TRANSIT BUS SAFETY IN THE U.S. AN ANALYSIS USING THE NATIONAL TRANSIT DATABASE Paper

Transcription

1 Perk, Staes, Begley, and Yegidis 1 TRANSIT BUS SAFETY IN THE U.S. AN ANALYSIS USING THE NATIONAL TRANSIT DATABASE Paper Victoria Perk (corresponding author) Senior Research Associate Center for Urban Transportation Research (CUTR) University of South Florida 4202 E. Fowler Avenue, CUT-100 Tampa, FL Phone: Fax: perk@cutr.usf.edu Lisa Staes Program Director Center for Urban Transportation Research (CUTR) University of South Florida 4202 E. Fowler Avenue, CUT-100 Tampa, FL Phone: Fax: staes@cutr.usf.edu Justin Begley Senior Research Associate Center for Urban Transportation Research (CUTR) University of South Florida 4202 E. Fowler Avenue, CUT-100 Tampa, FL Phone: Fax: begley@cutr.usf.edu Roberta Yegidis Affiliated Faculty Center for Urban Transportation Research (CUTR) University of South Florida 4202 E. Fowler Avenue, CUT-100 Tampa, FL Phone: Fax: yegidis@cutr.usf.edu Word Count: 5,648 (body) + 1,500 (6 tables) = 7,148 Submitted for presentation and publication August 1, 2014; re-submitted November 15, 2014

2 Perk, Staes, Begley, and Yegidis 2 ABSTRACT There is growing interest in transit safety research. A close examination of transit safety incidents requires, to some extent, reliance on the National Transit Database (NTD). Due to limited availability and some data limitations, it can be challenging to use NTD data. This study uses a sample of NTD safety data, specifically for the motorbus and demand-response modes, augmented with interviews from the 17 participating transit agencies. Among the key findings are that over 55 percent of the sampled incidents resulted from actions of other vehicle drivers. Another vehicle hitting a stopped transit vehicle was the cause of 21 percent of the sample incidents, resulting in 8 fatalities. Ten percent of the sample incidents involved the transit vehicle colliding with a pedestrian or bicyclist, resulting in 32 fatalities. Seven percent of the sample incidents resulted in five or more injuries on the transit vehicle, and in 25 percent of these incidents, the transit vehicle was stopped. Areas for further study include recommended changes to the NTD safety reporting, and lessons learned or best practices regarding transit agency safety programs, training, work rules, and pedestrian/bike/auto awareness campaigns.

3 Perk, Staes, Begley, and Yegidis 3 INTRODUCTION There is a growing interest in transit safety research. To closely examine transit safety incidents requires, to some extent, a reliance on data from the National Transit Database (NTD). Due to limited availability and some data limitations, it can be challenging to use the NTD data for research. This study uses NTD safety data augmented with transit agency interviews to take a closer look at the types and occurrences of safety incidents, specifically for the directlyoperated motorbus and demand-response modes. The study from which the information in this paper originated began with the Bus Safety Study: A Report to Congress, which analyzed the safety of public transit buses and over-the-road buses (OTRBs) (1). A particular conclusion from that study encouraged additional examination of available safety data for all motorbus and demand-response services. The new study, State of Transit Bus Safety in the US, comprises a comprehensive analysis of bus safety data (2). One objective was to summarize recent safety data for urban motorbus and demand-response providers, which is the topic of this paper. While the NTD does not directly address causality of incidents, the data were used as a starting point to understand the occurrences of different types of safety incidents. NTD safety data for a sample of 17 transit agencies were used summarized and augmented with the results of interviews with the sampled agencies to gain a better understanding of the incidents. DATA The primary source of data for the analysis presented in this paper was National Transit Database (NTD). The NTD was established by the U.S. Congress to be the country s primary source for information and statistics on the transit systems that operate in the U.S. Those agencies receiving grants from the Federal Transit Administration (FTA) under the Urbanized Area Formula Program ( 5307) or the Other than Urbanized Area (Rural) Formula Program ( 5311) are mandated by statute to report data to the NTD. Data from the Safety and Security NTD reports are used in this study to capture relevant information on major transit safety incidents. According to NTD definitions, an incident on a non-rail mode is reported as a Major Safety Incident if it meets at least one of the following thresholds (3): A fatality (30 days or less from the collision and not due to natural causes) An injury requiring immediate medical attention away from the scene Property damage greater than or equal to $25,000 Evacuations due to life safety reasons (imminent danger) The time period of 2008 through 2012 was selected because of the consistency in NTD definitions and reporting requirements during this time. From 2008 to 2012, there were nearly 25,000 major safety and security incidents reported in NTD for all reporters of the directlyoperated motorbus and demand-response modes. Approximately 60 percent of these, or 15,277, were major safety incidents.

4 Perk, Staes, Begley, and Yegidis 4 NTD Data Qualifiers The NTD Major Incident Reporting Form captures most of the important descriptors of an incident. However, current NTD reporting is not intended to capture all the details of incidents, such as causal factors (including, for example, operator errors or vehicle system failures or defects). While the form does allow for a narrative description of each event, there is no standard format in the presentation of the data and, therefore, there is wide variation in the level of detail provided, both among the participating agencies and within each agency over time. Further, the form must be submitted within 30 days of the incident date, so any information collected after the incident is reported to NTD is not available in the database. It is the objective of this study to collect this additional information for some of the sampled incidents, when available from the agencies. It is also important to remember that NTD safety data are self-reported by the agencies and, while the incident forms are checked by NTD analysts to be sure the reported data conform to NTD definitions, there is currently no formal process for the validation of the information. SAMPLE OF 17 TRANSIT AGENCIES Table 1 summarizes NTD motorbus operating statistics for the 17 transit agencies included in the study. The agencies were selected to reflect various scales of operation as well as geographic variation. Data from 2012 (the most recent data available) are shown for all of the agencies except Chatham Area Transit Authority, whose latest available data are from The table shows the range of the number of vehicles operated in maximum service for the agencies included in the study, spanning the 46-bus system at the Greater Bridgeport Transit Authority to the 3,772-bus system in New York.

5 Perk, Staes, Begley, and Yegidis 5 TABLE NTD Directly-Operated Motorbus Information Transit Agency Sample Transit Agency Alameda-Contra Costa Transit District (AC Transit), CA Vehicles Operated in Maximum Service Total Revenue Miles Total Unlinked Passenger Trips ,247,759 53,642,880 Capital Metro, Austin, TX 173 7,509,440 19,010,826 Central Ohio Transit Authority, Columbus, OH 257 9,689,684 18,423,352 Community Transit, Everett, WA 142 5,212,202 6,473,033 Chatham Area Transit Authority, GA 50 2,620,421 3,537,086 Greater Bridgeport Transit Authority, CT Greater Cleveland Regional Transit Authority, OH 46 2,164,688 5,851, ,912,864 38,487,169 Houston Metro, TX ,239,545 47,000,366 Indianapolis Public Transportation Corporation (IndyGO), IN Kansas City Area Transportation Authority, MO 120 6,171,704 9,947, ,794,854 15,988,034 Lane Transit District, Eugene, OR 86 3,127,919 11,465,053 Los Angeles Metro, CA 1,775 70,796, ,437,480 LYNX, Orlando, FL ,487,372 28,184,740 Massachusetts Bay Transportation Authority, Boston, MA ,507, ,340,048 Miami-Dade Transit, FL ,746,788 77,828,274 New York Metropolitan Transportation Authority Suburban Mobility Authority for Regional Transportation (SMART), Detroit, MI 3,772 95,129, ,381, ,299,807 10,423,733 SOURCE: National Transit Database (NTD) data for directly-operated motorbus modes, 2012 (except data for Chatham Area Transit Authority are from 2011).

6 Perk, Staes, Begley, and Yegidis 6 Characteristics of the Sample: Major Safety Incidents Table 2 provides data on major safety incidents, collisions, injuries, and fatalities for the sample of incidents compiled from the 17 transit agencies. There were a total of 3,374 major safety incidents for these agencies over the five-year study period for the directly-operated motorbus and demand-response modes. This sample represents approximately 22 percent of the total for these two modes for all NTD reporting agencies. This sample is large enough to ensure, with 95 percent confidence, that the sample means will represent the population means, within a margin of +/- two percent. Of the 3,374 incidents reported, Table 2 indicates that nearly all (3,287) are collisions. The remaining incidents include fires, hazardous material spills, and weather-related incidents. Injuries and fatalities are summarized for those on the transit vehicle (bus operators and passengers) and for pedestrians, bicyclists, and people waiting or leaving, which includes those individuals at stops, those who are about to board, or those who have just alighted the vehicle. There were only six fatalities on transit vehicles during this time period. There were a total of 33 fatalities to pedestrians, bicyclists, and those waiting at or leaving the vehicle or stop area. TABLE 2 NTD Safety Information for the 17-Agency Sample Year Total All Major Safety Incidents ,374 Collision Incidents ,287 Injuries to Bus Operators and Passengers 968 1, , ,928 Injuries to Pedestrians, Cyclists, and People Waiting or Leaving Fatalities to Bus Operators and Passengers Fatalities to Pedestrians, Cyclists, and People Waiting or Leaving SOURCE: National Transit Database, directly-operated motorbus and demand-response modes. Characteristics of the Sample: Collisions Table 3 presents information on collision types and associated injuries for the sample, based on the NTD collision categories. The majority of collisions can be characterized as other front impact, followed by rear-ended collisions. Rear-ended collisions are the most frequently reported collisions in demand response, followed by angle collisions.

7 Perk, Staes, Begley, and Yegidis 7 TABLE 3 NTD Collision Information for the 17-Agency Sample, NTD Collision Category Number of Collisions Injured Passengers Injured Operators Injured Waiting or Leaving Angle Head-On Other Front Impact 748 1, Rear-Ended Rear-Ending Side Impact Sideswipe Other Total 3,287 4, SOURCE: National Transit Database, directly-operated motorbus and demand-response modes. Characteristics of the Sample: Overall Injuries/Fatalities Table 4 summarizes all injury and fatality categories for this sample. Most injuries occur to the passengers on board the vehicles (59 percent). The next largest category is occupants of other vehicles involved (28 percent). However, most fatalities are associated with the occupants of the other vehicles (36 percent), while approximately nine percent are to bus occupants (operator and passengers), and another eight percent are those waiting/leaving.

8 Perk, Staes, Begley, and Yegidis 8 TABLE 4 NTD Major Safety Incidents for the 17-Agency Sample Injuries and Fatalities by Type Type % of Total Injuries Passenger % People Waiting or Leaving % Operator % Employee % Other Worker % Bicyclist % Pedestrian in Crossing % Pedestrian Not in Crossing % Other Vehicle Occupant % Other % Total 1,462 1,637 1,396 1,593 1, % Fatalities Passenger % People Waiting or Leaving % Operator % Employee % Other Worker % Bicyclist % Pedestrian in Crossing % Pedestrian Not in Crossing % Other Vehicle Occupant % Other % Total % SOURCE: National Transit Database, directly-operated motorbus and demand-response modes.

9 Perk, Staes, Begley, and Yegidis 9 Beyond the NTD Data Fields A major objective of this work was to ascertain contributing factors for incidents, including those relating to the transit vehicle and those relating to human factors, if possible. To accomplish this, it was necessary to go beyond the basic NTD data fields and read through the incident descriptions to try and determine these factors. As mentioned earlier in this paper, the NTD incident description field is open-ended; therefore, there is significant variation in the level of detail provided. An attempt was made to gain further insight from the transit agency interviews, discussed later in this paper. Additional Incident Classification From reading through each of the 3,374 incident descriptions on the NTD major incident forms for the sampled agencies, a refined set of incident classifications not seen before was developed. There are some caveats to this information. First, the incident description field is openended, thus a wide range of information and detail is provided, depending not only on the agency, but also the individual responsible for the form. Some agencies provided very clear and comprehensive descriptions while others were more sparsely described. Given that it is impossible to collect complete information on all incidents reported by the participating agencies, the study team used knowledge and insight regarding transit operations, guidance from the interview subjects, and professional judgment in categorizing the events. Table 5 summarizes these classifications representing the 17-agency sample.

10 Perk, Staes, Begley, and Yegidis 10 TABLE 5 Additional Classification of Major Safety Incidents for the 17-Agency Sample Based on Incident Descriptions, Category Other Vehicle Pulled In Front of (or Into) Transit Vehicle Number of Occurrences % of Total % Other Vehicle Hit Stopped Transit Vehicle % Other Vehicle Ran Stop Sign or Light % Other Vehicle Turned in Front of Transit Vehicle % Transit Vehicle in Collision with Pedestrian or Cyclist % Transit Vehicle Rear-Ended Another Vehicle % Transit Vehicle Hit Other Vehicle % Transit Vehicle Hit Object or Curb % Fire/Smoke % Transit Vehicle Made Left Turn Without Clearance % Other % Transit Vehicle Ran Stop Sign or Light % Collision Not Enough Information to Classify % Mechanical Problem with Evacuation % Involved Road Supervisor Vehicle % Total 3, % Mechanical/Vehicle Problem (not fire/smoke) % Operator Fatigue/Distraction/Medical Event % SOURCE: Based on National Transit Database Major Safety Incident descriptions. Other vehicle pulled into or in front of bus The largest number of incidents occurs when another vehicle pulls in front of or otherwise hits the transit vehicle, representing 24 percent of all incidents. This category is very broad and includes several types of collisions, including rear-ended collisions where the transit vehicle was in motion. Essentially, this category encompasses all other collisions that do not fit

11 Perk, Staes, Begley, and Yegidis 11 into any other category where another vehicle hit the transit vehicle. Some incidents falling into this category included vehicles exiting driveways or streets and subsequently hitting the transit vehicle in the front, middle, or rear sections. Other incidents in this category involved other vehicles changing lanes into the path of the transit vehicle, or other vehicles hitting the transit vehicle traveling in the opposite direction. Other vehicle hit stopped bus Over 20 percent are due to another vehicle colliding with a stopped transit vehicle. This includes when the transit vehicle is rear-ended while stopped (either at a transit stop, traffic light, stop sign, rail crossing, or in traffic congestion). But, it also includes other contact to the transit vehicle when stopped. For some of the agencies in the sample, this percentage (of other vehicles hitting the transit vehicle while stopped) represented a majority of their incidents. Other vehicle ran a traffic light/stop sign Nearly 14 percent of the incidents involved another vehicle running a stop sign or red light. The occurrence of this type of collision did vary for the individual agencies included in the sample, many of which indicated that this type of collision is much more frequent and therefore represents a higher percentage of total incidents at their agencies. An incident from the sample was placed into this category if the NTD description specifically indicated that the collision occurred as the result of another motor vehicle not complying with a traffic control device, such as a traffic signal or stop sign. Many of these types of incidents resulted in side impact collisions with the motorbus. Other vehicle turned in front of bus To be included in this category, the NTD incident description had to clearly indicate that another vehicle turned in front of the transit vehicle. In most cases, this occurs when a motor vehicle believes it can beat the bus, either by turning left in front of it, or by going around it to turn right in front of it. This type of incident occurred in 10 percent of the sample. Transit vehicle in collision with pedestrian/bicyclist In this case, the same definition as the NTD reporting guidelines is used, when a transit vehicle makes contact with a pedestrian/bicyclist. A bus made contact with a pedestrian in a crosswalk in 10 percent of the sample. Transit vehicle rear-ends another vehicle This category represents the NTD category of rear-ending collisions, wherein the transit vehicle makes contact with the rear of the vehicle in front of it. Seven percent of the sample incidents involved the transit vehicle rear-ending another vehicle. Transit vehicle hits another vehicle This includes all instances of the transit vehicle making contact with another vehicle, except for rear-ending collisions, discussed above. In many cases, these incidents are determined to be preventable by the transit agency, although it must be noted that the determination of preventability varies among agencies. These incidents represented just five percent of the sampled incidents.

12 Perk, Staes, Begley, and Yegidis 12 Transit vehicle hits fixed object or curb Matching the NTD definition, this category includes incidents where the transit vehicle makes contact with a curb or other fixed object. Three percent of the sample fit this category. Fire/smoke This category represents incidents that meet the NTD Major Incident threshold for fires, including an evacuation and use of fire suppression. Approximately two percent of the sample involved fires or smoke that met these criteria. Transit vehicle makes left turn without clearance This type of incident involves the transit vehicle turning left without proper clearance from oncoming traffic. Approximately two percent of the sample involved this action. Transit vehicle runs a traffic light or stop sign As should be expected, this type of incident occurs rarely (less than one percent of the sample), but there are cases where the transit vehicle operator is found to have not properly complied with a traffic control device. Mechanical Problem with Evacuation An evacuation when there is imminent danger to life is also a threshold for an NTD Major Incident. If the evacuation is not for a fire or smoke, it is included in this category. Mostly, these incidents are related to fuel leaks. Less than one percent of the sample involved this type of evacuation. Other This classification includes all other incidents in the sample that are not collisions and that do not fit into any other categories. These other incidents comprise one percent of the sample. Examples include tree branches and construction equipment falling on the transit vehicle, objects thrown at the transit vehicle, and cracked glass breaking and injuring passengers, among others. Collision but not enough information to classify In a few cases, it was clear that an incident in the sample was a collision, but there was not enough information to categorize it within one of the other categories. These cases represent only a very small number of the total (less than one percent). For some, but not all, the study team was able to find out more information from the transit agency interviews (discussed later in this paper). Additional information on possible mechanical or vehicle issues Possible mechanical or other vehicle issues (not resulting in fire or smoke) were identified in 33 incidents, representing one percent of the sample. Possible operator issues, including fatigue, distraction, or a medical event were specifically identified in the NTD descriptions of 16 incidents.

13 Perk, Staes, Begley, and Yegidis 13 More on Pedestrian/Bicyclist Incidents The various circumstances leading to collisions with pedestrians and bicyclists deserve additional attention. Of the overall data sample of 3,374 incidents, 336 involved the transit vehicle colliding with a pedestrian or bicyclist (10 percent). Approximately 28 percent of these occurred while the transit vehicle was making a turn (62 cases involved hitting a pedestrian in a crosswalk while making a left turn, with 3 fatalities). Approximately 15 percent occurred when the transit vehicle was making a stop, leaving a stop, or stopped. When approaching or leaving a stop, a common occurrence is to make contact with an individual who is about to board, or has just alighted. There are several incidents where a bicyclist or pedestrian collided with the transit vehicle while it was stopped. In some of these cases, the pedestrian/bicyclist was determined to be intoxicated; overall approximately four percent of the 336 incidents were confirmed to involve an intoxicated pedestrian or bicyclist, according to the NTD incident descriptions. Pedestrians hitting their heads on the transit vehicle s mirrors while the vehicle is stopped, as well as the bus operator making contact with pedestrians with the mirrors while pulling into or leaving a stop, occurred in nine percent of the total pedestrian/bicyclist incidents. There were also several incidents involving pedestrians coming into contact with a transit vehicle s bike racks. While it is logical that the transit vehicle operator is at fault in most of the collisions that involve pedestrians or bicyclists, several of these incidents occurred when the transit vehicle was stopped (as described above), and approximately ten percent of the cases involved unusual, unfortunate events that could not necessarily be foreseen, including people chased into the street (including a shoplifter), bicyclists losing control and falling into the path of the transit vehicle (in some cases even from the sidewalk), pedestrians colliding with bicyclists then falling into the bus path, pedestrians falling off the curb into the bus path, and another vehicle hitting a pedestrian/cyclist who was then pushed into the bus path. Other unusual occurrences involved the following: The operator beginning to move when someone was still removing a bike from the rack. A bicyclist hanging on to the rear of the moving bus then colliding with a parked car. A bicyclist approaching a stopped bus after the operator had honked at him, and kicking the bus until he fell, claiming injury. High-Injury Incidents Part of the data analysis involved a closer examination of incidents that resulted in five or more injuries on the transit vehicle (passengers and operators). Of the total sample of 3,374 incidents, 230 fit this criterion (seven percent). Clearly, there is some correlation between the number of injuries on a vehicle and the number of passengers on board. NTD data do not provide any information about how many passengers were on board a transit vehicle when an incident occurred. However, it is evident from the data that many of these high injury incidents occurred at the larger transit agencies in the sample. Of the 230 high-injury incidents, 214 were collisions with another motor vehicle. Collisions with fixed objects represented only 11 of the 230 incidents and involved a bus jumping a curb or crashing into a building (one occurrence), or hitting other fixed objects such as poles and fire hydrants. The five remaining high-injury incidents involved the transit vehicle being hit by construction equipment or debris, falling scaffolding, or hitting similar non-fixed

14 Perk, Staes, Begley, and Yegidis 14 objects. Table 6 shows how these high-injury incidents are categorized according to the collision types identified in the NTD as well as the additional refined incident classifications. From Table 6, other front impact was the most common type of high-injury incident, occurring in 26 percent of these incidents. The second most common type of high-injury incident, occurring in 22 percent of the cases, was when the transit vehicle was rear-ended. A transit vehicle may be rear-ended while stopped or while moving. When it is rear-ended while stopped at a transit stop, there is a greater chance of more injuries if there are many passengers boarding and alighting at the time of the incident because more people are standing or walking on or very near the vehicle. Collisions can also occur, of course, when a transit vehicle is stopped at an intersection, in response to a traffic control device, or in congested traffic. Overall, 25 percent of the high-injury incidents occurred while the transit vehicle was stopped.

15 Perk, Staes, Begley, and Yegidis 15 TABLE 6 NTD Collision Type and Additional Classification of High Injury* Incidents for the 17-Agency Sample, Based on Incident Descriptions, Category Number of Occurrences % of Total NTD Collision Type Other Front Impact % Rear-Ended % Angle % Side Impact % Head-On % Rear-Ending % Sideswipe % Other % Blank % Total % Additional Classification Other Vehicle Hit Stopped Transit Vehicle % Other Vehicle Pulled In Front of (or Into) Transit Vehicle % Other Vehicle Ran Stop Sign or Light % Other Vehicle Turned in Front of Transit Vehicle % Transit Vehicle Rear-Ended Another Vehicle % Transit Vehicle Hit Object or Curb % Transit Vehicle Hit Other Vehicle % Other % Transit Vehicle in Collision with Pedestrian or Cyclist % Transit Vehicle Ran Stop Sign or Light % Total % SOURCE: Based on National Transit Database Major Safety Incident descriptions. *High injury incidents are defined as those with five or more injuries on the transit vehicle (passengers and operators). DISCUSSIONS WITH PARTICIPATING TRANSIT AGENCIES After reading through all of the 3,374 incident descriptions, the study team flagged several incidents for follow up with the transit agencies. The information gathered from these 17 urban agencies is summarized in the following sections, according to incidents related to vehicle design, vehicle maintenance, and vehicle operations.

16 Perk, Staes, Begley, and Yegidis 16 Incidents Related to Vehicle Design Based on the incident descriptions, there were relatively few incidents in the sample that appear to be related to vehicle design. However, there were several incidents that involved the transit vehicle s mirrors. Specifically, 44 cases involved either the transit vehicle s mirrors contacting another vehicle or its mirrors, another vehicle contacting the transit vehicle s mirrors, or the two vehicle s mirrors coming into contact with each other (represents one percent of the sample). An additional 28 incidents (less than one percent of the sample) occurred when the transit vehicle hit either an object or a pedestrian with its mirrors, and in six cases a pedestrian walked into a transit vehicle s mirrors. It cannot be determined from this information to what extent the placement of the mirrors on the transit vehicles contributed to these incidents; however, two of the agencies that were interviewed specifically acknowledged there were some problems with older style mirrors that extend 15 inches from the bus. Incidents involving the mirrors can cause injuries to the faces and eyes of those hit by broken glass, including passengers, operators, and people at bus stops. Two agencies specifically indicated that they have had issues with older style mirrors that extend 15 inches from the bus. These types of mirrors make contact with other vehicles (as well as poles, shelters, and even pedestrians) more likely. Incidents Related to Vehicle Maintenance A total of 36 incidents were initially flagged from the NTD descriptions as possibly involving a mechanical issue or other vehicle issue (not resulting in fire or smoke). This represents just one percent of the total 3,374 incidents. After agency follow-up, three of these incidents were found to be related instead to the operator driving too fast for the conditions or not following proper procedures for setting the brake and parking the bus. A little more than one-third of the remaining incidents were attributed to brake issues, according to the NTD incident descriptions. Claiming brake failure is common after a collision and it is the general practice of the agencies to follow up with a comprehensive examination of the brakes and other vehicle performance. However, during the agency interviews, only three of twelve incidents could be confirmed to be caused by brake failure. Unfortunately, no additional information could be found by the agencies for the remaining nine. There were four incidents where steering problems were reported in the incident descriptions. From the agency follow-ups, only two of those could be confirmed (no information was available for the other two). While fires or smoke (meeting the thresholds of a major safety incident, discussed previously) represented less than two percent of the total number of incidents, it should be noted that some agencies experienced several of these incidents while others (including some of the larger agencies in the sample) had none at all. Most fires are engine-related or otherwise originate at the rear of the vehicle. No fatalities were associated with the 60 fire/smoke incidents, and only 9 injuries. In a few cases, it was documented that the injuries were due to passengers shoving each other as they evacuated. Most agencies believe their maintenance practices have been successful in preventing incidents due to mechanical failures. Two participating agencies specifically noted that an aging fleet and maintenance issues were impacting safety. Examination of the data did find some examples of mechanical failure as a contributing factor for some incidents, although the number was quite small overall (one percent of the total sample). Still, it is clear that some agencies have significantly higher rates of engine fires and brake problems.

17 Perk, Staes, Begley, and Yegidis 17 Regarding vehicle age, it is not surprising that older vehicles are more susceptible to mechanical problems. However, most agencies indicate that good preventive maintenance addresses these problems before they can cause an accident. According to one agency, problems with vehicle age cause more service interruptions than safety incidents. Incidents Related to Vehicle Operations Based on the NTD incident descriptions for the sample, only 17 incidents appeared to be related to operator distraction, fatigue, or a medical event. It should be noted that it is possible that more incidents are related to fatigue or distraction, in particular, but there is not enough information in the available data to make a conclusive determination. There are many incidents due to general carelessness or inattentiveness. After agency follow-up, one of these 17 incidents was found to be related to such carelessness, as the operator s foot came off the brake when he reached up to change the head sign code (it was described in the NTD as the foot slipping off the brake, which can also possibly indicate a medical condition or fatigue). In five of the remaining cases, the operator lost consciousness. Two of these involved the operator choking and passing out. In another case, the operator became dizzy and passed out, striking a building and injuring six passengers and himself. In one of the instances of the operator falling asleep (which was confirmed by bus video), the time of the occurrence, 4:00 a.m., may have been a factor. Interestingly, in one case, the operator allowed himself to get involved in a road rage incident with a driver of another motor vehicle. Distractions from Personal Electronic Devices or Vehicle Technology Eleven of the 17 agencies cited distraction as an issue. Cell phone use is ubiquitous today, and transit agencies have had to move quickly to address inappropriate use of personal cell phones by operators. Agencies generally have specific rules regarding cell phone use, including a zero-tolerance policy for use while operating the transit vehicle (typically resulting in immediate termination). However, at least one agency acknowledges that enforcement can be difficult. Most policies allow an operator to carry a cell phone with them, but that it must be turned off and out of view (in a personal bag, for example) while driving. When a driver may use their phone varies slightly among agencies; but, in most cases, the phone can only be turned on and used when the transit vehicle is parked and the operator is off of the vehicle (such as on a layover or break). Although, a few agencies just require that the operator is out of the driver s seat (which should indicate that the bus is stopped and properly parked). The technologies now utilized on board transit vehicles (many of which are in the driver s field of vision) are also possible sources of distraction. Most vehicles are now equipped with mobile data terminals (MDTs). While agencies generally do not require an operator to answer a message when the bus is moving, they are still a distraction. One agency is enhancing their MDT messaging to provide information to the operator on whether a message is urgent or whether it can wait until the operator can safely pull over. However, most acknowledge that a certain level of vehicle technology is increasingly necessary, and it is simply something that must be addressed through work rules and training. Technology is not the only potential source of operator distraction. Another type of distraction noted by the agencies is when operators conduct excessive conversations with passengers while driving. While agencies encourage customer-friendliness, and many operators do get to know their regular customers, it is important that operators understand that they should not be having conversations with passengers while driving. At one agency, it was found that

18 Perk, Staes, Begley, and Yegidis 18 people standing up front, forward of the safety line and talking with operators, was a contributing factor in a several incidents. Many agencies find that operational characteristics related to interlined routes and tight schedules create stress for operators that can result in distractions. These agencies work with their planning departments to minimize these issues where possible. Another agency has been looking into other human administrative factors within the operating environment, such as the potential negative impact of operator distraction resulting from the issuance of driver disciplinary notices, or other information that may be received less than favorably, at the beginning of a shift. This seems to be a previously overlooked factor that can be addressed through adjustments in the timing of such procedures, so as to reduce the incidence of operators starting their shifts with a certain level of anger, irritation, worry, or embarrassment that may result in distraction once on the road. Other factors in the operating environment, such as operator interactions with dispatch, supervisors, etc. may also play a role in potential distractions. Operator Fatigue/Medical Issues As with distraction, the participating agencies have varied experiences with operator fatigue. Some consider it a very minor issue, while it has proven to be more significant at other agencies. Overall, 11 of the 17 agencies identified operator fatigue as a concern (not the same 11 that identified issues with distraction). Sometimes fatigue can result from a medical issue such as sleep apnea or other sleep disorders. One agency indicated that some operators are directed to undergo sleep evaluations or tests to determine whether they have sleep apnea or other medical related issues that prevent them from getting adequate rest. In several cases at this agency, operators who were found to have nodded off or who reported their foot slipping off the brake prior to an incident were sent for medical evaluations and sleep assessments. Wellness programs and additional training were noted as methods that can help bring awareness of how one s overall health can impact the level and quality of rest and therefore, the level of fatigue. Another source of fatigue is operators who work split shifts and/or work at another job involving driving. In these cases, the operator may be working the proper number of hours for the transit agency per regulations, but is exceeding allowable driving hours when added to their driving hours for outside employment. Most agencies require operators to inform supervisors of other employment, but it can be difficult to track and enforce. Safety Impacts of Stop Location Stop location was cited as a potential safety problem at 6 of the 17 agencies interviewed. At least one agency has been working over the last few years to relocate stops that have been the source of safety problems. There are both advantages and disadvantages to near-side and farside stops, but several agencies prefer near-side stops to minimize the queuing of vehicles in the intersection. The trade-off is that other vehicles will often try to pass the bus at a near-side stop and turn in front of it, many times resulting in collisions. SUMMARY AND FINDINGS This paper has provided a comprehensive summary of a national sample of major safety incidents (as defined by the NTD and with additional incident classifications developed by the authors for this work) occurring at 17 urban transit agencies from 2008 through 2012, for the directly-operated motorbus and demand-response modes. In addition, results of discussions with

19 Perk, Staes, Begley, and Yegidis 19 participating agencies regarding their specific incidents, as well as more general safety issues have been compiled and presented. Key Findings Highlights from the data summarized in this paper are listed below. From 2008 to 2012, there were 4,928 injuries and 6 fatalities to transit passengers and operators. There were 342 injuries to pedestrians and cyclists with 33 fatalities. Occupants of other vehicles suffer the largest number of fatalities as a result of these major incidents, followed by pedestrians in crosswalks. Overall, fatalities are relatively rare; there are approximately 0.03 fatalities per one million miles of revenue service. Using NTD collision fields, the collision type of other front impact results in the highest number of injuries to passengers, operators, and people waiting at or leaving a stop. The second and third highest number of injuries result from rear-ended and rearending collisions, respectively. Sideswipes result in the least number of these injuries. The analysis in this paper went beyond the NTD data fields to ascertain additional contributing factors of major incidents. To accomplish this, each NTD incident description in the sample was read, and several incidents required follow-up with the participating agencies to further clarify the information found in the NTD. Additional highlights are below. Over 55 percent of the sample incidents resulted from the actions of other vehicle drivers. (Many of these incidents may still have been deemed preventable, depending on the agency s definitions). Approximately 24 percent of the sample incidents are the result of another vehicle pulling into or in front of the transit vehicle. These incidents resulted in 10 total fatalities. Another vehicle hitting a stopped transit vehicle was the cause of 21 percent of the sample incidents, resulting in 8 fatalities. Another vehicle running a traffic light or stop sign caused another 14 percent of the incidents, with 3 fatalities. Ten percent of the sample incidents involved the transit vehicle colliding with a pedestrian or bicyclist, resulting in 32 fatalities (4 waiting or leaving, 8 cyclists, 10 pedestrians in crosswalks, and 10 pedestrians not in crosswalks). Fifteen percent of the collisions with pedestrians or cyclists occurred when the transit vehicle was stopped. In 62 incidents, the transit vehicle hit a pedestrian in a crosswalk while making a left turn. Seven percent of the sample incidents resulted in five or more injuries on the transit vehicle. In 25 percent of these incidents, the transit vehicle was stopped. One percent of the sample incidents could be directly related to vehicle maintenance issues. Out of 12 possible brake problems cited, 3 were confirmed as brake failures. Out of four possible issues with steering, two were confirmed as failures. In addition, there were fuel/fluid leaks, blown tires, and incidents where parts fell off the transit vehicle. Approximately one-half percent of the sample incidents could be attributed to an operator issue such as distraction, fatigue, or a medical event. There were five instances of the operator losing consciousness, and two operators fell asleep while driving. It is possible that more of the sample incidents could be due to distraction, but it cannot be confirmed with the available data. More information on distraction was provided in the agency interviews.

20 Perk, Staes, Begley, and Yegidis 20 Overall, safety incidents directly caused by mechanical problems or maintenance issues do occur, but at very low rates. Operator issues and errors with distraction and fatigue vary among agencies, but most agree that such factors impact safety. Still, a relatively low number of incidents from the total sample could be attributed directly to distraction or fatigue. A number of other incidents were found to be due to operator carelessness or inattentiveness that was determined not to be the result of distraction or fatigue. Contributing factors in these incidents include the bus operator s failure to park or secure the bus properly, not paying attention to turning radius or mirrors, and not accurately judging the speed of other vehicles, as examples. Further study into these factors is needed but will likely require changes in NTD reporting designed to specifically capture this information. There are ways to do so without adding significant reporting burden to the transit agencies. Additional helpful research to further the understanding of transit bus safety should include lessons learned or best practices regarding transit agency safety programs, training, work rules, and pedestrian/bike/auto awareness. ACKNOWLEDGEMENT The research described in this paper was sponsored by the Federal Transit Administration, which provided the NTD data used in this paper. REFERENCES (1) Bus Safety Study: A Report to Congress, prepared for the Federal Transit Administration by the Center for Urban Transportation Research at the University of South Florida, FTA Report No. 0051, November (2) State of Transit Bus Safety in the U.S., prepared for the Federal Transit Administration by the Center for Urban Transportation Research at the University of South Florida, FTA Report No. (TBD), June (3) National Transit Database 2013 Safety and Security Reporting Manual, Office of Budget and Policy, Federal Transit Administration, U.S. Department of Transportation, January 2013.