SP-1c: Investigation of Frequency, Severity and Nature of Bus Accidents, with Comparison of Injuries to Wheelchair-Seated and Able-Bodied Occupants

Task Leader: C. Greg Shaw, Ph.D.

Co-investigators: none

Other Participants: Tom Songer, PhD (epidemiologist), Jim Patrie PhD (statistician), Mike Buckel (transit bus safety engineer), Graduate Student.

Duration/Staging of Task: This is a 24-month research task that will be conducted in the first two years of the 60-month RERC cycle.



Importance of the problem

It is a well-recognized fact of injury biomechanics that, in addition to a belt-type occupant restraint, the vehicle seat is an important component of the occupant-protection system. For people who remain seated in their wheelchair when using a public transit vehicle, the wheelchair must serve as the vehicle seat. While there can be little disagreement that wheelchair users deserve the same level of crash protection as passengers who ride on vehicle seats (DOT 1981), there is considerable disagreement as to what constitutes a similar level of crash protection. In large part, this is because of the large variety of wheelchair types and models, with a wide range of physical characteristics, size, and mass, that may become vehicle seats. It is also because people who use wheelchairs may have different abilities to remain seated without a restraint system, than do able-bodied travelers.

Beyond the issue of offering an equal level of occupant protection to wheelchair users is the issue of offering the opportunity for an equal level of injury risk. The risk of an injury is not only a function of crash and restraint conditions, but of the injury tolerance of the occupant or segment of the occupant population. Although there is a wide range of injury tolerance for people with physical disabilities, just as there is for able-bodied travelers, it is likely that the distribution of injury tolerance is significantly different and generally lower for people who must remain seated in their wheelchair.

The required level of seat and restraint-system performance is a function of expected levels of crash dynamics, which, in turn, are a function of vehicle size and mass, as well as the travel speeds and types of transportation service. In order to determine conditions of wheelchair securement and occupant restraint that provide an equivalent, or enhanced, level of occupant protection and safety for wheelchair-seated travelers in public- and school-bus transportation, information on the frequency, magnitude, and nature of vehicle impacts, and of the frequency and severity of injuries to able-bodied and wheelchair-seated occupants in different types and sizes of vehicles, is needed. However, as noted by Shaw (2000), there is very little such information available, especially for wheelchair passengers, to make informed decisions regarding policies and procedures for safely transporting people in wheelchair, and for setting a reasonable level of occupant protection for these individuals.

As noted in the Introduction to Priority SP1, it has been assumed in developing WTORS standards to date that the manufacturer does not, and cannot, control the type or size of vehicle in which the WTORS equipment is installed. Thus, development of U.S. and ISO standards, have used worst-case frontal crash conditions, similar to those used in federal standards for passenger vehicles, in setting the dynamic performance requirements for WTORS. The wheelchair securement system that is most effective in complying with the 30-mph, 20-G frontal impact requirements set forth in Appendix A of SAE J2249, Wheelchair Tiedown and Occupant Restraint Systems for Use in Motor Vehicles, is the four-point, strap-type tiedown. This is also a system that has tremendous versatility for securing a wide range of wheelchair types and models that have not been designed for securement in a motor vehicle. It is, however, a securement system that is time consuming and difficult to use, especially on wheelchairs that do not have clearly designated and accessible securement points as required by ANSI/RESNA WC/19.

The requirements for WTORS set forth in SAE and ISO standards are clearly unnecessarily high for equipment used in larger transit vehicles and many school buses. Evidence also suggests that ADA wheelchair tiedown strength requirements for transit vehicles (ATBCB 1992) are overly demanding for large buses that operate on city streets (transit buses). A statement recently issued by The Transportation Research Board (TRB) (6/2001) clearly states the problem.

In order to enact provisions mandated by the ADA, the Architectural and Transportation Barrier Compliance Board has issued guidelines, covered by USDOT regulations, concerning the securement of wheelchairs on-board buses. The ADA securement standard was developed based on experience with smaller vehicles with lower mass (e.g. van conversions) and resulted in a complex system involving four separate securement belts plus occupant restraint belts. Although extremely secure in design, its use has raised a number of challenging practical issues for transit agencies and for wheelchair users on-board transit buses. As a result of these issues, some wheelchair users feel uncomfortable and constrained in their use of transit buses, and for some others, safety may be compromised by improper or haphazard use of the securement system.

A similar statement can probably be made for many school transportation situations, where it is likely that the current WTORS standards are higher than necessary for many larger vehicles. Furthermore, unlike the situation for personal licensed vehicles, in many public transit environments, efficiency and independence in securing wheelchairs, in order to meet route schedules and minimize intrusion into a rider’s personal space, are important factors that must be considered, along with safety, when providing transportation to wheelchair occupants.

The primary hypothesis of this research Task is that a better understanding of real-world injury risk and accident exposure and conditions for users of larger transit- and school-bus type vehicles, with a special emphasis on injuries and risks to wheelchair-seated occupants in these environments, will help in making good decisions about the need for improved safety versus the need for improved efficiency and ergonomics in transporting people in wheelchairs. It is also hypothesized that a better understanding of accidents and injuries in larger vehicles will lead to establishing more appropriate performance criteria for WTORS used in these vehicles. This, in turn, should lead to the development and use of wheelchair securement systems that are more compatible with the need for operational efficiency and user independence.

Research Objectives

The objective of this Task is to assemble information from a wide range of sources regarding the nature and frequency of bus accidents and occupant injuries for both wheelchair riders and able-bodied passengers, where different types of restraint and seat/wheelchair securement are used. Although the study will address all bus sizes and types, it will focus on large intra-city public transit buses. This information will be used, along with information from bus crash tests, to make recommendations on performance levels for wheelchair securement in different bus environments. Finally, this Task will examine and report on the appropriateness of current proposals for using very-low-g wheelchair securement systems in selected vehicle and transportation environments.

Anticipated Outcomes

This study will provide a retrospective overview of bus accidents that result in injuries to wheelchair and vehicle-seated passengers. It will provide information about accident severities that will be useful in estimating the risk of injury to wheelchair riders relative to the general user population, and will facilitate decision making relative to the level of crash protection that is required. A level of crash protection for large transit buses will be recommended, and it is anticipated that this will encourage the development of wheelchair tiedown and occupant restraint systems that provide a more appropriate level of crash safety, along with improved ease of use in comparison to present ADA- and SAE-compliant WTORS.

References

ATBCB (1992). Buses, Vans & Systems, Technical Assistance Manual, U.S. Architectural & Transportation Barriers Compliance Board, October 1992.

Booz Allen (1977). Transbus Safety and Human Factors Summary Report. Booz, Allen, and Hamilton, Bethesda, Maryland.

DOT (1981). Wheelchair Securement Systems in Transit Vehicles: A Summary Report. Department of Transportation report prepared by Brenner, E; and Giangrande, RV.

Elias (2001) J, Sullivan L, McCray L. Large School Bus Safety Restraint Evaluation. ESV Conference Paper # 345. June 2001.

Langwieder (1985) K, Danner M, Hummel Th. Collision Types and Characteristics of Bus Accidents- Their Consequences for Bus Passengers and the Accident Opponent. ESV Conference Proceedings pp 585-913 July 1985.

Lawrence (2001) G. Study of Improved Safety for Minibuses by Better Seat and Occupant Retention. ESV Conference Paper # 334. June 2001.

Martin, P (1996) UVA SOWHAT Project: Accident Analysis to Determine Delta-V Distributions for Vans. University of Virginia Transportation Engineering Center unpublished report.

NHTSA (1994) Traffic Safety Facts. Department of Transportation National Highway Traffic Safety Administration 1994.

NTSB (1999) Bus Crashworthiness Issues. Highway Special Investigation Report NTSB/SBIR-99/04 National Transportation Safety Board 1999.

Schneider (1991) L. Rationale, Historical Synopsis of the Literature, and Bibliography, Vol 2. Prepared for the Canadian Standards Association, Document # R91-13.

Shaw (2000). Wheelchair Rider Risk in Motor Vehicles: a Technical Note. VA Journal of Rehabilitation Research and Development Vol. 37 No. 1.

Zeeger (1993) C, Huang H, Stutts J, Rogman E, Hummer J, Fruin J. Characteristics and Solutions Related to Bus Transit Accidents. Southeastern Transportation Center / University of North Carolina Highway Safety Research Center. March 1993


Progress report May 1, 2003

In order to establish a reasonable level of occupant protection for wheelchair riders in large transit buses, we have begun a retrospective investigation of bus accidents and occupant injuries that will involve a comprehensive review of past and present studies, and a survey of a representative sample of U.S. transit providers.

Subtask 2 – Examine and monitor data from previously reported and ongoing studies - was completed and documented as a paper approved for publication in the VA Journal of Rehabilitation Research and Development. Subtask 3 – Search and Review U.S. National Databases – also was completed during this time period. In addition to two U.S. databases (GES and SAMIS), a Canadian database search was completed. Subtask 4- Survey Transit Providers – also was completed after receiving a disappointing response from initially receptive sources. Very good information was received from Washington State and the State of New York Department of Transportation. In March 2003 an additional subtask was added to the study. This six month effort will involve crash reconstruction of bus crashes to determine deceleration levels. Subtask 5 – Analyze and Summarize Information – was begun in April 2003.


Progress report May 1, 2004

In order to establish a reasonable level of occupant protection for wheelchair riders in large transit buses, we conducted a retrospective investigation of bus accidents and occupant injuries. This review involved a comprehensive review of past and present studies, which was documented as a paper published in the VA Journal of Rehabilitation Research and Development. We also searched U.S. and Canadian regional and national databases specifically for bus crashes involving accelerations greater than 1g. An additional study involved reconstruction of bus crashes to determine deceleration levels.


Progress report May 1, 2005

Data extracted from Florida police accident reports were used to identify 1322 Florida transit buses involved in a traffic accident during 2000 and 2001 that resulted in a fatality, an injury, or property damage of $500 or more. Of these, 337, or about 25%, involved an injury to at least one bus passenger. Most injuries to bus passengers were minor and 57.8% of the passengers on the buses were not injured. Collisions in which transit bus passengers were injured were much more likely to occur along the longitudinal axis of the bus, with about 74.2% of transit buses with injured passengers being struck in the front or rear. The most common impact point for buses with seriously injured passengers is the front of the bus, while the most common for buses with minor injuries is the rear.

Of the 337 buses with one or more injured passengers, sufficient data were available to calculate a "peak contact velocity," or PCV, for 241 transit buses by taking into account the masses of the vehicles, relative velocity, and angle of impact. Estimates of PCV were used as a rough estimate of the crash severity, or bus delta V, in mph. Of these 241 cases, the PCV was less than 3 mph in 164, or 68% of the cases, and it was less than 5 mph in 216, or 90% of the cases. A maximum PCV of about 20 mph was estimated for one transit-bus crash.

Data from staged bus crash tests were used to determine a relationship between crash delta V and peak vehicle deceleration measured at the floor of the vehicle. This resulted in a linear regression line, whereby G's=0.48325*PCVkph-0.0302, which was used to estimate the distribution of crash severity by peak g level for the 241 crashes for which delta Vs were estimated. Of the 241 cases with PCV estimates, the estimate of peak g level was less than 1 g in 102 or 42% of the cases, and it was less than 3 g in 203 or 84% of the cases. In one case, the estimated maximum crash deceleration was 15.5 g. The mean estimated peak g in frontal impacts is 2.12, compared with 1.14 for side impacts and 2.60 for rear impacts.

Estimates for vehicle miles traveled (VMT) obtained from several counties in Florida were used to determine the frequency of different crash severities by VMT. Overall, the probability of a being involved in a bus crash with a PCV, or delta V, over 19 mph is once per 500 million miles and over 5 mph is one in 17.5 millions miles. Also, the probability of being involved in a bus crash with a peak deceleration over 10 g is once per 500 million miles, while the probability of being involved in a crash with a peak deceleration over 3 g is one in 27.0 million miles. Work is currently under way to compare these estimates for crash severity by VMT to corresponding numbers for passenger vehicles.

Progress report May 1, 2006

Frequencies of different collision severities by millions of vehicle miles (VMT) traveled for transit buses were compared to those for passenger vehicles based on crashes with injured occupants in the National Automotive Sampling Systems (NASS). A paper on this study was submitted to the 2005 International Truck & Bus Safety & Security Symposium and a presentation was made at the symposium on November 14th in Alexandria, VA.

Of the 242 transit-bus cases, the PCV was less than 3 mph in 164, or 68% of the cases, and it was less than 5 mph in 216, or 90% of the cases. By comparison, only 3% of passenger car crashes resulting in an injury to an occupant of the vehicle are less than 5 mph and 34% are greater than 15 mph delta V. The mean estimated PCVs for transit-bus crashes that resulted in an injury to one or more bus occupants are 2.76 for frontal impacts, 1.51 for side impacts, and 3.28 for rear impacts. A maximum PCV of about 20 mph was estimated for one transit-bus crash. The estimated of peak-g level was less than 1 g in 102 or 42% of the cases, it was less than 3 g in 203 or 84% of the cases, and it was less than 5 g in 225 or 93% of the cases. The mean estimated peak g in frontal impacts is 2.12, compared with 1.14 for side impacts and 2.60 for rear impacts.

Overall, the probability of a being involved in a bus crash with a PCV (or delta V) over 20 mph is once per 455 million vehicle miles, and involvement in a crash over 5-mph delta V is one in 17.5 millions miles. By comparison, the probabilities that an injured occupant in a passenger car is involved in crash with severities over 5 and 20 mph delta V are once every 1.1 millions miles and once every 7.2 million miles, respectively. Using the correspondence between PCV and peak g level, the corresponding frequencies for transit bus involvement in crashes of different severities by peak-g level are once every 3.3 million vehicle miles peak gs of 1 or more, once every 26.8 million miles for peak gs of 5 or more, and once every 455 millions miles for peak gs of 10 or more.


5 year report June 1, 2006

The purpose of this study was to use data from police accident reports of transit-bus accidents in Florida over a two-year period to estimate the distribution and frequency of transit bus collision severities and occupant injuries. The results will be used to guide the development of performance requirements and test methods of new standards for equipment and procedures used to transports wheelchair-seated occupants in fixed-route transit buses. 

Florida police accident reports (PARs) in years 2000 and 2001 document 1322 transit-bus accidents that resulted in a fatality, an injury, or property damage of $500 or more.  Of these, 337 accidents, or about 25%, involved an injury to at least one bus passenger.  Most injuries to bus passengers were minor and 57.8% of the passengers on the buses were not injured. Collisions in which transit bus passengers were injured were much more likely to occur along the longitudinal axis of the bus, with about 74.2% of transit buses with injured passengers being struck in the front or rear.  The most common impact point for buses with injured passengers is the front of the bus, while the most common impact for buses with minor injuries is the rear of the bus.

Of the 337 buses with one or more injured passengers, sufficient data, including vehicle masses, relatively velocity, and angle of impact, were available on the 242 PARs to calculate a “peak contact velocity,” or PCV, which is a rough estimate of crash severity or delta V.  This method for estimating delta V was validated using cases in the National Automotive Sampling System (NASS) for which delta V was also estimated using vehicle crush measurements and reconstruction computer programs such as WinSmash. 

Of the 242 transit-bus cases, the PCV was less than 3 mph in 164, or 68% of the cases, and it was less than 5 mph in 216, or 90% of the cases.  By comparison, only 3% of passenger car crashes resulting in an injury to an occupant of the vehicle are less than 5 mph and 34% are greater than 15 mph delta V.   The mean estimated PCVs for transit-bus crashes that resulted in an injury to one or more bus occupants are 2.76 for frontal impacts, 1.51 for side impacts, and 3.28 for rear impacts.  A maximum PCV of about 20 mph was estimated for one transit-bus crash.

Data from staged crash tests of transit buses were used to determine a relationship between crash delta V and peak vehicle deceleration at the floor of the vehicle.  This resulted in a linear regression of peak-g  = 0.48325*PCVkph-0.0302 which was applied to 241 of the 242 cases for which PCV had been estimated.  The estimated of peak-g level was less than 1 g in 102 or 42% of the cases, it was less than 3 g in 203 or 84% of the cases, and it was less than 5 g in 225 or 93% of the cases. In one case, the estimated maximum crash deceleration was 15.5 g.  The mean estimated peak g in frontal impacts is 2.12, compared with 1.14 for side impacts and 2.60 for rear impacts.

Estimates for vehicle miles traveled (VMT) obtained from several counties in Florida were used to determine the frequency of different crash severities by VMT. Overall, the probability of a being involved in a bus crash with a PCV (or delta V) over 20 mph is once per 455 million vehicle miles, and involvement in a crash over 5-mph delta V is one in 17.5 millions miles.  By comparison, the probabilities that an injured occupant in a passenger car is involved in crash with severities over 5 and 20 mph delta V are once every 1.1 millions miles and once every 7.2 million miles, respectively.  Using the correspondence between PCV and peak level, the corresponding frequencies for transit bus involvement in crashes of different severities by peak-level are once every 3.3 million vehicle miles peak gs of 1 or more, once every 26.8 million miles for peak gs of 5 or more, and once every 455 millions miles for peak gs of 10 or more.

Although involvements in crashes of higher severity are dramatically less frequent for occupants of transit buses compared to occupants of passenger cars, the rates for occupant injury per 100-million passenger miles traveled are not as different. For example, occupants of passenger cars sustain incapacitating (A injury on PAR) about 9 times every 100-million passenger miles and sustain 27 non-incapacitating (B injury on PAR) and 71 minor (C injury on PAR) injuries every 100-million passenger miles.  By comparison, occupants of transit buses sustain about 1.9 incapacitating injuries, 14 non-incapacitating injuries, and 52 minor injuries every 100-million passenger miles. These transit-bus injury rates are only about one-fifth, half, and three-fourth the rates for the rates for passenger vehicles at the three injury levels of incapacitating, non-incapacitating, and minor, respectively.

Last updated: August 18, 2006

Comments