Chapter 6 - Safety

Chapter 6 - Safety


Promoting "public health and safety by working toward the elimination of transportation-related deaths, injuries, and property damage" is a high priority of the U.S. Department of Transportation (DOT) [1]. The United States has made much progress in reducing the number of transportation-related deaths, but crashes and incidents involving transportation vehicles, vessels, aircraft, and pipelines still claimed over 44,000 lives and injured more than 3 million people in 2000. Transportation accidents are the ninth single leading cause of death in the United States (figure 1). However, motor vehicle crashes are the leading cause of death for people between 4 and 33 years of age [2].

Motor vehicle collisions account for about 95 percent of transportation-related deaths and an even higher percentage of transportation injuries. Human behavior-such as alcohol and drug use, reckless operation of vehicles, failure to properly use occupant protection devices, and fatigue-is a major factor in a high proportion of crashes.

DOT has set specific targets for the next few years to improve transportation safety. These include goals to lower the U.S. commercial air carrier fatal crash rate by 80 percent by 2007, reduce the highway fatality rate to 1.0 per 1 million vehicle-miles traveled by 2008, and reduce commercial truck-related fatalities by 50 percent by 2010. Specific safety initiatives for rail, transit, maritime, and pipelines are also in place.

1. U.S. Department of Transportation, Performance Plan, Fiscal Year 2003 (Washington, DC: March 2002).
2. U.S. Department of Transportation, National Highway Traffic Safety Administration, Traffic Safety Facts 2000: Overview (Washington, DC: October 2001).

Transportation Fatalities: A Modal Picture

Over the last 25 years, the total number of fatalities on the nation's roads, rails, and waterways and in the skies declined (figure 1). Despite progress, transportation crashes and incidents claimed 44,314 lives in 2000, of which 41,945 involved highway vehicles. Occupants of passenger cars and light trucks (i.e., sport utility vehicles, vans and minivans, and pickup trucks) accounted for over 70 percent of the transportation fatalities in 2000; pedestrians, motorcyclists, bicyclists, and others involved in motor vehicle collisions accounted for most of the remaining deaths (table 1).

Of the 2,369 transportation fatalities in 2000 that did not involve highway vehicles, recreational boating and general aviation (e.g., private planes for individual and business use) together claimed the lives of 1,293 people. Commercial carriers (airlines, trains, waterborne vessels, and buses) accounted for slightly under 1,000 fatalities. Many of these were bystanders and others outside of vehicles.

Transportation Fatality Rates

The more people travel, the greater the risk they incur. Thus, using the absolute numbers of fatalities to compare the safety of a given mode over time (table 1) can be misleading, since any change in the fatality numbers might be explained by a change in the amount of transportation activity. A clearer picture can be derived from exposure rates. Exposure rates are calculated by dividing the absolute numbers of fatalities (or other adverse outcome) by an activity measure, such as number of trips, number of miles traveled, or number of hours of vehicle operation.

Figure 1 shows fatality rates for selected modes for a time period of two decades or more. It is clear that safety in most modes has improved over the last 25 years. However, for several of the modes, the greatest improvement in fatality rates tended to occur in the earlier years of the period.

The activity measures used as denominators are not the same for all modes. For highway travel, exposure to risk is approximately proportional to distance traveled, hence the use of vehicle-miles as the denominator. For aviation, the greatest proportion of crashes occurs during takeoff and landing; hence risk is approximately proportional to the number of operations (measured as departures). Data on departures are not available for general aviation for recent years, so hours flown is used instead. For some means of travel, there are no good measures of the risks entailed. For example, while over 4,700 pedestrians were struck by motor vehicles and died in 2000, exposure measures are lacking because good data are not available for the amount of time, distances, or other circumstances of pedestrian travel.

Highway submodes show considerable improvement in fatality rates since 1975, when the federal government began to collect systematic national data from states. While all highway submodes show improved rates, there is much variation among them. Occupants of passenger cars and light trucks (including pickup trucks, vans, and sport utility vehicles) have much higher fatality rates than occupants of large trucks. Motorcycle riders have the highest fatality rate by far among the highway submodes (27.65 fatalities per 100 million vehicle-miles. The fatality rates per 100 million vehicle-miles are 1.31 and 1.22 for passenger car and light-truck occupants, respectively, versus 0.37 for large-truck occupants. A large number of factors influence the difference in fatality rates. For example, the greater size and mass of large trucks serves to protect the occupants of these vehicles in crashes with smaller vehicles or less massive objects.

Many factors may interact to explain the decreasing fatality rates. For highway modes, promotion of safety belt, child safety seat, and motorcycle helmet usage, and measures to discourage drunk driving have all had a beneficial effect. So, too, have improvements in vehicle and highway design and greater separation of traffic. Finally, some of the decrease in transportation fatalities may be a consequence of better and prompter medical attention for victims of transportation crashes and accidents.

Highway Crash Characteristics

The overwhelming majority of highway fatalities occur as a result of single-vehicle crashes and crashes involving two vehicles. For example, in 2000, 42 percent of traffic crash fatalities were vehicle occupants (including drivers) killed in single-vehicle crashes and 38 percent of fatalities occurred as a result of two-vehicle crashes (table 1). Crashes in which three or more vehicles were involved caused only 7 percent of traffic fatalities in 2000.

An average of one-third of all motor vehicle crash fatalities nationwide result from single vehicle run-off-the-road (ROR) crashes, and two-thirds of these ROR fatalities occur in rural areas. It has been estimated that 40 to 60 percent of these crashes are due to driver fatigue, drowsiness, or inattention. The Federal Highway Administration recommends the use of rumble strips along the roadway shoulder as an effective way to reduce these incidents. The noise produced by vehicle tires on these rumble strips warns drivers that they are leaving the roadway. Studies of the effectiveness of shoulder rumble strips indicate that they can reduce the overall rate of ROR crashes between 15 and 70 percent [1]. In the future, the development of in-vehicle technologies that detect driver drowsiness and inattention and suitably warn the driver may further reduce the incidence of such crashes.

Traffic crashes between light trucks or vans and passenger cars is of increasing concern. Since the early 1980s, the category of light trucks and vans (LTVs) has grown dramatically (figure 1). LTVs include pickup trucks, vans, minivans, truck-based wagons, and sport utility vehicles (SUVs).

Differences in vehicle size, weight, and geometry in multivehicle crashes can put occupants of passenger cars at greater risk in a crash with a light-duty truck than in a crash involving two or more passenger cars. For example, a study done for NHTSA by the University of Michigan Transportation Research Institute shows that when an SUV strikes a passenger car in a frontal crash, occupants of the car are almost twice as likely to have fatal injuries as the occupants of the SUV. In frontal collisions between two cars of similar weight, the ratio of deaths is 1:1. The same study found that, in side impact crashes, SUVs are more injurious as a striking vehicle than are passenger cars. For example, when SUVs strike passenger cars on the left side, the risk of death to the car driver can be 25 times greater than the risk to the SUV occupant. However, in the same type of crash involving two cars, the risk of death to the driver of the car being struck is only 10 times greater than the occupant of the other car [3].

Another issue related to SUVs is their propensity to rollover during certain steering maneuvers. SUVs are constructed with higher ground clearance for occasional offroad use and, thus, have a higher center of gravity. SUV height, along with other factors, contributes to the average rate of 98 rollover fatalities per million registered vehicles compared with 44 such fatalities per million registered vehicles for all other light vehicle types [4]. Also, in fatal crashes in 2000 SUVs were twice as likely to rollover when compared to passenger cars, increasing the risk of occupant ejection, fatality, or injury [2].

1. U.S. Department of Transportation, Federal Highway Administration, Effectiveness of Rumble Strips, available at, as of Dec. 20, 2001.
2. U.S. Department of Transportation, National Highway Traffic Safety Administration, Fatality Analysis Reporting System, available at, as of April 2002.
3. U.S. Department of Transportation, National Highway Traffic Safety Administration/University of Michigan Transportation Research Institute, Fatality Risks in Collisions Between Cars and Light Trucks (Washington, DC: September 1998).
4. U.S. Department of Transportation, Office of the Assistant Secretary of Public Affairs, News Release: DOT Requires Upgraded Rollover Warning Label for Sport Utility Vehicles, Mar. 5, 1999, available at, as of Dec. 20, 2001.

Highway Crashes on Rural and Urban Roads

Two- and three-lane rural roads make up the majority of the highway system in the United States. If Interstate highways are excluded, these rural roads represent four times the highway mileage of urban roads in the U.S. highway system [1].

Almost 60 percent of all fatal crashes in 2000 occurred on rural roads. Seventy-one percent of these crashes were on roads with speed limits of 55 mph or more (table 1). Conversely, almost 70 percent of urban highway crashes occur on roads with speed limits under 55 mph. Irrespective of speed limits, most rural and urban highway crashes occur on principal arterial and other roads rather than on Interstate highways; 88 percent of crashes in the case of rural roads and 86 percent, for urban roads [2].

Road conditions contribute to the greatest proportion of fatal crashes in rural areas. In particular, two-way traffic on roads posted for high speed limits is a concern. Rural drivers often must deal with challenging road geometry (e.g., width, alignment, and sight distances) and challenging geography (e.g., steep grades and mountain passes). Adverse weather can further affect rural road conditions and sparse and patchy telecommunications infrastructure can slow emergency response time when a crash occurs.

1. Transportation Research Board, National Cooperative Highway Research Program, Accident Mitigation Guide for Congested Rural Two-Lane Highways, Report 440 (Washington, DC: 2000).
2. U.S. Department of Transportation, National Highway Traffic Safety Administration, Fatality Analysis Reporting System (FARS) database, available at, as of April 2002.

Alcohol-Related Highway Crashes

In 2000, 41 percent of the 41,945 highway fatalities were related to alcohol, a 2 percent increase over 1999. In 1982, the first year for which data are available, 25,165 people died in alcohol-related motor vehicle crashes-57 percent of all highway fatalities. By 2000, alcohol-related fatalities had dropped to 17,380 (figure 1). The U.S. Department of Transportation has a goal of reducing alcohol-related fatalities to no more than 11,000 by 2005 [2].

Improved state and local education programs, stricter law enforcement, adoption of a 0.08 blood alcohol concentration (BAC) by 33 states and the District of Columbia, higher minimum drinking ages, more stringent license revocation laws, and reduced tolerance for drinking and driving have all been cited as factors in reducing alcohol-related deaths [1]. Despite improvements, 22 percent of passenger car drivers, 20 percent of light truck drivers, 1 percent of large truck operators, and 29 percent of motorcycle operators involved in fatal crashes in 2000 were legally intoxicated with a BAC of 0.10 or greater [3].

Just over 38 percent of the drivers between the ages of 21 and 24 who were involved in a fatal motor vehicle highway crash in 2000 had a BAC of 0.01 or more; over 32 percent had a BAC of 0.08 or more (figure 2). While the highest of any age group, this does represent a decline from 46 percent and 39 percent, respectively, in 1990. Overall, the percentage of all drivers with any alcohol content (BAC of 0.01 or more) involved in a fatal highway crash declined from 33 percent to 26 percent between 1990 and 2000 [3].

Alcohol-related fatalities declined more quickly in the 1980s than in the 1990s. Between 1994 and 2000, the percentage of highway fatalities attributed to alcohol declined by only 1 percent-from 42 percent to 41 percent. Moreover, while alcohol-related fatalities among drivers 16 to 20 years of age decreased, alcohol consumption in this age group increased every year from 1993 to 2000 [2].

In 2000, Congress enacted legislation that provides strong encouragement for states to adopt the 0.08 BAC [5]. States have until October 1, 2003, to pass the stricter limit or face the withholding of 2 percent of their federal highway construction funds. After 2003, states that fail to pass the 0.08 BAC will lose an additional 2 percent of their federal funding every year. By October 1, 2006, and each year thereafter, states that still have not adopted 0.08 BAC laws will lose 8 percent of their funding [4].

Fatality rates vary by state (see map). It is illegal in every state and the District of Columbia to drive a motor vehicle while under the influence of alcohol. In addition, every state has laws that make it illegal for a person to drive a motor vehicle with a specific amount of alcohol in his or her blood. As of November 2002, 17 states defined intoxicated driving as 0.10 BAC-the level at which a person's blood contains 1/10th of 1 percent of alcohol. Thirty-three states and the District of Columbia have enacted 0.08 BAC laws1 [1].

Highway safety advocates have encouraged states to take a systems approach to reducing drunk driving. Some states have enacted a combination of measures. In addition to 0.08 BAC limits, such measures include stringent license revocation laws (under which a person deemed to be driving under the influence has his or her driving privileges suspended or revoked), comprehensive screening and treatment programs for alcohol offenders, vehicle impoundment, and zero tolerance BAC and other laws for youths [6].

1Massachusetts has adopted a 0.08 BAC law but does not meet federal requirements to avoid sanctions in 2003 under the federal 0.08 law.

1. Insurance Institute for Highway Safety/Highway Loss Data Institute, DUI/DWI Laws as of November 2002, available at http://www., as of November 2002.
2. U.S. Department of Transportation, FY 2000 Performance Report/FY 2002 Performance Plan (Washington, DC: 2001), also available at, as of Dec. 3, 2001.
3. U.S. Department of Transportation, National Highway Traffic Safety Administration, personal communication, Nov. 1, 2002.
4. _____. "Congress Agrees to 0.08% Blood Alcohol as the Legal Level for Impaired Driving," NHTSA Now Newsletter, Oct. 16, 2000.
5. U.S. Department of Transportation, Office of Public Affairs, "Statement by U.S. Transportation Secretary Rodney Slater Upon Signing of Transportation Appropriations Act by President Clinton," Oct. 23, 2000.
6. U.S. General Accounting Office, Resources, Community, and Economic Development Division, Highway Safety: Effectiveness of State 0.08 Blood Alcohol Laws(Washington, DC: June 1999).

Occupant Protection: Safety Belts, Air Bags, and Child Restraints

The National Highway Traffic Safety Administration (NHTSA) estimates that, in 2000, safety belts saved the lives of 11,889 passenger vehicle occupants over 4 years old (figure 1). NHTSA also estimates that 21,127 lives could have been saved that year if all passenger vehicle occupants aged 4 and older wore safety belts [1].

The number of lives saved has increased dramatically since 1984 when states began to enact safety belt laws. A June 2001 NHTSA survey showed that 73 percent of passenger vehicle occupants used safety belts [3]. Usage rates differ noticeably among the states based on how the safety belt laws are enforced (see map). There are three levels of enforcement: primary enforcement allows a police officer to stop and cite someone for not wearing a safety belt; secondary enforcement allows a police officer to cite someone for not wearing a safety belt only if they have been stopped for some other infraction; and no enforcement [1]. Usage in the 17 states with primary enforcement was 78 percent as opposed to 67 percent in the 33 states with secondary enforcement laws (see map). Safety belt usage in New Hampshire, which does not require adults to wear safety belts, was 56 percent.

Beginning in September 1997 (model year 1998), all new passenger vehicles were required to have driver and passenger air bags. The following year, the same requirement was applied to light trucks. NHTSA estimates that, as of 2000, more than 106 million air-bag-equipped passenger vehicles were on the road, including 81 million with dual air bags. In 2000, an estimated 1,584 lives were saved by air bags. From 1987 through 2000, an estimated total of 6,553 lives were saved [1].

According to NHTSA, air bags, combined with safety belts, offer the most effective safety protection available today for passenger vehicle occupants. Air bags are supplemental protection and are designed to deploy in moderate-to-severe frontal crashes. Adults and some children riding in front seats have been injured or killed by air bags inflating in low severity crashes. While far more lives have been saved by air bags than have been lost, since 1990, 195 deaths from injuries caused by air bags have occurred. This includes 119 children riding in the front seat [2]. If children under the age of 13 ride in the back seat of passenger vehicles and are secured by appropriate restraint systems, risk of injuries or death from air bags can be avoided [1].

In 2000, 541 passenger vehicle occupant fatalities were reported among children less than 5 years of age [1]. NHTSA estimated that in 2000, use of child restraint systems saved the lives of 316 children under the age of 5. An additional 143 lives could have been saved-for a total of 458-if every child under age 5 had been properly restrained in a child safety seat. From 1975 through 2000, an estimated 4,816 lives were saved by child restraints [1].

1. U.S. Department of Transportation, National Highway Traffic Safety Administration, Traffic Safety Facts 2000: Occupant Protection (Washington, DC: 2001); and personal communication, Nov. 1, 2002.
2. ____. Air Bag Fatalities and Serious Injury Report (Washington, DC: 2001), also available at, as of Dec. 10, 2001.
3. ____. "Research Note: Observed Shoulder Belt Use from the June Mini National Occupant Protection Use Survey (NOPUS), August 2001.

Cell Phones and Motor Vehicles

As cell phones have gained in popularity, they have become a growing concern with respect to highway safety. Nearly 120 million Americans now subscribe to a cell phone service, up from 7.6 million in 1991 (figure 1). Many Americans use their phones while driving. The National Highway Traffic Safety Administration (NHTSA) estimates 3 percent of passenger vehicle drivers are talking on hand-held cell phones at any given time. Of the 54 percent of drivers who usually have a wireless phone in their vehicle, 73 percent reported using their phone while driving. [5]

The precise effects of cell phone use on public safety have not been determined, but preliminary studies suggest that cell phones distract drivers. A forthcoming study by the National Safety Council found that drivers engaged in cell phone conversations missed twice as many simulated traffic signals and took longer to react to those signals they detected. The results were the same for hand-held and hands-free devices [2].

State police forces have not been required to collect information on cell phones in vehicle accident reports until recently. Within the past two years 18 states have enacted laws requiring police to include information on cell phone usage in vehicle accident reports. Louisiana, New Jersey, New Mexico, Pennsylvania, and Virginia have approved cell phone and driving studies [1].

State legislators have also begun to regulate cell phone use in motor vehicles (table 1). New York became the first state to ban drivers from using hand-held devices. Massachusetts requires that drivers keep one hand on the steering wheel at all times, which may encourage increased use of hands-free devices. Due to a concern that hands-free devices might interfere with drivers' ability to hear noises from the surrounding environment, Illinois and Florida have prohibited the use of headsets except for single-sided headsets. Arizona and Massachusetts school bus drivers are prohibited from using cell phones while driving, and several other states have proposed similar legislation. As of late 2001, an additional 20 states were debating laws related to cell phones and motor vehicles.

Despite safety concerns due to driver distraction, wireless technologies provide some clear safety and traffic management benefits. The Cellular Telephone and Internet Association estimates that motorists on cell phones place 139,000 emergency calls every day. Most laws restricting cell phone use have recognized this benefit by allowing motorists to place calls in an emergency. In a 1997 study, NHTSA found that state police are generally appreciative of the quick notification capabilities afforded by cell phones [4]. Furthermore, that study and a 1997 study published in the New England Journal of Medicine found that cell phones can reduce emergency response times and save lives [3]. In addition to being helpful in emergency situations, the NHTSA study found that cell phones enable motorists to quickly notify authorities of road hazards, congestion, or problem drivers. And in the case of roadside mechanical problems, cell phones enhance drivers' personal security by allowing them to contact help immediately.

1. National Conference of State Legislators, Cell Phones and Highway Safety: 2001 State Legislative Update, available at programs/esnr/2001cellph.htm, as of Nov. 19, 2001.
2. National Safety Council, "Does Cell Phone Conversation Impair Driving Performance?" preliminary report, available at, as of Nov. 14, 2001.
3. Cohen P.J., K.P. Quinlan, O. Paltiel, A. Ambrose, D.A. Redelmeier, R.J. Tibshirani, M. Maclure, and M.A. Mittleman. "Cellular Telephones and Traffic Accidents," New England Journal of Medicine vol. 336, No. 7, Feb. 13, 1997, pp. 453-458, abstract available at, as of Nov. 19, 2001.
4. U.S. Department of Transportation, National Highway Traffic Safety Administration, An Investigation of the Safety Implications of Wireless Communications in Vehicles, available at, as of Nov. 14, 2001.
5. U.S. Department of Transportation, National Highway Traffic Safety Administration, "National Occupant Protection Use Survey 2000," available at, as of Nov. 14, 2001.

Large Trucks

Crashes involving large trucks1 resulted in 5,282 fatalities in 2000. Annually, the number of such fatalities varies, from a low of 4,462 in 1992 to a high of 6,702 in 1979 (figure 1). The number of drivers and occupants of large trucks killed in crashes has declined since the late 1970s, when fatalities averaged about 1,200, compared with 672 in the 1990s. The overwhelming majority of people killed in large truck collisions-78 percent in 2000-were occupants of other vehicles or nonmotorists [1].

In two-vehicle crashes involving a large truck and a passenger vehicle, driver-related crash factors were cited by police officers at the scene for 25 percent of the truck drivers involved and for 82 percent of the passenger vehicle drivers. Table 1 shows the percent of crashes in which either the large-truck driver or the passenger-vehicle driver or both were cited for one or more of the top 12 factors identified.

Large truck safety issues have received increased attention in recent years. In 1999, Congress passed the Motor Carrier Safety Improvement Act, which created the Federal Motor Carrier Safety Administration2 within the U.S. Department of Transportation. Among other provisions, the legislation calls for increased roadside inspections, compliance reviews and enforcement actions, improvements in safety data, and additional research into crash causes. About 24 percent of the over 2.4 million motor carrier vehicles inspected in 2000 were taken out of service (figure 2 and figure 3).

1Trucks with a gross vehicle weight greater than 10,000 pounds.

2 In addition to large trucks, the Federal Motor Car-rier Safety Administration oversees passenger vehicles designed to transport eight or more persons and vehicles used to transport hazardous materials.

1. U.S. Department of Transportation, National Highway Traffic Safety Administration, Traffic Safety Facts 2000: Large Trucks (Washington, DC: 2001).


Almost 700 bicyclists and other pedalcyclists1 were killed in crashes with motor vehicles in 2000, a 31 percent reduction since 1975 (figure 1). This continues a 25-year trend of a steady decline in the number of pedalcycle deaths and, to a lesser extent, in the pedalcyclist portion of the overall motor vehicle fatalities. Pedalcyclist fatalities in 2000 amounted to only 1.6 percent of all 41,945 highway fatalities; whereas, in 1975, they were 2.2 percent. A similar trend exists for pedalcyclist injuries. An estimated 51,000 were injured in traffic crashes in 2000 compared with 67,000 in 1995 [5, 6, 7].

It is difficult to fully assess pedalcyclist fatality and injury trends, because exposure data for the mode-that is, the number of trips pedalcyclists make or the amount of time they spend cycling-are limited and vary by season. For instance, various Bureau of Transportation Statistics Omnibus surveys show that in the summer months about 20 percent of adults ride bicycles but in the winter months only about 10 percent do [4].

Along with the decreasing number of fatalities in the last 25 years, another trend is apparent. Most pedalcycle deaths (involving motor vehicles) now occur among riders 15 years of age and older, rather than among children and young teens (figure 2). The fatalities in 1975 attributed to child pedalcyclists (under 15 years old) represented 61 percent of the total that year. This age group's fatalities in 2000, however, comprised only 26 percent of the total while the 25 to 54 age group sustained the highest (43 percent). In 1975, the latter age group represented only 10 percent of the total fatalities. The same phenomenon has occurred in the 55 to 64 and over 65 age groups. When the pedalcycle fatalities are compared with the number of people by age group in 1975 and 2000, a different pattern emerges (table 1). While there were almost 5 pedalcycle fatalities per million people in 1975, by 2000 fatalities were down to less than 3. In 1975, these fatality rates were much higher for those under 15 years of age, while in 2000 the rate (or likelihood of being killed) was about the same for all age groups. The median age of pedalcyclist fatalities shifted from 13 years old in 1975 to 36 years old in 2000.

Factors other than age appear to affect pedalcycling fatalities. In 2000, for instance, most pedalcycle fatalities (63 percent) occurred in urban areas, with two-thirds of these arising at nonintersection locations. In addition, most of the pedalcyclists injured or killed in 2000 were males: 78 percent and 88 percent, respectively [2]. In more than one-third of pedalcyclist fatalities, alcohol use was reported; one-fifth (21 percent) of pedalcyclists killed were intoxicated [8]. Nearly one-third of pedalcyclists involved in crashes were riding against traffic [1, table 37]. In fact, a study that calculated relative risk based on exposure rates found that pedalcycling against traffic increased the risk of a collision with a motor vehicle by a factor of 3.6 [9].

Although 90 percent of pedalcycle fatalities involve a collision with a motor vehicle, most pedalcycle injuries do not. There are about 500,000 pedalcycle-related emergency room visits annually [3]. Most of these pedalcycle mishaps involve falls and collisions with fixed objects; collisions with motor vehicles account for just 15 percent of the visits [2]. Collisions with pedestrians, other pedalcycles, and animals are apparently prevalent, but states do not generally record these data since they do not involve motor vehicles.

1A pedalcyclist is a person on a vehicle powered solely by pedals. Pedalcycles may have one to four wheels. Bicycles (pedalcycles with two wheels) are the dominate type.

1. Hunter, W.W., J.C. Stutts, and W.E. Pein, Pedalcycle Crash Types: A 1990s Informational Guide, FHWA-RD-96-104 (Washington, DC: U.S. Department of Transportation, Federal Highway Administration, 1997).
2. Rivara, F.P., D.C. Thompson, and R.S. Thompson, Circumstance and Severity of Pedalcycle Injuries (Seattle, WA: Snell Memorial Foundation, Harborview Injury Prevention and Research Center, 1996).
3. Tinsworth, D., C. Polen, and S. Cassidy, Pedalcycle-Related Injuries: Injury, Hazard, and Risk Patterns, Technical Report (Washington, DC: U.S. Consumer Product Safety Commission, 1993).
4. U.S. Department of Transportation, Bureau of Transportation Statistics, Omnibus Survey, August 2000-December 2001.
5. U.S. Department of Transportation, National Highway Traffic Safety Administration, Fatality Analysis Reporting System (FARS) database, 2000.
6. ______. Traffic Safety Facts 1996-Overview (Washington, DC: 1996).
7. ______. Traffic Safety Facts 2000-Overview (Washington, DC: 2000).
8. ______. Traffic Safety Facts 2000-Pedalcyclists (Washington, DC: 2000).
9. Wachtel, A. and D. Lewiston, "Risk Factors for Pedalcycle-Motor Vehicle Collisions at Intersections," ITE Journal, September 1994, pp. 30-35.


In 2000, 4,763 pedestrians were killed in crashes involving motor vehicles, a 37 percent reduction since 1975 (figure 1). Many factors can contribute to motor vehicle-related pedestrian fatalities (table 1).

Data evaluating exposure risks faced by pedestrians are very limited. State data on pedestrian fatalities per 100,000 population show that the levels of fatalities vary across the country (figure 2).

Pedestrians comprised less than 3 percent of the 3,189,000 people injured in motor vehicle crashes in 2000, but just over 11 percent of the fatalities involving motor vehicles. The majority of pedestrian fatalities in 2000 occurred in urban areas (71 percent), at nonintersection locations (78 percent), in normal weather conditions (91 percent), and at night (64 percent). Additionally, males accounted for about 68 percent of the pedestrian fatalities in 2000. An estimated 31 percent of pedestrians killed in traffic crashes in 2000 were intoxicated (with a blood alcohol concentration of 0.10 or more), whereas only 13 percent of the drivers involved in fatal pedestrian crashes were [1].

1. U.S. Department of Transportation, National Highway Traffic Safety Administration, Traffic Safety Facts 2000: Pedestrians (Washington, DC: 2001).

Commercial Aviation

Despite the tragic loss of four U.S. airliners to a terrorist attack in September 2001, aviation continues to be a remarkably safe mode of transportation. U.S. air carriers experience less than one fatal crash for every million flights [1]. Preliminary statistics show that trend unchanged in spite of September 11.

There were 36 air carrier accidents in 2001, including the 4 crashes on September 11 (figure 1). Despite the loss of four aircraft on September 11 and American Airlines Flight 587 on November 12, the air carrier fatality rate has remained stable. Fatalities resulting from criminal acts are not included in the calculation of accident fatality rates (see box). Air carriers experienced 0.22 deaths per 100,000 flight hours in 2001, the same fatality rate in 1991 and less than that in 2000 (0.32 deaths per 100,000 hours) (figure 2).

However, the six fatal air carrier accidents in 2001 resulted in 531 fatalities (table 1). That is the highest number of air carrier fatalities in over 40 years. On September 11, all 265 people on board 4 planes were killed when the planes were crashed by terrorists. American Airlines Flight 587 was en route from John F. Kennedy International Airport in New York City to Santo Domingo in the Dominican Republic when it crashed shortly after takeoff on November 12. All 260 people on board and 5 people on the ground were killed by the crash. One ground worker was struck and killed by an airplane propeller in August.

In addition to air carriers, there are two other categories of commercial aviation flights: scheduled commuter flights and on-demand air taxis.1 There were 2 fatal U.S. commuter plane accidents and 18 fatal .on-demand air taxi accidents during 2001, resulting in 73 deaths [1]. Overall, commercial aviation, including air carrier, commuter, and on-demand air taxi flights, accounted for 52 percent of all air-related fatalities in 2001. The remaining fatalities resulted from general aviation accidents.

The overall accident rate for all three types of commercial aviation operations combined is 0.58 accidents per 100,000 flight hours. However, differences in the accident rates among the three types of operations do exist (figure 2). For example, the accident rate for air carriers has historically been well below that of commuter carriers and air taxis.

Finally, although the overall accident and fatality rates for commercial aviation remain low, the continued growth forecast for U.S. aviation in the coming decade raises concern. The Federal Aviation Administration (FAA) estimates that commercial aviation aircraft (excluding air taxis) will fly more than 24 million hours in 2007, a 37 percent increase over 1999. Commercial aviation (excluding air taxis) experienced an average of six fatal accidents each year in the United States between 1994 and 1996. If the projected growth in flight hours occurs and the fatal accident rate is not reduced, aviation experts estimate that the number of fatal commercial aviation accidents could rise to nine per year by 2007. To address this potential danger, FAA's "Safer Skies" program has a goal of reducing the number of fatal commercial accidents per million flight hours by 80 percent by 2007 [2].

1For safety reporting and analysis, commercial aviation consists of air carriers (those with aircraft having 10 or more seats), cargo haulers, commuter carriers (those with aircraft having 9 seats or fewer in scheduled service), air taxi service (those carriers with aircraft having 9 seats or fewer in unscheduled service), and helicopter service.

1. National Transportation Safety Board, Accidents, Fatalities, and Rates: 1982-2001 (Washington, DC: 2002), also available at, as of April 2002.
2. U.S. Department of Transportation, Federal Aviation Administration, Safer Skies: A Focused Agenda, 2000, available at http://www/, as of Sept. 20, 2000.

General Aviation

Most aviation accidents involve general aviation (GA) aircraft1 (table 1); however, GA fatalities and fatality rates have decreased over the last quarter century (figure 1). In 1975, general aviation experienced 1,252 fatalities-over twice as many as the 553 reported in 2002 (preliminary data). Moreover, the fatality rate (expressed as fatalities per 100,000 hours flown) declined from 4.35 to 1.22 over the same period [2].

The major causes of fatal general aviation accidents are weather, pilot loss of control or other maneuvering errors made during flight, and accidents on approach to the airport [5]. Nearly one-quarter of all general aviation accidents between 1989 and 1999 were related to weather [3]. Furthermore, the number of fatalities also varies a great deal by month, with fewer fatalities generally occurring in the winter months because of fewer flights (figure 2).

Another area of concern is the growing number of runway incursions2 involving GA aircraft. In 1999, GA pilot error caused 139 (76 percent) of the 183 runway incursions [1].

Changes in flight hours can also affect accident rates. The Federal Aviation Administration (FAA) estimates that GA flight hours will increase to about 36 million hours by 2007-nearly 19 percent higher than 1999. Although general aviation accidents and fatalities have been trending downward for 25 years, aviation experts believe these numbers will rise over the next decade with the projected increase in flight hours. Because of the potential safety implications associated with rapid growth in both commercial and GA flight hours, FAA initiated the "Safer Skies" program in 1998 with the goal of reducing aviation accident rates [4].

1General aviation includes a wide variety of aircraft, ranging from corporate jets to small piston-engine aircraft used for recreational purposes, as well as helicopters, gliders, and aircraft used in operations such as firefighting and agricultural spraying.

2 A runway incursion is any occurrence on a runway involving an aircraft, vehicle, or pedestrian that creates a collision hazard for aircraft taking off, intending to take off, landing, or intending to land.

1. Deyoe, Robin, Runway Safety Program Office, Federal Aviation Administration, U.S. Department of Transportation, personal communication, Sept. 13, 2000.
2. National Transportation Safety Board, Accidents, Fatalities, and Rates: 1982-2000 (Washington, DC: 2001), also available at, as of Apr. 17, 2001.
3. U.S. Department of Transportation, Federal Aviation Administration, Weather Study Index, available at, as of June 2002.
4. ____. Safer Skies: A Focused Agenda, 2000, available at http://www/, as of Sept. 20, 2000.
5. U.S. General Accounting Office, Resources, Community, and Economic Development Division, Aviation Safety: Safer Skies Initiative Has Taken Initial Steps to Reduce Accident Rates by 2007 (Washington, DC: June 2000).

Commercial Maritime Vessel Incidents

About 50,000 commercial vessels carrying freight and passengers call at U.S. ports every year. In 2000, there were almost 7,000 verified U.S. and foreign vessel incidents1 in U.S. waters. Over the last six years, the number of commercial vessel incidents in U.S. waters has declined (table 1). Approximately 90 percent of these incidents occurred among 10 vessel types, and this concentration has been increasing since 1997.

Towboats and tugboats have ranked as the number one vessel type involved in incidents since 1994. Prior to 1994, fishing vessels ranked number one; they now rank second. However, the number of incidents involving both of these vessel types has been declining in recent years [5]. Towboats and tugboats primarily push and pull barges on U.S. inland waterways and provide tug assist services in ports and along coastal areas. Towboats and tugboats, which can handle as many as 35 barges at a time, have limited maneuverability, especially when the crew is involved in maneuvering barges [4]. People falling overboard account for the majority of the fatalities in the inland towing industry [2].

A study of U.S. maritime incident data revealed that in 2000 the highest proportion (42 percent) of all maritime fatalities occurred among commercial fishing vessels. The next highest proportion of fatalities were among towboats and barges (11 percent), freight ships (10 percent), and passenger vessels (10 percent) [1]. The U.S. Coast Guard, which estimates that there are between 100,000 to 120,000 vessels in the U.S. commercial fishing fleet, believes the industry to be one of the most hazardous in the nation [3]. The number of fishing vessel worker fatalities may be climbing after a drop in 1997 (see figure 1). This may be due to increased economic pressure and competition in the commercial fishing industry, which encourages risk taking [3].

The number of recreational boats involved in commercial vessel incidents has been climbing since 1996 (figure 2). The safety of these boaters can be dependent on their ability to identify commercial vessels, particularly tugboats and towboats, and accurately assess their movements [2].

1 Incidents are defined as collisions, groundings, and "allisions" (when two vessels sideswipe each other).

1. Ungs, Timothy J. and Michael L. Adess, U.S. Department of Transportation, U.S. Coast Guard, Water Transportation and the Maritime Industry, available at, as of February 2001.
2. U.S. Department of Transportation, U.S. Coast Guard, "Epilogue," American Waterways Operators, available at, as of February 2001.
3. ____. U.S. Coast Guard Marine Safety and Environmental Protection Business Plan FY 2001-2005, available at, as of February 2001.
4. U.S. Department of Transportation, U.S. Coast Guard, Marine Safety Office, Providence, RI, available at, as of Feb. 24, 2001.
5. U.S. Department of Transportation, U.S. Coast Guard, Resources Management Directorate, Data Administration Division, personal communication, February 2001.

Recreational Boating

Most fatalities, injuries, and accidents on the water involve recreational boating. In 2000, the U.S. Coast Guard (USCG) reported a total of 7,740 recreational boating accidents and 4,355 injuries (figure 1). Personal watercraft and open motorboats account for the highest number of these accidents. Although fatalities remain high, the number has declined from 865 in 1990 to 701 in 2000. More than one-third of recreational boating accidents involved collisions with other vessels in 2000 (table 1). Substantially more drownings were related to the use of open motorboats than for any other type of recreational craft (table 2).

The majority of recreational boating accidents occurred during vessel operation and were caused by operator error, such as recklessness, inattention, and speed (table 3). Alcohol involvement accounted for 6.8 percent of accidents due to operator error in 2000. USCG found that 84 percent of all boating fatalities occurred on boats where the operator lacked safe boating education [1].

Regardless of the cause of the accident or the type of boat involved, boaters can improve their chances of survival by wearing life jackets or using other personal flotation devices (PFDs). Eight out of 10 fatal boating accident victims were not wearing a PFD. USCG estimates that the use of life jackets could have saved the lives of 445 drowning victims in 2000 [1].

1. U.S. Department of Transportation, U.S. Coast Guard, Office of Boating Safety, Boating Statistics 2000 (Washington, DC: 2001).


Most railroad fatalities occur on railroad rights-of-way and at highway-rail grade crossings, not on trains. (Railroad casualties include people killed and injured in train and nontrain incidents and accidents on railroad-operated property.) Of the 937 people killed in accidents and incidents involving railroads in 2000, only 4 were train passengers. As major train accidents are relatively infrequent, the number of fatalities fluctuates from year to year (table 1). The fatality rate per million train-miles changed little between 1978 and 1993, but since that time has dropped by about 40 percent (figure 1).

Although far fewer people die in highway-rail grade-crossing accidents than in the past, the toll is still large (figure 2). Of the 425 lives lost in 2000 in this type of accident, none were passengers on trains; all were in motor vehicles or on foot [1].

Trespassers not at grade crossings (people on railroad property without permission) accounted for 463 (49 percent) of the railroad deaths in 2000. Better understanding of trespassing and its motivations could be essential to addressing this high toll.

1. U.S. Department of Transportation, Federal Railroad Administration, Railroad Safety Statistics Annual Report 2000 (Washington, DC: July 2001).


Pipelines carry vast quantities of liquids and gases to fuel the nation's commercial and consumer demands. Transmission pipelines, which total about 325,000 miles, transport natural gas over long distances from sources to communities. Distribution pipelines, which then move natural gas to residential, commercial, and industrial customers, consist of about 1.7 million miles and are primarily intrastate. Hazardous liquid pipelines total about 156,000 miles and transport mostly crude oil and refined petroleum products to terminals and facilities [2].

Pipelines are a relatively safe way to transport energy resources and other products, but they are subject to forces of nature, human actions, and material defects that can cause potentially catastrophic accidents [3]. The U.S. Department of Transportation issues regulations covering pipeline design, construction, operation, and maintenance for both natural gas and interstate hazardous liquid pipelines.

The number of fatalities related to pipeline incidents varies from year to year, reflecting the high consequences associated with a limited number of failures (figure 1). The 38 pipeline fatalities in 2000 were more than twice the number recorded in 1975 [1]. One natural gas pipeline rupture was responsible for killing 12 people near the Pecos River in Carlsbad, New Mexico, in August 2000. It was the deadliest pipeline incident in the continental United States in almost 25 years. Overall, there were 234 gas pipeline incidents and 146 liquid pipeline accidents in 2000.1

Major causes of pipeline accidents include excavation and other outside force damage, material failure, and corrosion. However, the causes vary by year and by type of pipeline. Most gas pipeline incident reports consistently cite external damage (i.e., "damage by outside force") (table 1 and table 2). However, for liquid pipelines the "other" category most often designates the cause, depending on the year (table 3).

Major advances in the materials used for pipes and welding, inspections, and the installation process over the past 25 years have reduced the number of leaks and made those that take place less severe. New corrosion coatings and new application processes have produced dramatically longer lives for pipes.

See box for Pipeline Accident and Incident Data Reporting.

1The use of the terms "gas pipeline incident" and "liquid pipeline accident" does not imply different types of events but rather is a consequence of the official reporting forms.

1. National Transportation Safety Board, We Are All Safer, SR-98-01, 2nd ed. (Washington, DC: July 1998), also available at, as of Dec. 20, 2001.
2. U.S. General Accounting Office, The Office of Pipeline Safety is Changing How it Oversees the Pipeline Industry, GAO/RCED-00-128 (Washington, DC: May 2000).
3. U.S. Department of Transportation, Research and Special Programs Administration, Office of Pipeline Safety, Pipeline Statistics (Washington, DC: 2000), also available at, as of Dec. 20, 2001.

Hazardous Materials Transportation Accidents and Incidents

Like all modes of transportation, the movement of hazardous materials comes with the risk of accidents and incidents, including the threat of explosion, fire, or contamination of the environment. The safe transportation of hazardous materials has long been an area of governmental concern and oversight. The U.S. Department of Transportation (DOT), together with the Nuclear Regulatory Commission (for radioactive materials), are responsible for developing safety regulations for the transportation of hazardous materials, including training and packaging requirements, emergency response measures, enforcement, and data collection.

In 2000, over 17,000 incidents were reported to DOT's Hazardous Materials Information System (HMIS), the primary source of national data on hazardous materials transportation safety (table 1). These incidents resulted in 13 deaths and 244 injuries directly attributable to the materials being transported [1]. More than 85 percent of reported incidents occurred on the nation's highways.

Modal incident data reported to the HMIS database can be sorted into various hazard classes. There are nine broad hazard classes defined in the hazardous materials regulations. They include several categories of explosives; flammable, nonflammable, and poisonous gases; flammable and combustible liquid; flammable, spontaneously combustible, and dangerous when wet material; oxidizers and organic peroxide; poisonous materials and infectious substances (see box); radioactive material; corrosive material; and miscellaneous hazardous materials. The 9 classes are further disaggregated into 22 divisions to enhance the usefulness of the data for analytical purposes. These data, along with flow/exposure data provided by the Commodity Flow Survey,1 can assist in evaluating the risk of transporting hazardous materials.

For air, highway, and rail modes, two categories of materials constitute the great majority of incidents: flammable combustible liquids and corrosive materials (table 2). This is true for water shipments, as well, but HMIS data for this mode are sparse, because only those water incidents involving packaged hazardous materials, generally in nonbulk packages, are subject to the HMIS incident reporting. Most water spills are reported to the U.S. Coast Guard (see the Energy and the Environment chapter).

1 The Commodity Flow Survey (CFS) is conducted every five years by the Bureau of Transportation Statistics in partnership with the U.S. Census Bureau as part of the Economic Census. Data covering 1993 and 1997 are available at Data and information from the 2002 CFS will be available in late 2003.

1. U.S. Department of Transportation, Hazardous Materials Information System database, available at, as of July 19, 2001.

Worker Fatalities

Occupational risk from transportation incidents is often overlooked in safety analyses. However, these incidents are the largest single cause of occupational fatalities. Of the 5,915 occupational fatalities for all workers in 2000, 43 percent of them occurred as a result of a transportation incident (table 1). Among transportation occupational fatalities, 74 percent resulted from transportation incidents (table 2).

While the number of occupation fatalities for all workers in the United States fell 5 percent between 1992 and 2000, transportation causes rose 4 percent. Much of this increase may be due to highway incidents, which increased 18 percent during the period, and accounted for over half of the transportation-related occupational deaths in 2000. Just over 14 percent of all occupational fatalities in 2000 were truck drivers and 706 (83 percent) of them died in transportation incidents. Since the number of airplane accidents can vary greatly from year to year, so too do occupational deaths attributed to air transportation. The 280 deaths in 2000 were 21 percent below those in 1992, but 26 percent higher than the 223 deaths in 1998 [1].

The risk of being killed while working in a transportation occupation is more than five times the average for all occupations. Among transportation occupations, airplane pilots and navigators and taxicab drivers were at highest risk. Bus drivers had the lowest risk of being killed on the job. The risk of work-related fatalities for all occupations, on average, was 45 fatalities for every million employees in 2000 (table 3). In contrast, the risk for transportation occupations was 255 fatalities per million employees.

Between 1993 (the first year for which data are available) and 2000, the national average risk of work-related fatalities decreased 19 percent. For transportation occupations as a whole, however, it decreased only 4 percent, though the risk for taxicab drivers, rail transportation occupations, and water transportation occupations decreased appreciably. In 1993, the risk for taxicab drivers was 1,658 fatalities per million employees, the highest among all transportation occupations. By 2000, it fell to 538, much lower than that for airplane pilots and navigators.

See box for Transportation Worker Fatalities and September 2001.

1. U.S. Department of Labor, Bureau of Labor Statistics, Census of Fatal Occupational Injuries, 1992, 1993, 1998, and 2000.