were found for most of the counties that make up the metro areas of these states. The metro area boundaries were those that were in existence as of 2002 (U.S. Census Bureau, County and City Data Book, 2000).

Using the U.S. National Highway Traffic Safety Administration's (NHTSA) Fatal Accident Reporting System (FARS), fatal traffic crashes and EMS times were divided as having occurred in two types of areas: urban and rural, or ex-urban. FARS uses the Federal Highway Administration's (FHWA) definition of urban and rural. This definition follows the U.S. Bureau of the Census definition of urban but "allows responsible state and local officials in cooperation with each other, and subject to approval by the Secretary of Transportation, to adjust the Census boundaries outward, as long as they encompass, at a minimum, the entire Census designated area" (http://www.fhwa.dot.gov/planning/census/faqa2cdt.htm, U.S. DOT 2003). All else is defined as "rural." With the 2000 Census, the U.S. Census Bureau classifies populations and land areas as "urban," "rural, non-farm" and "rural, farm" according to population density. Before 2000, the definition of urban hinged upon municipal incorporation and total population, not on population density thresholds. For the 2000 Census, an urbanized area or urban cluster "consists of core census block groups or blocks that have a population density of at least 1,000 people per square mile and surrounding census blocks that have an overall density of at least 500 people per square mile" whereas the classification "'rural' consists of all territory, population, and housing units located outside of urbanized areas and urban clusters" which means a population density below 500 per square mile (http://www.census.gov/geo/www/ua/ua_2k.html). A census block group and a census block are sub-components of a census tract and are smaller than an actual census tract. FARS therefore follows the Census Bureau's designation of what is considered urban and rural in each area as far as population density is concerned unless state and local officials have gone beyond urbanized areas and urbanized cluster boundaries.

All of the Metropolitan Statistical Areas' (MSA) urban core counties (those with cities of population of 50,000 or more) for which we had data, had "rural" fatal crashes and "rural" EMS run times. (2) For example, in 2002, Miami-Dade County, Florida had 226 fatal crashes classified as urban and 83 classified as rural or ex-urban. According to FARS, the urban EMS time from notification until arrival was 5.9 minutes for urban fatal crashes and 7.2 minutes for rural or ex-urban accidents for Miami-Dade. In 2002, the U.S. average EMS response time from notification to arrival was 6.51 minutes for urban fatal crashes and 12.11 minutes for rural fatal crashes (NHTSA FARS 2002). For the metro counties studied in our paper, the corresponding EMS times were 7.6 minutes for urban and 10.7 minutes for ex-urban from notification to time of arrival. For fatal crashes per 10,000 population, the average was 2.5 for urban fatal crashes and 6.3 for ex-urban fatal crashes.

Like other studies, the number of fatal traffic crashes was adjusted on a per 10,000 population basis. Again, using Miami-Dade County as an example, approximately 81% of the county's land mass is classified as rural, non-farm because of low population density whereas in Louisville-Jefferson County, Kentucky, around 32% of the county's land mass is identified as rural, non-farm. The 226 urban, fatal crashes that occurred in Miami-Dade County were divided by the total population per 10,000 of Miami-Dade County classified as urban, whereas the 83 rural, fatal crashes were divided by the total population per 10,000 of Miami-Dade County...

To continue reading

Request your trial