Roadside Assistance Providers Fatally Struck by Vehicles at the

ROADWAY SYSTEMS & DRIVERS

TECHNICAL REPORT

Roadside Assistance Providers

Fatally Struck by Vehicles

at the Roadside: Incidence

and Characteristics

JAN 2024

607 14th Street, NW, Suite 701

Washington, DC 20005

202-638-5944

AAAFoundation.org

Title

Roadside Assistance Providers Fatally Struck by Vehicles at the Roadside: Incidence and

Characteristics

(January 2024)

Authors

Brian C. Tefft

, Anqi Wei

, and Rebecca Steinbach

1. AAA Foundation for Traffic Safety

2. Georgia Institute of Technology

iii

Foreword

When a vehicle breaks down or a minor crash has occurred, drivers and their families

can become stranded on the side of the highway as traffic speeds past. Roadside

assistance providers courageously dedicate themselves to helping motorists during this

time of great vulnerability, often placing themselves in harm’s way to ensure that others

are able to return home safely. Unfortunately, every year, far too many roadside

assistance providers are stuck by passing vehicles and are unable to return home to their

own families.

Since its founding in 1947, the AAA Foundation for Traffic Safety has been dedicated to

preventing injuries and saving lives on our roads. However, as a foundation created and

supported by AAA, finding ways to protect the roadside assistance providers who

dedicate themselves to protecting the rest of us is uniquely personal. This report

describes the results of research that seeks to understand the characteristics of crashes

in which roadside assistance providers have died in the line of duty, to help us find ways

to prevent such tragedies from happening in the future. This report should be of interest

to roadside assistance professionals as well as first responders, road authorities with

responsibility for traffic incident management, policymakers, and all motorists.

C. Y. David Yang, Ph.D.

President and Executive Director

AAA Foundation for Traffic Safety

About the Sponsor

AAA Foundation for Traffic Safety

607 14

Street, NW, Suite 701

Washington, D.C. 20005

202-638-5944

www.aaafoundation.org

Founded in 1947, the AAA Foundation for Traffic Safety in Washington, D.C., is a

nonprofit, publicly supported charitable research and educational organization

dedicated to saving lives by preventing traffic crashes and reducing injuries when

crashes occur. Funding for this report was provided by voluntary contributions from

AAA/CAA and their affiliated motor clubs, individual members, AAA-affiliated insurance

companies, and other organizations or sources.

This publication is distributed by the AAA Foundation for Traffic Safety at no charge, as a

public service. It may not be resold or used for commercial purposes without the explicit

permission of the foundation. It may, however, be copied in whole or in part and

distributed for free via any medium, provided the Foundation is given appropriate credit

as the source of the material. The AAA Foundation for Traffic Safety assumes no liability

for the use or misuse of any information, opinions, findings, conclusions, or

recommendations contained in this report.

If trade or manufacturer’s names are mentioned, it is only because they are considered

essential to the object of this report and their mention should not be construed as an

endorsement. The AAA Foundation for Traffic Safety does not endorse products or

manufacturers.

Table of Contents

List of Abbreviations and Acronyms ........................................................................... vi

Executive Summary ....................................................................................................... 1

Introduction ................................................................................................................... 3

Method ........................................................................................................................... 5

Records of Roadside Assistance Providers who Died ........................................................... 5

Data on All Fatal Crashes in the United States ....................................................................... 6

Results ............................................................................................................................ 9

Overall Trends and Patterns ..................................................................................................... 9

Characteristics of Crashes ........................................................................................................ 11

Characteristics of Striking Vehicles and Drivers ................................................................. 15

Discussion .................................................................................................................... 18

Limitations .................................................................................................................................. 23

Conclusion and Recommendations ............................................................................. 25

References .................................................................................................................... 27

List of Abbreviations and Acronyms

BLS Bureau of Labor Statistics

DOT Department of Transportation

ERSI Emergency Responder Safety Institute

FARS Fatality Analysis Reporting System

FHWA Federal Highway Administration

MMUCC Model Minimum Uniform Crash Criteria

NHTSA National Highway Traffic Safety Administration

OSHA Occupational Safety and Health Administration

RAP Roadside Assistance Provider

Executive Summary

Roadside assistance providers including motor vehicle towing personnel, mobile

mechanics, and safety service patrollers often work on or alongside highways to help

stranded motorists after vehicle breakdowns or crashes. This places them at risk of being

struck by passing vehicles. Research is needed to understand the circumstances of

crashes involving roadside assistance providers, to inform and prioritize efforts to

protect them.

Unfortunately, when roadside assistance providers are injured or killed in

crashes, the sources of data used most in traffic safety research do not reliably report the

fact that the victim was a roadside assistance provider. When they are struck while

working outside of their vehicle, they typically are reported simply as pedestrians,

rendering them indistinguishable from other pedestrians injured or killed in crashes.

This has hindered past efforts to study roadside assistance provider safety.

In the current study, industry records were used to identify roadside assistance

providers struck and killed by vehicles. Those records were matched to a federal

database of all fatal crashes nationwide to determine which crash victims in the

database were roadside assistance providers. These matched data were then used to

examine the characteristics of the crashes in which roadside assistance providers were

killed, with a focus on types of factors relevant to the selection or prioritization of

potential countermeasures to protect roadside assistance providers.

The study identified 123 roadside assistance providers fatally struck while

working in the road or at the roadside in years 2015 (the first year of victim records

available) through 2021 (the latest year of national fatal crash data available). There

appears to have been an increasing trend in the number of roadside assistance providers

killed each year; however, trends should be interpreted with caution as it is possible that

some cases were missed. The vast majority of these deaths occurred on Interstate

highways or other high-speed limited-access highways. Most crashes occurred during

good weather on roads that were not slippery. Nearly two-thirds occurred in darkness;

however, information regarding use of high visibility safety apparel, emergency lighting,

and other safety equipment was unavailable in most cases. Approximately one in four

crashes occurred in the travel lanes of the roadway, highlighting the importance of

increasing compliance with Slow Down, Move Over laws, which require drivers to move

over one lane or reduce speed when approaching roadside assistance providers, police,

firefighters, emergency medical service providers, and in some states also other workers

working at the roadside. However, more than half of all crashes occurred when the

striking driver departed the roadway and struck the roadside assistance provider on the

shoulder or roadside, suggestive albeit not necessarily indicative of the involvement of

factors such as impairment, fatigue, or distraction. More than one-third of striking

drivers who were tested for alcohol were alcohol-positive; however, nearly half were not

tested.

The current study represents the most comprehensive compilation of data on

roadside assistance provider fatalities of which the AAA Foundation for Traffic Safety is

aware. Records of 123 roadside assistance providers fatally struck by vehicles were

identified; slightly more than reported in a federal occupational fatality database and far

more than in national fatal motor vehicle crash data.

Based on the findings of this research, as well as the limitations encountered, this

report makes the following recommendations:

• There is a need to reinforce public awareness of and increase motorist

compliance with Slow Down, Move Over laws; research to investigate the most

effective approaches to increase compliance is also needed.

• Countermeasures are needed to protect roadside assistance providers and first

responders from out-of-control vehicles that depart the roadway by preventing

the original road departure, by protecting those at the roadside from being struck,

or by reducing the severity of any impact that may occur. Research is needed to

determine what countermeasures are most effective and practical to meet this

need.

• Training for roadside assistance providers should emphasize the importance

avoiding working or standing on the traffic-facing side of the incident scene

whenever possible and minimize time spent on the traffic-facing side of the scene

when it cannot be avoided altogether, as well as provide strategies for how to do

so.

• In cases where countermeasures are site-specific or where deployment must be

prioritized, results of the current study suggest that efforts should focus on

protecting roadside assistance providers working on high-speed limited-access

highways.

• State police crash report forms should include data fields designed to report

whether a crash victim was an incident responder, and type of responder when

applicable, as called for in the current edition of the Model Minimum Uniform

Crash Criteria. Additionally, states should consider collecting additional

information in crashes in which roadside assistance providers and other

emergency response personnel are struck by vehicles, or at least in those that

result in the injury or death of a responder.

Introduction

Roadside Assistance Providers (RAPs) including motor vehicle towing personnel,

emergency roadside technicians, mobile mechanics, and safety service patrollers work to

assist motorists whose vehicles have become disabled due to mechanical failures,

crashes, or other causes. Whereas law enforcement, firefighting, and emergency medical

service personnel are classified “first responders,” RAPs are often termed “second

responders.” When RAPs are called to assist a motorist at the roadside, they are often

required to work under high-risk conditions, such as on the shoulder or even in the

travel lane of a high-speed roadway that remains open to traffic, placing them at risk of

being struck by passing vehicles. Using occupational fatality and employment data from

the Bureau of Labor Statistics (BLS) between 2005 and 2015, a previous study by Bunn et

al. (2018) found that the motor vehicle towing industry had more than double the

occupational fatality rate of other transportation incident management industries

including law enforcement and fire protection.

Although RAPs clearly face substantial risk in the line of duty, little research has

examined characteristics of crashes in which RAPs are injured or killed (Chandler &

Bunn, 2019). Studies examining first responders as well as RAPs often include relatively

little data specifically regarding RAPs, likely due at least in part to crash report forms

and data reporting conventions that hinder the identification of crash victims as RAPs

(Horrey et al., 2021). Although the Model Minimum Uniform Crash Criteria (MMUCC)

(National Highway Traffic Safety Administration [NHTSA], 2017, 2024) calls for the

collection of data including whether each crash-involved person was an incident

responder and the type of responder if applicable, NHTSA reported that state police

crash reports collected fewer than 50% of the MMUCC data elements as of 2022 (Agency

Information Collection Activities, 2022). It is evident from the few studies that have

attempted to examine injuries or fatalities of RAPs that the current state of data

collection and reporting has hindered research on the topic.

Yu et al. (2013) examined a sample of 265 crashes in Wisconsin between 2000 and

2010 in which emergency responders were struck by vehicles; however, most involved

law enforcement officers and few were RAPs. (The article did not specifically report the

number of RAPs; however, calculations from reported data indicate it was fewer than

10.) The authors noted that the state’s police crash report form did not provide any data

fields to note that a crash-involved person was a tow-truck driver, likely leading to

undercounting of towing operators struck by vehicles. Carrick & Srinivasan (2023)

examined Florida crashes between 2011 and 2020 in which a driver was cited for a

violation of the state “Slow Down, Move Over” law. In most of the crashes examined, only

the responder’s vehicle, not the responder, was struck. The authors reported that of 519

crashes examined, only 34 involved a responder struck as a pedestrian, the majority of

which were law enforcement officers. Of the 34, only nine were RAPs (seven towing and

two safety service patrol), thus limiting the depth of analysis that could be conducted

regarding the characteristics of crashes in which RAPs were struck. The authors noted

that the state’s crash database lacked detail necessary to identify all emergency roadside

response personnel, and that their method of identifying only crashes in which a driver

was cited for violating the state’s Slow Down, Move Over law would not identify all

crashes in which responders were struck. Chandler & Bunn (2019) examined injuries and

deaths of towing operators reported in an Occupational Safety & Health Administration

(OSHA) database between 2002 and 2017, and identified 106 total injuries or deaths of

towing operators, including 39 in which the towing operator was a pedestrian struck by

a passing vehicle. However, the authors noted that OSHA is not required to investigate all

occupational injuries that occur on public roadways and thus that OSHA’s database likely

did not include all such injuries and deaths.

Bunn et al. (2018) examined traffic incident management occupational fatalities in

Kentucky between 2005 and 2016 identified through the state’s National Institute for

Occupational Safety and Health (NIOSH)-funded Fatality Assessment and Control (FACE)

occupational fatality surveillance program. Of 29 such traffic incident management

fatalities identified, 11 were RAPs (six towing and five transportation agency personnel,

e.g., safety service patrol or similar). The article noted that the state’s crash database only

correctly identified the occupation of four of the 29 fatalities in total, including only

three of the 11 RAPs. Thus, had crash data been the only source of information used, a

large majority of all deaths of traffic incident management personnel, including but not

limited to RAPs, would not have been identified as having involved traffic incident

management personnel or RAPs. Importantly, however, the study noted that the

occupations of all but one of the traffic incident management fatalities were identified

correctly through a different source of data: media reports.

Yang et al. (2023) examined media reports describing over 5,000 incidents in

which emergency responders including RAPs were struck or nearly struck by vehicles

between 2001 and 2020. Although not stated explicitly in the article, the study data

appeared to include approximately 85 RAPs who died and roughly an additional 200 who

were struck or nearly struck. Although the study notes that results could potentially be

biased by differential media reporting in relation to incident severity or responder type,

results suggest that incidents involving RAPs were significantly more likely to be fatal

than incidents involving other types of responders. This study was able to make effective

use of data from media reports to identify crashes involving emergency responders

including RAPs—at least partially overcoming this key shortcoming of state and national

crash databases discussed previously. However, many crash-related details almost

universally available in crash databases (e.g., location relative to the roadway, driver

actions contributing the crash, even the time of day) were not reported in a substantial

proportion of the media articles and thus were not available for analysis.

Given the well-documented limitations of crash databases that hinder the

identification of crash victims as RAPs, as well as the noted success of identifying them

through media reports, the current study sought to use data from both sources to

identify and examine the characteristics of crashes in which RAPs were struck and killed

by vehicles.

Method

This purpose of the current study was to conduct a descriptive analysis of the

characteristics of crashes in which RAPs were struck and killed by vehicles. Because

RAPs are not consistently identifiable in crash databases, RAPs struck and killed by

vehicles were first identified using records compiled by two organizations that

memorialize RAPs who died. These records were then matched to a database of all fatal

crashes in the U.S. to obtain additional information about the crashes. These matched

records were then analyzed to describe the characteristics of the crashes in which RAPs

were killed in years 2015–2021.

Records of Roadside Assistance Providers Who Died

Records of roadside assistance providers who died were obtained from two

sources: The National Tow List’s “Fallen Operators” (2015–2021) and the Emergency

Responder Safety Institute’s struck-by-vehicle fatality incidents data (2019–2021).

The National Tow List maintains a free list of towing providers globally on their

website and maintains an up-to-date list of all towing related news on their publicly

available Facebook page

. Each year starting in 2016, the National Tow List has compiled

a list of “Fallen Operators” in their Facebook photo albums, which details the name, age,

date, location (city and state), and circumstances of death for tow operators who died in

the preceding year. The list of Fallen Operators from years 2015–2021 included a total

334 deaths of tow truck drivers, including deaths unrelated to being struck by a vehicle

(e.g., medical events, incidents involving equipment). The authors manually reviewed

the descriptions of the circumstances of each death. Based on this review, 122 deaths for

which circumstances were described as “struck” or similar were included in the current

study.

The Emergency Responder Safety Institute (ERSI) maintains a list of struck-by-

vehicle fatalities of all types of emergency responders on its website

(ERSI, 2023). The list

is compiled from daily searches of internet sources, news reports, social media pages,

and through contacts in the police, fire, EMS, and towing industries. The list details the

name, date of the incident and death, location (state and county), type of emergency

responder, on/off-duty status, organization/employer, and activity of struck-by fatalities

along with a hyperlink to a news story or source. The first full year for which ERSI’s

struck-by-vehicle fatality incident data is available was 2019. ERSI receives funding and

support from the U.S. Fire Administration, the Department of Homeland Security, the

Department of Justice’s Office of Justice Programs, and the Federal Highway

Administration.

Data obtained from ERSI included 155 records of responder deaths, including law

enforcement, EMS/firefighters, tow truck drivers, road service technicians, mobile

mechanics, and safety service patrollers. From these records, 68 were identified as RAPs

(i.e., towing, road service technicians, mobile mechanics, and safety service patrollers),

and were included in the current study.

Records from the National Tow List and ERSI were combined into a single list

(hereafter referred to as RAP records). This included 59 deaths reported by the National

Tow List in 2015–2018, 63 deaths reported by both organizations in 2019–2021, and five

additional deaths reported only by ERSI in 2019–2021. After removing duplicates, the

final combined list of RAP records included records of 127 roadside assistance providers

struck and killed by vehicles in 2015–2021.

Data on All Fatal Crashes in the United States

Data on all fatal crashes in the U.S. were obtained from the NHTSA’s Fatality

Analysis Reporting System (FARS) (NHTSA, 2023a). FARS includes a record of every crash

that involves a motor vehicle in-transport, occurs on a public roadway, and results in the

death of at least one involved person within 30 days of the crash. Data in FARS are

compiled from police crash reports and include detailed information about all vehicles

and people involved in fatal crashes, as well as information about the characteristics of

the road, environmental conditions at the time of the crash, and the sequence of events

that led up to the crash.

Matching RAP Records to Fatal Crash Data

RAP records were matched to corresponding records in FARS through a multi-

stage process. Data available in both RAP records and in FARS included the victims’ age,

date of death, and the state in which the crash occurred. (The National Tow List also

included the city of the crash; ERSI also included the crash date and the county.)

In the first round of matching, the date of death, victim’s age, and state from RAP

records uniquely identified 95 records in FARS. An additional seven RAP records each

matched two FARS records on the basis of these variables.

For records with multiple possible matches, the authors sought media reports

about the crashes to obtain additional information that could potentially help to

determine which FARS record was the correct match. For RAP records obtained from

ERSI, the links provided by ERSI were used. For RAP records obtained from the National

Tow List, the authors performed Google searches for relevant media articles using search

terms including the victim’s name, state, date of crash or death, and search terms crash

OR accident OR hit OR struck OR killed OR died OR fatal). When articles were located, the

authors manually abstracted information such as the approximate time of the crash, the

name of the road on which the crash occurred, and sometimes additional location

information. These additional data were then used to determine which of the FARS

records, if any, was the correct match for each RAP record. Using this procedure, the

seven RAP records each initially matched to two FARS records were uniquely matched to

a single FARS record, yielding 102 matched records, after which 25 remained

unmatched.

For the 25 remaining unmatched records, online news reports were searched as

described above, both to obtain additional information as described above and to

confirm or correct information in the original RAP records. Several corrections were

made to victim’s ages (in most cases by ±1 year) and crash and death dates (usually by ±1

day). Such corrections to ages in dates in the RAP records resulted in three additional

matches to FARS, leaving 22 RAP records still unmatched.

For the remaining unmatched records, matching criteria were relaxed to allow

discrepancies of ±1 day in the crash date and date of death, age was not required to

match, and new potential matches were examined manually to check for agreement with

respect to the crash city or county, time, and the road on which the crash occurred and

determine which, if any, was the correct match. This process yielded 17 additional

matches including several with age discrepancies of 1 year, one with an age discrepancy

of 2 years, one with an age discrepancy of 4 years, and six with discrepancies of 1 day in

the date of death. In one additional case, RAP records listed a fatally injured towing

operator as 61 years of age but the matched FARS record (matched on state, county, road,

date, time, and number and type of vehicles in the crash) listed the victim’s age as 34.

The authors identified the same crash in the relevant state crash database, and in it the

victim’s age was listed as 61. The authors thus concluded that this was the correct match,

and that the victim’s age was entered incorrectly into FARS.

Of five cases remaining unmatched, no online news article could be found

pertaining to two of the crashes/victims. Another was determined through online news

articles to be outside the scope of state and national traffic fatality data because the

victim died several months after the crash. In one case, it was unclear whether the

towing operator had died or not, as multiple sources of data provided conflicting

information. For this crash, an article on a crowdsourced website described a crash in

which a 35-year-old male driver of a sedan struck a stopped tow truck, killing the 48-

year-old male towing operator. The FARS record of what was clearly the same crash

(matched on date, time, state, city, road, makes and models of vehicles involved, and

approximate ages of their drivers) indicated that the 35-year-old male driver of the

striking vehicle died while the 47-year-old male driver of the struck tow truck survived.

Given this discrepancy, the authors searched a death notice for the towing operator and

found none. However, a search for the striking driver yielded a death notice for a person

of the same name (per media report), age, and state who died on the date of the crash.

Given these conflicting data, this crash was not included in the current study.

The last unmatched case was confirmed through multiple mainstream television

and print media reports to have been a tow operator who was struck and killed by a

vehicle on a public road thus was clearly within the scope of FARS; however, FARS

contained no record of any fatal crash in the corresponding city within several weeks of

the crash date indicated in the RAP record and media reports. The authors uniquely

identified the relevant crash in the state Department of Transportation’s crash database

based on the crash date, time, county, city, and road. However, the crash was coded in

the state’s database as a non-fatal crash. In attempt to reconcile RAP record and media

reports with the state’s coding of the crash as non-fatal and its absence from FARS, the

authors requested and obtained a redacted copy of the police crash report for the

relevant crash. In the police report, the police officer’s narrative clearly stated, “Unit #1

did not see the tow truck…and struck the tow truck driver who was standing next to Unit

#2 killing him.” However, the tow truck driver was erroneously omitted from the portion

of the police report that lists the injury severity of persons who were injured or killed,

and a checkbox at the top of the form intended to be marked for fatal crashes was not

marked, collectively explaining the miscoding of the crash in the database as non-fatal

and consequent omission from the FARS database. This case was included in the study.

The authors used data abstracted from the police report to populate relevant variables in

the study database. Ultimately, 123 RAP records were matched to their corresponding

records in crash databases.

Identification of Roadside Assistance Providers Using FARS Alone

Before 2019, the FARS database did not specifically identify motor vehicle towing

personnel, safety service patrollers, or other RAPs when they were struck while outside

of their vehicle (NHTSA, 2023b). However, it included a code for emergency services

personnel including but are not limited to RAPs. Of the 123 RAPs identified in the current

study through RAP records, 55 were killed in 2015–2018, and only 9 of them were

identified in FARS as emergency services personnel.

The NHTSA added new codes to the FARS database in 2019 to facilitate the

identification of crash victims as towing operators or other transportation workers

(including but not limited to RAPs, e.g., state highway safety service patrollers). In years

2019–2021, 26 crash victims were coded as towing operators and 16 as transportation

workers. By comparison to RAP records, 22 of the 26 persons coded in FARS as towing

operators and 3 of 16 persons coded in FARS as transportation workers were confirmed

to have been RAPs. Media reports regarding the other four crash victims identified in

FARS as towing operators indicated that two were not struck by vehicles; no reports

were identified regarding the remaining two. Of the 13 transportation workers identified

in FARS but not in the RAP records, 12 were determined to have not been RAPs (most

were road construction/maintenance workers) and one could not be determined.

In sum, FARS only identified as RAPs 34 of the 123 crash victims identified

through RAP records, including those who were identified in FARS as emergency services

workers as well as the those identified more specifically as towing operators or

transportation workers. Moreover, the new variables added to FARS in 2019 identified

only 25 of the 68 RAPs identified through RAP records in years 2019–2021 and between

zero and three additional RAPs not identified through the RAP records. Thus, it was

determined that FARS alone was insufficient to identify RAPs struck and killed by

vehicles, even with the new variables added in 2019.

Analysis

The 123 RAP records and matched FARS records were analyzed to describe the

characteristics of the crashes in which RAPs were fatally struck by vehicles, with a focus

on factors potentially relevant to the development or prioritization of countermeasures

to protect RAPs.

Results

Overall Trends and Patterns

Figure 1 shows the distribution of the 123 RAP deaths identified by year. The data

are indicative of a statistically significant increasing trend in the number of RAPs

identified who were fatally struck by vehicles each year (P=0.001 for linear trend). While

the total number of people killed in all crashes each year also increased over the study

period, fatalities identified as RAPs also increased significantly as a proportion of total

traffic fatalities (P=0.0064 for linear trend in proportion), indicating that the increasing

trend in RAPs fatally struck by vehicles was even greater than the overall increasing

trend in all traffic fatalities. Note that trends should be interpreted with caution; if RAP

records were less complete in earlier years than in later years, this could result in the

magnitude of the trend being overestimated.

Figure 2 shows the number of RAP fatalities identified in each state over the study

period. Unsurprisingly the three most populous states (California, Texas, and Florida)

accounted for the largest numbers of RAP fatalities. Differences between states should be

interpreted with caution, however, as the methodology used to identify cases could have

resulted in the unintentional omission of some RAP fatalities, and even a small number

of additional cases could impact the distribution of fatalities by state significantly.

Figure 1. Annual Number of Roadside Assistance Providers

Struck and Killed by Vehicles, United States, 2015–2021.

Figure 2. Number of Roadside Assistance Providers Struck and Killed by Vehicles,

by State, 2015–2021.

2015 2016 2017 2018 2019 2020 2021

Figure 3 shows the ages of RAPs fatally struck by vehicles. Their average age was 41

years, the youngest was 21, and the oldest was 80. All of them were men.

Figure 3. Ages of Roadside Assistance Providers Struck and

Killed by Vehicles, United States, 2015–2021.

Characteristics of Crashes

Roadway and Environment

Figure 4 shows the types of roads on which RAPs were struck and killed by

vehicles. In total, 85% of all deaths of RAPs struck by vehicles occurred on high-speed

limited-access highways. More than two-thirds of all cases occurred on Interstate

highways; the majority of the remainder occurred on other limited-access highways.

Nearly 90% of all RAP fatalities occurred at locations with speed limits of 55 mph or

higher; only six occurred at locations with speed limits below 45 mph.

Figure 4. Functional Classification and Speed Limit of Roads on which Roadside Assistance

Providers were Struck and Killed by Vehicles, United States, 2015–2021.

30 30

21-30 31-40 41-50 51-60 61-70 71-80

Age of Roadside Assistance Provider

Interstate

84 (68%)

Other

Freeway

21 (17%)

Arterial

15 (12%)

Collector

3 (2%)

≤ 40

45-50 55-60 65+

Speed Limit (miles per hour)

Figure 5 shows the time of day and lighting conditions present when RAPs were

struck and killed by vehicles. The majority occurred during nighttime hours. Nearly two-

thirds occurred during darkness, most of which occurred at unlit locations. Nearly one-

third of all RAP fatalities occurred between the hours of 9 p.m. and 1 a.m.

Figure 5. Time of Day and Lighting Conditions when Roadside Assistance Providers were

Struck and Killed by Vehicles, United States, 2015 – 2021.

Figure 6 shows the weather and roadway surface conditions present at the

locations where RAPs were fatally struck. The road was reported as dry and weather was

reported as clear or cloudy without precipitation in 103 of the 123 crashes examined

(84%). Precipitation was only noted in 16 crashes (13%). Similarly, the road was only

noted as wet in 13 crashes (11%) and icy in seven (6%).

Figure 6. Road Surface and Weather Conditions Present when

Roadside Assistance Providers were Struck and Killed by Vehicles,

United States, 2015–2021.

Daylight

42 (34%)

Dawn/Dusk

4 (3%)

Dark-Lit

27 (22%)

Dark-Unlit

50 (41%)

5-9a 9a-1p 1-5p 5-9p 9p-1a 1-5a

103

Dry road,

no precip.

Wet road,

no precip

Wet road,

raining

Icy road,

no precip.

Icy road,

snow/sleet/

freezing rain

Pre-Crash Circumstances

Figure 7 shows the locations of RAPs when they were struck. All but four of them

were outside of their vehicles. Sixty-three percent (77 of 123) were standing or working

outside of the travel lanes on the shoulder or roadside. Approximately one-third (39 of

123) were working or standing in the travel lanes. Three were known to be outside of

their vehicles, but it was unclear whether they were located in the travel lanes or on the

shoulder/roadside. Of the four RAPs who were killed while inside of their roadside

assistance vehicle, three were parked on the shoulder or roadside and one was reported

in FARS as in-transport on the roadway.

Figure 7. Location of Roadside Assistance Providers Struck and Killed by Vehicles,

United States, 2015–2021.

Table 1 shows the general sequence of events in crashes in which RAPs were

struck and killed by vehicles, up to the point at which the RAP was struck. Note that

crashes that shared similar major events were grouped together, the details provided are

intended to be representative examples of a general type of crash; they do not

necessarily fully capture all of the details of every crash in the same group. Also note that

events that occurred after the RAP was struck were not examined, as the focus was on

understanding events that occurred leading up to the RAP being struck.

Outside vehicle,

on shoulder/roadside/etc.

Outside vehicle,

in travel lanes

Outside vehicle,

location unknown

Inside vehicle,

on shoulder/roadside/etc.

Inside vehicle,

in travel lane

Table 1. General Sequences of Events in Crashes in which RAPs were Struck and Killed,

United States, 2015–2021.

Crash Type

General Description

Specific Details

Number (%)

A-1

Vehicle in transport

strikes RAP directly

RAP struck in travel lanes of

roadway

24 (20%)

A-2

Vehicle departs road first, then

strikes RAP outside of travel lanes

36 (29%)

B-1

Vehicle in transport

strikes RAP vehicle

and/or disabled

vehicle, then one of

those vehicles

strikes RAP

RAP vehicle or disabled vehicle

struck in travel lanes of roadway

4 (3%)

B-2

Striking vehicle departs road first,

initial impact occurs outside of

travel lanes

39 (32%)

B-3

Striking vehicle departs road and

strikes RAP vehicle (RAP inside of

vehicle)

3 (2%)

Vehicles collided on the road first, then one of those

vehicles strikes the RAP

16 (13%)

RAP vehicle struck while re-entering traffic with

disabled vehicle in tow

1 (1%)

In the most common general type of crash, labelled A in Table 1, a vehicle in

transport (i.e., a vehicle being driven on the roadway) struck the RAP directly. This

general type of crash accounted for 60 of the 123 RAP fatalities examined (49%). In crash

type A-1, which accounted for 40% of type A and 20% of all RAP fatalities, the RAP was

standing or working in the travel lanes of the roadway when the vehicle struck him. In

type A-2 (60% of crash type A and 29% of all RAP fatalities), the vehicle first traveled out

of the travel lanes and onto the shoulder or roadside before striking the RAP. As noted

previously, these generalized descriptions are intended only to capture the key details

leading up to the RAP being struck and do not fully capture every detail of every crash.

For example, in some of these crashes, the striking vehicle may have struck the RAP

vehicle, the disabled vehicle, a guardrail, etc., after striking the RAP; however, data

indicate that the striking vehicle struck the RAP directly, before any other impacts in the

crash. Finally, of particular note, in 25 of these crashes, only the RAP worker and neither

the RAP vehicle nor disabled vehicle was struck.

In crash type B, a vehicle in transport first struck the RAP vehicle or a disabled

vehicle, leading to a chain of events resulting in the RAP being struck and killed. This

general crash type accounted for 43 of the 123 RAP fatalities examined (37%). Most

commonly, in 39 of the cases examined, a vehicle in transport first departed the

roadway, and then struck the RAP vehicle or disabled vehicle outside of the travel lanes

of the roadway, after which one of these vehicles (the vehicle in transport, the RAP

vehicle, or the disabled vehicle) fatally struck the RAP (crash type B-2 in Table 1). These

include cases in which the RAP was standing on the shoulder or roadside when struck as

well as cases when the RAP was standing on the RAP vehicle (e.g., standing on the flatbed

actively securing the disabled vehicle to the flatbed) and cases where the RAP was

standing between the RAP vehicle and disabled vehicle and became pinned when the

disabled vehicle was struck. In multiple cases, media reports or other sources described

the vehicle in transport driving up the ramp of the flatbed, becoming airborne, and

striking the RAP upon landing. Also included within crash type B were four cases that

occurred within the travel lanes of the road (i.e., where the striking vehicle did not

depart the road prior to the crash) and three in which the RAP was seated inside the cab

of the parked RAP vehicle and was killed when the vehicle was struck.

In a less common yet still significant scenario, 16 cases (13% of all RAP fatalities)

followed an initial collision between two or more vehicles in transport on the road (type

C in Table 1). In most of these cases, one of the vehicles involved in the initial collision

then struck the RAP directly, though a few involved more complex chains of events (in

two cases, one of the original vehicles then struck the RAP vehicle, which then struck the

RAP, and one case involved multiple impacts between multiple vehicles in transport on

the roadway before one of those vehicles subsequently struck the RAP).

Finally, in one case, a vehicle in transport struck the RAP vehicle while the RAP

vehicle was attempting to re-enter traffic with the disabled vehicle in tow.

Characteristics of Striking Vehicles and Drivers

The characteristics of the vehicles and drivers who struck RAPs were examined to

gain additional insight into the safety risks experienced by RAPs. Note that in cases

where the RAP was struck in a complex sequence of events that involved multiple

vehicles, this analysis focused on the vehicle and driver that crash data indicate initiated

the sequence of events that led to the RAP being struck, even if it was not necessarily the

same vehicle that physically struck the RAP. Also note that 26 of the drivers who fatally

struck RAPs (21%) left the scene of the crash; some still had not yet been identified before

the FARS data were finalized, resulting in some missing information about striking

drivers and vehicles.

Figure 8 shows the types of vehicles that struck and killed RAPs. Slightly more

than one-third were pickup trucks, vans, or SUVs and slightly fewer than one-third were

passenger cars. Nearly one in four was a large truck. Vehicle type was unknown for eight

striking vehicles.

Figure 8. Types of Vehicles that Struck and Killed

Roadside Assistance Providers, United States, 2015–2021.

Figure 9 shows the ages of the drivers of the vehicles that struck and killed RAPs.

Age was unknown for nine drivers who left the scene and were not subsequently

identified. Excluding those of unknown age, the median age of these drivers was 38

years, the youngest was 17 and the oldest was 80. Excluding nine for whom sex was not

determined, 82 (72%) were men and 32 (28%) were women.

Figure 9. Ages of Drivers who Struck and Killed Roadside

Assistance Providers, United States, 2015–2021.

Figure 10 shows data regarding alcohol involvement among drivers who struck

RAPs. Of drivers who were tested for alcohol and had confirmed results (58 of 123), 20

(16% of all drivers; 34% of those with alcohol data) tested positive for alcohol, and 16

(13% of all; 28% of those with alcohol data) had a blood alcohol concentration higher

than the legal limit, including nine whose blood alcohol concentration was more than

double the legal limit. Notably, however, 48 of 123 drivers (39%) who struck and killed

RAPs were not tested for alcohol, and no information about alcohol testing and/or results

were available for another 17 (14%).

Car

39 (32%)

Pickup/

Van/SUV

46 (37%)

Large

Truck

30 (24%)

Unknown

8 (7%)

16-20 21-30 31-40 41-50 51-60 61-70 71-80 Unknown

Age of Driver of Striking Vehicle

Figure 10. Blood Alcohol Concentration of Drivers who Struck and

Killed Roadside Assistance Providers, United States, 2015–2021.

Figure 11 shows the driver license status of drivers who struck and killed RAPs.

Among those whose license status could be determined (113 of 123), one in eight lacked a

valid driver’s license. Three were completely unlicensed, nine were driving with a

suspended or revoked license, and two had an expired driver’s license. Note, however,

previous research has shown that drivers who leave the scene of crashes are

disproportionately likely to lack a valid license (Benson et al., 2021), thus it is likely that

at least some of the 10 whose license status was not determined may have lacked a valid

license. Among those who were driving large trucks requiring a commercial driver’s

license, all held the appropriate class of license for their vehicle.

Figure 11. Driver’s License Status of Drivers who Struck and

Killed Roadside Assistance Providers, United States, 2015–2021.

0.00

0.01–0.07 0.08–0.15 ≥0.16

Not Tested Unknown

Blood Alcohol Concentration of Striking Driver

Valid

99 (80%)

Suspended/Revoked: 9 (7%)

Expired: 2 (2%)

Unlicensed: 3 (2%

)

Unknown: 10 (8%)

Figure 12 shows selected items on the driving records of the striking drivers in the

5 years preceding the crash for 110 of the 123 striking drivers; driving history

information was unavailable for 13. The data indicate that 26 of them (24% of those with

available data) had one or more convictions for speeding violations in the past 5 years,

and 24 (22%) had one or more convictions for other moving violations. Twenty-one

striking drivers (19% of those with driving history data) had had their license suspended

at least once in the past 5 years, 18 (16%) had been involved in at least one crash, and

eight (7%) had at least one previous driving under the influence (DUI) conviction. Note

that the previous statistics are not mutually exclusive and should not be added. In total,

more than half of all striking drivers with available driving history data (57 of 110) had

at least one conviction for speeding, DUI, another moving violation, a license suspension,

or a crash on their recent driving records.

Figure 12. Driving History in Previous 5 Years, Drivers who Struck

and Killed Roadside Assistance Providers, United States, 2015–2021.

Note: Speeding, other moving violation, and DUI refer to convictions only;

violations not resulting in convictions are not included. Data are based on

110 drivers with driving history data available.

Discussion

Although previous research had examined particular subsets of crashes involving

traffic incident management or roadside assistance personnel (Carrick & Srinivasan,

2023; Chandler & Bunn, 2019; Yu et al., 2013), the current study is, to the best of the

authors’ knowledge, the first study that has sought to identify and characterize all

crashes in which RAPs, including motor vehicle towing personnel, safety service

patrollers, emergency roadside technicians and mobile mechanics, were fatally struck by

passing vehicles while working at the roadside.

Speeding

Other Moving

Violation

License

Suspension

Crash

DUI

1 or More

of the Above

Previous research noted that RAPs were not readily identifiable in state crash

databases (Bunn et al., 2018) due to limitations in police crash report forms (Carrick &

Srinivasan, 2023; Yu et al., 2013), and are not comprehensively covered in an OSHA

database of occupational fatalities because OSHA is not required to investigate

occupational fatalities on public roads other than in work zones (Chandler & Bunn,

2019). However, a study of NIOSH-funded state-level occupational fatality surveillance

concluded that traffic incident management fatalities including those of RAPs were

reliably identified by media reports (Bunn et al., 2018). While media reports are useful to

identify incidents involving RAPs, they lack much of the contextual information typically

available in crash databases (Yang et al., 2023). Thus, the current study identified RAPs

fatally struck by vehicles using lists previously compiled by two industry organizations,

largely based on media reports in conjunction with other sources, to memorialize towing

operators and emergency responders who died. These records were then linked to

NHTSA’s FARS database—a database of all fatal crashes in the U.S.—to obtain additional

information to enable the examination of crash characteristics.

Results indicate that at least 123 RAPs were struck and killed by vehicles on U.S.

roads in 2015–2021. The true number could be higher as it is possible some cases were

missed. The data suggest that there was an increasing trend over the study period in the

annual number of RAPs struck and killed by vehicles. Moreover, while there was an

overall national increase in traffic fatalities over the study period, the increase in

fatalities of RAPs struck by vehicles was shown to be significantly greater. The trend

should be interpreted with caution given the possibility that some cases could have been

missed; however, unless a very large number of cases were missed disproportionately

among the earlier years of the study, it does appear that the annual number of RAPs

stuck and killed by vehicles is increasing.

Results include many similarities but also differences relative to past research and

provide several insights that may help to inform countermeasures to protect RAPs. In a

study of Florida crashes in which a driver was cited for a Move-Over Law violation, none

resulted in the death of a responder (Carrick & Srinivasan, 2023). In a Wisconsin study of

road worker and incident responder struck-by crashes, <10% resulted in a worker or

responder’s death (Yu et al., 2013). Not unexpectedly, crashes examined in the current

study—all of which resulted in the death of a RAP—differed from those in many respects.

In the current study, 68% of RAP struck-by deaths occurred on Interstate

highways and 18% occurred on other limited-access highways, and nearly 90% occurred

at locations with posted speed limits of 55 mph or higher, all far higher than in previous

studies. In another study of crashes involving disabled vehicles not focused specifically

on RAPs, Spicer et al. (2021) found that fatalities of pedestrians attending to disabled

vehicles increased by 27% from 2014 to 2018. While that research did not distinguish

between RAPs and other motorists (e.g., the drivers of the disabled vehicles), it highlights

that substantial proportions of such fatalities occur late at night and on Interstates,

consistent with findings from the current study. While it is important for motorists and

responders to take precautions at all times, when considering ways to prevent future

struck-by fatalities of RAPs, it is important to recognize that the overwhelming majority

of RAP struck-by fatalities occur on high-speed limited access highways. While

Interstates and other limited access highways often provide wide shoulders and long

sight distances to improve safety, working outside of a vehicle on such facilities is clearly

dangerous to RAPs, both because drivers do not ordinarily expect to encounter

pedestrians on freeways or Interstates, and because high traffic speeds increase the risk

that a crash will be fatal (Tefft, 2013). These results indicate that, in the context of any

potential countermeasures that are site specific or that can only be deployed on a limited

basis, the need to protect RAPs from being struck and killed by passing vehicles is clearly

greatest on high-speed limited-access highways.

In contrast to previous studies of mainly non-fatal crashes (Carrick & Srinivasan,

2023; Yu et al., 2013), a substantial majority of crashes examined (63%) occurred in

darkness, of which nearly twice as many occurred at unlit locations as at locations with

lighting. While FARS does not include details regarding responders’ use of emergency

lighting, flares, high-visibility safety apparel, or other countermeasures to increase their

conspicuity to passing motorists, the high proportion of RAP struck-by fatalities

occurring in darkness highlights the potential importance of countermeasures to

increase the visibility and conspicuity of the scene and workers.

In a study of occupational safety data from OSHA, Chandler & Bunn (2019) noted

that a substantial proportion of cases in which towing operators were struck occurred

while the operator was actively loading or securing the disabled vehicle onto the tow

truck—a scenario in which the authors noted that the raised flatbed or the disabled

vehicle could potentially block approaching drivers’ view of the vehicle’s emergency

lighting. It would be useful to investigate the prevalence of this scenario in a larger

sample of crashes such as those examined in the current study; however, such details are

beyond the scope of FARS. Multiple media reports of crashes examined in the current

study described the striking vehicle driving up the raised ramp of the flatbed and then

either landing on the tow operator or pushing the disabled vehicle (already on the

flatbed) into him; however, most crashes were not described in sufficient detail to

determine how many of the crashes examined involved this or similar scenarios. This

suggests, albeit not conclusively, that there may be a particular need to ensure the

visibility and conspicuity of tow trucks and towing personnel while a disabled vehicle is

actively being loaded onto the tow truck.

Carrick & Srinivasan (2023) noted that many responders struck by vehicles were

positioned on the traffic-facing side of the incident scene when struck. Chandler & Bunn

(2019) also identified entering/exiting the vehicle and working on the traffic-facing side

of the scene as prominent factors in injuries and fatalities of towing operators reported

in an OSHA database. Although FARS data only indicate the responders’ location as in the

travel lane versus on the shoulder and not specific position relative to vehicles, data

suggest that many of the fatally injured RAPs were likely positioned on the traffic-facing

side of the scene as well. In nearly half of the crashes examined, the first harmful event

of the crash was a vehicle in transport striking the RAP (i.e., before having struck the

RAP’s vehicle or disabled vehicle). In fact, in approximately one-third crashes examined,

only the RAP and not the RAP’s vehicle nor the disabled vehicle was struck. It is possible

that some of these RAPs were positioned elsewhere. For example, the narrative

description of one case from the current study that was also identified in an OHSA

database of workplace fatality inspections stated that the RAP was walking behind his

vehicle removing cones from the road when he was struck. However, the data would

appear to suggest that a substantial proportion of RAPs who were struck and killed by

vehicles were likely positioned on the traffic-facing side of the scene. Another case

identified in the above-mentioned OSHA database stated that the RAP was attempting to

change a tire on the traffic-facing side of a semi-trailer when he was struck. Current

safety training practice for towing operators emphasizes the importance of minimizing

exposure to traffic and suggests strategies such as entering and exiting the vehicle on the

passenger side and using passenger side controls wherever possible. However, in some

circumstances working at least briefly on the traffic-facing side of the scene may be

inevitable. Nonetheless, collectively, these findings suggest that reinforcing RAP training

to avoid working on traffic-facing side of the scene whenever possible may be

particularly important to protecting RAPs from being fatally struck by vehicles.

The current study identified 29 fatal crashes in which either the RAP or their

vehicle was struck directly in the travel lanes of the road, killing the RAP. Slow Down,

Move Over laws are a key countermeasure to protect RAPs as well as other first

responders. These laws generally require drivers to move over one lane whenever

possible, and to reduce speed if moving over is impossible, when passing emergency

response personnel, RAPs, and in some states also certain other types of vehicles stopped

on the side of the road (AAA, 2023). Compliance with Slow Down, Move Over laws—which

require drivers to move over one lane or reduce speed when approaching roadside

assistance providers, police, firefighters, emergency medical service providers, and in

some states also other workers working at the roadside—should prevent such crashes

and deaths. Previous AAA Foundation research has found that use of vehicle-mounted

variable message signs (Liu et al., 2023) and cones and flares (Blomberg et al., 2023)

increase the likelihood of drivers moving over, though both studies reported that

compliance was not universal even when these countermeasures were used.

Another important finding of this research, however, is that 60% of incidents in

which RAPs were struck and killed by vehicles were preceded by the striking vehicle

driving off the road onto the shoulder or roadside. In a study of crashes in which RAPs,

as well as first responders, were struck by vehicles, Yang et al. (2023) reported that

crashes in which responders were killed were significantly more likely to involve out-of-

control vehicles, compared with crashes in which the responder survived. While not

conclusive, these observations suggest that many drivers who drove off the road before

striking a RAP might have been impaired, fatigued, or distracted. In the current study,

approximately one-third of the striking drivers who were tested for alcohol tested

positive, although nearly 40% were not tested. While not investigated in the current

study, the roles of fatigue and distraction are known to be substantially underreported in

crash data (Tefft, 2012). Although none of these features are unique to RAP fatalities,

RAPs are particularly vulnerable to such crashes given their exposure to traffic while

working on the side of high-speed highways.

Protecting RAPs from apparently out-of-control vehicles leaving the road and

striking them on the roadside presents a difficult challenge. If a substantial proportion of

these drivers are indeed impaired, fatigued, distracted, or otherwise not actively

controlling their vehicles, awareness of Slow Down, Move Over laws and motivation to

comply may not be sufficient to prevent these crashes. Moreover, in addition to the 60%

of RAP fatalities preceded by another motorists departing the road, another 13% were

preceded by a collision between multiple vehicles in transport, after which one of the

vehicles involved in the original crash then struck the RAP. While preventing vehicles

from departing the road or colliding on the road in the first place would be optimal, most

countermeasures to prevent road departure are vehicle-based (e.g., advanced driver

assistance systems such as lane departure warning or automatic emergency braking) or

infrastructure-based (e.g., rumblestrips to alert drivers that road departure is imminent

and given them an opportunity to redirect their vehicle). Beyond general

countermeasures intended to prevent road departure, there are also countermeasures

measures to provide advance warning to approaching motorists through in-vehicle GPS

or mobile apps, giving them more opportunity to respond by changing lanes or slowing

down. Whether such measures would effectively reduce the risk of other motorists

departing the road or colliding near the original incident scene (e.g., by helping the

approaching motorists to regain situational awareness sooner) is unclear, likely depends

on the causes of the original road departure or collision, and would also be dependent

upon the approaching drivers having and using relevant technology. While such

countermeasures should help to protect RAPs, they are outside of the immediate control

of the RAPs working on the side of the road. Blomberg et al. (2023) provides a review of

countermeasures that RAPs can deploy in the field and notes that many require

additional evaluation and some pose questions of practicality. Thus, results of the

current study suggest that countermeasures deployed by RAPs in the field to redirect out-

of-control vehicles, prevent them from striking a RAP, or mitigate the severity of any

impact, may be especially important for preventing RAP fatalities. More research is

needed, however, to identify countermeasures that are both effective and practical to

deploy in the conditions in which RAPs work.

Results also highlight shortcomings in current crash data collection and reporting

practices and needs for improvement with respect to data related to RAPs. Only 34 of the

123 RAPs identified in the current study were identifiable in FARS as RAPs or emergency

services personnel. Before 2019, NHTSA’s FARS database lacked sufficient detail to

identify crash victims as RAPs, and identified only 9 of the 55 in the current study as

emergency services personnel. In 2019, NHTSA added new data elements to FARS to

identify crash victims who were specific types of incident responders including tow

operators and other transportation workers (NHTSA, 2023b). However, even after this

addition, FARS identified only 25 of the 68 fatally struck RAPs who were identified

through RAP records in the current study, and at most three additional RAPs not

identified through those RAP records. Thus, although FARS now includes these new data

elements, it will not be possible to rely solely on FARS to identify fatally injured RAPs

unless and until individual states update their police crash report forms to record this

information, as individual police crash reports are the main source of information used

by NHTSA to enter information into FARS.

Moreover, while the methodology used for the current study was premised on the

belief that all RAPs struck and killed by vehicles on public roads would be present in

FARS (but not necessarily identifiable as RAPs), the authors identified one RAP fatality

that was absent from FARS altogether. Further investigation revealed that its absence

from FARS was due to the corresponding state police crash report erroneously not being

marked as fatal, even though the officer’s narrative clearly stated that a tow truck driver

was killed. The incidence of fatal crashes missing from FARS due to errors in the original

crash reports is unknown.

Limitations

This research is subject to several limitations that should be noted. No existing

source of research data reliably identifies all RAPs fatally struck by vehicles (Horrey et

al., 2021). The current study relied on records compiled by two industry groups that

memorialize RAPs who died. It is possible that some fatally injured RAPs were

overlooked, as those organizations’ records were compiled from news reports, social

media, and word of mouth. The BLS reports that 90 persons employed in the motor

vehicle towing industry were killed in “pedestrian–vehicular incidents” during the 2015

to 2021 study years excluding 2019 (corresponding BLS data were unavailable for 2019)

(Bureau of Labor Statistics Data, n.d.); the current study found records of 106 fatally

injured RAPs, including 98 towing operators and eight other RAPs during the same years.

Thus, even if the current study failed to identify some RAPs who should have been

included, it appears that the current study nonetheless examined a more comprehensive

set of RAP fatalities than had ever previously been assembled. Until all states update

their police crash report forms to enable the identification of crash victims as towing

operators or other roadside service providers as provided in the MMUCC (National

Highway Traffic Safety Administration, 2017), there remains a need for research using

other sources of data both to produce accurate statistics regarding RAP fatalities and

injuries and to examine crash circumstances in greater depth to inform and prioritize

efforts to protect them.

In addition, while FARS contains extensive data about the circumstances of

crashes, it was not designed specifically to collect information at the level of detail that

would be desirable to understand the circumstances of crashes in which RAPs were

struck by vehicles. For example, while this research identifies that 63% of fatalities of

RAPs occurred in darkness, it provides little information about use of vehicle lighting,

high visibility safety apparel, or temporary traffic control devices such as cones and

flares, all of which are important to the safety of RAPs working at the roadside.

Additionally, while results suggest that nearly half of fatally injured RAPs may have been

on the traffic-facing side of the incident scene when struck—a finding with important

implications for RAP training—the data lack the level of detail needed to confirm this

conclusively. As noted previously, research has suggested that risk of being struck may

be particularly high while actively loading the disabled vehicle onto the raised flatbed,

possibly due to the RAP vehicle’s emergency lighting being obscured (Chandler & Bunn,

2019); however, the data available for the current study lacked sufficient detail to

determine the prevalence of this scenario. If pertinent data were routinely collected in

investigations of crashes in which RAPs (and other emergency response personnel) were

struck by vehicles, or at least in those resulting in severe injuries or fatalities, it would

provide valuable data for researchers as well as practitioners that would help to inform

and prioritize efforts to protect them. Such information could be collected in a

supplemental form conceptually analogous to supplemental forms used for crashes

resulting in fatalities and crashes involving commercial vehicles. While it is

acknowledged that states are sensitive to increasing the length and complexity of forms

that law enforcement officers are required to complete when investigating crashes, it

should also be noted that far fewer than 1% of all crashes involve RAPs or emergency

response personnel, thus any such supplemental data form would be used extremely

rarely.

Finally, while fatalities represent the worst possible outcome, the current study

was unable to examine non-fatal RAP injuries. The current study relied on memorials of

RAPs who died to identify cases—a methodology clearly not generalizable to non-fatal

crashes. In a study of crashes related to disabled vehicles but not specific to RAPs, Spicer

et al. (2021) estimated that an average of 1,111 pedestrians were injured and 294 were

killed each year while working on, entering, exiting, or attending to disabled vehicles

nationwide in 2014 to 2018. If assumed to be applicable to RAPs, the ratio of non-fatal

injuries to fatalities in that study would suggest that for every RAP fatally struck, there

may be perhaps three to four additional RAPs who are struck as pedestrians and suffer

non-fatal injuries, or perhaps roughly 100 annually in recent years. Clearly more

research is needed to quantify and examine the characteristics of RAPs who are struck

by vehicles yet survive, both because any of these crashes could have been fatal had

circumstances been different, and because non-fatal injuries can nonetheless result in

inability to work, lost wages, and even life-altering disability. Contrasting the results of

the current study to previous studies of predominantly non-fatal crashes of incident

responders (Carrick & Srinivasan, 2023; Yu et al., 2013), it appears likely that the

characteristics of non-fatal crashes in which RAPs survive crashes differ in many ways

from fatal crashes.

Conclusion and Recommendations

This research represents the most comprehensive examination of struck-by

fatalities of roadside assistance providers of which the AAA Foundation for Traffic Safety

is aware. Results indicate that at least 123 roadside assistance providers were struck and

killed by vehicles between 2015 and 2021. There appears to have been an increasing

trend in fatalities of roadside assistance providers struck by vehicles during these years,

even larger than the increasing trend in traffic fatalities overall over the same period.

Results highlight the importance of reinforcing motorist compliance with Slow Down,

Move Over laws. However, results also indicate that a substantial proportion of these

incidents involve out-of-control vehicles that departed the road prior to striking roadside

assistance providers, suggesting an additional need for countermeasures to prevent

vehicles from departing the road in the first place, redirecting out-of-control vehicles to

prevent them from striking roadside assistance providers, and/or mitigating the severity

of any impact that may occur. Results also suggest, albeit inconclusively, that some of the

roadside assistance providers who died may have been working on the traffic-facing side

of the scene, reinforcing the importance of training to avoid operating on the traffic-

facing side of the scene to the greatest extent possible. Comparison of the number of

cases identified through industry memorial websites versus through variables coded in

crash data highlight the need for improvements in data collection forms and reporting

procedures to ensure that roadside assistance providers who are injured or killed in the

line of duty are identifiable in crash reports and databases, as such data are critical to

researchers as well as practitioners to identify, develop, and prioritize ways to protect

roadside assistance providers.

Based on the findings of this research, as well as the limitations encountered, this

report makes the following recommendations:

• There is a need to reinforce public awareness of and increase motorist

compliance with Slow Down, Move Over laws; research to investigate the most

effective approaches to increase compliance is also needed.

• Countermeasures are needed to protect roadside assistance providers and first

responders from out-of-control vehicles that depart the roadway by preventing

the original road departure, by protecting those at the roadside from being struck,

or by reducing the severity of any impact that may occur. Research is needed to

determine what countermeasures are most effective and practical to meet this

need.

• Training for roadside assistance providers should emphasize the importance

avoiding working or standing on the traffic-facing side of the incident scene

whenever possible and minimize time spent on the traffic-facing side of the scene

when it cannot be avoided altogether, as well as provide strategies for how to do

so.

• In cases where countermeasures are site-specific or where deployment must be

prioritized, results of the current study suggest that efforts should focus on

protecting roadside assistance providers working on high-speed limited-access

highways.

• State police crash report forms should include data fields designed to report

whether a crash victim was an incident responder, and type of responder when

applicable, as called for in the current edition of the Model Minimum Uniform

Crash Criteria (NHTSA, 2024). Additionally, states should consider collecting

additional information in crashes in which roadside assistance providers and

other emergency response personnel are struck by vehicles, or at least in those

that result in the injury or death of a responder.

References

AAA. (2023). Slow Down/Move Over – AAA Exchange.

https://exchange.aaa.com/automotive/roadside-assistance/slow-down-move-over/

Agency Information Collection Activities; Notice and Request for Comment; Evaluation of

the Model Minimum Uniform Crash Criteria Program. (2022, March 29). Federal Register.

https://www.federalregister.gov/documents/2022/03/29/2022-06496/agency-information-

collection-activities-notice-and-request-for-comment-evaluation-of-the-model

Benson, A. J., Tefft, B. C., Arnold, L. S., & Horrey, W. J. (2021). Fatal hit-and-run crashes:

Factors associated with leaving the scene. Journal of Safety Research, 79, 76–82.

https://doi.org/10.1016/j.jsr.2021.08.007

Blomberg, R. D., Wright, T. J., Finstad, K., Brunsen, E., Van Houten, R., Radmehr, N., Yepez,

J., & Johnson, T. (2023). Protecting Roadside Workers: Field Evaluation of Flares, Cones,

and Tow Truck Light Patterns. https://aaafoundation.org/wp-

content/uploads/2022/12/Protecting-Roadside-Workers_Field-Evaluation-of-Cones-Flares-

Lighting-Report.pdf

Bunn, T. L., Slavova, S., Chandler, M., Hanner, N., & Singleton, M. (2018). Surveillance of

traffic incident management-related occupational fatalities in Kentucky, 2005-2016.

Traffic Injury Prevention, 19(4), 446–453. https://doi.org/10.1080/15389588.2018.1432042

Bureau of Labor Statistics Data. (n.d.). Retrieved July 24, 2023, from

https://data.bls.gov/pdq/SurveyOutputServlet

Carrick, G., & Srinivasan, S. (2023). Characterizing Incident Responder Crashes Involving

Move Over Law Violations. Transportation Research Record, 2677(2), 741–752.

https://doi.org/10.1177/03611981221108385

Chandler, M. D., & Bunn, T. L. (2019). Motor vehicle towing: An analysis of injuries in a

high-risk yet understudied industry. Journal of Safety Research, 71, 191–200.

https://doi.org/10.1016/j.jsr.2019.10.006

Emergency Responder Safety Institute. (2023). Yearly Fatality Reports 2019-2021

[dataset]. https://www.respondersafety.com/news/struck-by-incidents/yearly-fatality-

reports/

Horrey, W. J., Tefft, B. C., Arnold, L. S., & Villavicencio, L. (2021). Improving Roadside

Responder Crash Data: Outcomes from an Expert Roundtable Discussion (p. 10) [Research

Brief]. AAA Foundation for Traffic Safety. https://aaafoundation.org/wp-

content/uploads/2021/12/Improving-Roadside-Responder-Crash-Data-December-2021.pdf

Liu, J., Penmetsa, P., Yang, C., Hainen, A., & Barnett, T. (2023). Protecting Roadside

Workers: Field Evaluation of a Vehicle-Mounted Variable Message Sign and Examination of

Worker Perceptions and Use of Countermeasures. https://aaafoundation.org/wp-

content/uploads/2022/10/Protecting-Roadside-Workers-Field-Evaluation-of-Perceptions-

and-Countermeasures.pdf

National Highway Traffic Safety Administration. (2017). MMUCC Guideline: Model

Minimum Uniform Crash Criteria, Fifth Edition (DOT HS 812 433); pp. 60–61.

https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812433

National Highway Traffic Safety Administration. (2023a). Fatality Analysis Reporting

System 2015-2021 [dataset]. https://www.nhtsa.gov/file-

downloads?p=nhtsa/downloads/FARS/

National Highway Traffic Safety Administration. (2023b). Fatality Analysis Reporting

System Analytical User’s Manual 1975-2021 (DOT HS 813 417); pp. H-4. National Highway

Traffic Safety Administration.

National Highway Traffic Safety Administration. (2024). MMUCC Guideline: Model

Minimum Uniform Crash Criteria, 6th Edition (DOT HS 813 525a); pp. 244–246. National

Highway Traffic Safety Administration.

https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/813525

Spicer, R., Bahouth, G., Vahabaghaie, A., & Drayer, R. (2021). Frequency and cost of

crashes, fatalities, and injuries involving disabled vehicles. Accident Analysis &

Prevention, 152, 105974. https://doi.org/10.1016/j.aap.2021.105974

Tefft, B. C. (2012). Prevalence of motor vehicle crashes involving drowsy drivers, United

States, 1999-2008. Accident; Analysis and Prevention, 45, 180–186.

https://doi.org/10.1016/j.aap.2011.05.028

Tefft, B. C. (2013). Impact speed and a pedestrian’s risk of severe injury or death. Accident

Analysis & Prevention, 50, 871–878. https://doi.org/10.1016/j.aap.2012.07.022

Yang, C., Liu, J., Li, X., & Barnett, T. (2023). Analysis of first responder-involved traffic

incidents by mining news reports. Accident Analysis & Prevention, 192, 107261.

https://doi.org/10.1016/j.aap.2023.107261

Yu, L., Bill, A. R., Chitturi, M. V., & Noyce, D. A. (2013). On-Duty Struck-By Crashes:

Characteristics and Contributing Factors. Transportation Research Record, 2386(1), 112–

120. https://doi.org/10.3141/2386-13