Toll Facility Safety Study Report to Congress
3. Findings - Characteristics of Accidents and Injuries Occurring at Toll Plazas
To address the first objective of this effort, which was to study the incidence of accidents and injuries occurring in the vicinity of highway toll collection facilities, the team gathered and analyzed available data on accidents and injuries occurring in the vicinity of toll plazas. This section presents the findings of the accident and injury data analysis.
The study team obtained data on 10,322 vehicular crashes representing 7 agencies, and data on 2,662 worker injuries representing 15 agencies. A summary of the records for both vehicular crashes and worker injuries and the time periods for which data was made available to the study team is shown in Tables 3-1 and 3-2 below. The amount of data provided by each agency for vehicular crashes varied significantly, with one agency providing as few as 15 records and another providing as many as 5,114 records; as a result, it is important to note that data from only two agencies comprises 88 percent of the vehicular crash records in the database.
The findings of the accident and injury data analysis are presented in the following sections.
|Agency||Period Covered||Number of Records|
|Agency 2||January 2000 - July 2006||646|
|Agency 3||January 2005 - July 2006||12|
|Agency 4||January 2002 - December 2006||4|
|Agency 6||February 2002 - December 2006||59|
|Agency 7||January 2002 - February 2005||80|
|Agency 8||July 2001 - September 2006||29|
|Agency 9||January 2004 - December 2006||87|
|Agency 10||January 2005 - September 2006||183|
|Agency 11||January 2006 - November 2006||15|
|Agency 12||January 2001 - November 2006||191|
|Agency 13||January 1996 - October 2006||139|
|Agency 14||January 2001 - December 2003||615|
|Agency 15||January 1975 - December 1990||9|
|Agency 16||January 1997 - December 2006||399|
|Agency 17||January 2000 - November 2006||194|
|Agency||Period Covered||Number of Records|
|Agency 1||January 2005 - December 2006||665|
|Agency 2||June 2001 - May 2006||3,955|
|Agency 3||April 2002 - August 2006||16|
|Agency 4||January 2003 - December 2006||73|
|Agency 5||September 1994 - September 2006||484|
|Agency 6||January 2006 - November 2006||15|
|Agency 7||January 2001 - December 2006||5,114|
3.1 Worker Injuries
The worker injury data provided to the team generally included fields for location, date, time, and, in most cases, a general description that allowed the study team to categorize the data. The descriptions were useful in determining the cause of the injury, the specific resulting injury, and the resulting body parts that were injured.
Some agencies' data provided information on the time of day that the injury occurred. Table 3-3 suggests that, for these agencies, injuries increase in the morning rush hour period and remain higher throughout the midday and afternoon before dropping off again in the late evening. This is also likely the portion of the day when the highest numbers of toll collectors are working and when traffic volumes are greatest.
|Hour||Number of Injuries||% of Injuries|
|12:00 AM - 1:00 AM||15||2.20%|
|1:00 AM - 2:00 AM||6||0.90%|
|2:00 AM - 3:00 AM||11||1.60%|
|3:00 AM - 4:00 AM||12||1.80%|
|4:00 AM - 5:00 AM||11||1.60%|
|5:00 AM - 6:00 AM||21||3.10%|
|6:00 AM - 7:00 AM||46||6.90%|
|7:00 AM - 8:00 AM||44||6.60%|
|8:00 AM - 9:00 AM||40||6.00%|
|9:00 AM - 10:00 AM||38||5.70%|
|10:00 AM - 11:00 AM||50||7.50%|
|11:00 AM - 12:00 PM||42||6.30%|
|12:00 PM - 1:00 PM||25||3.70%|
|1:00 PM - 2:00 PM||40||6.00%|
|2:00 PM - 3:00 PM||36||5.40%|
|3:00 PM - 4:00 PM||24||3.60%|
|4:00 PM - 5:00 PM||24||3.60%|
|5:00 PM - 6:00 PM||33||4.90%|
|6:00 PM - 7:00 PM||41||6.10%|
|7:00 PM - 8:00 PM||21||3.10%|
|8:00 PM - 9:00 PM||19||2.80%|
|9:00 PM - 10:00 PM||25||3.70%|
|10:00 PM - 11:00 PM||28||4.20%|
|11:00 PM - 12:00 AM||17||2.50%|
An additional analysis was performed to look at the number of toll collector injuries that involved a vehicle. From the data provided, 224 of 1,931 (12 percent) injuries involved a vehicle, although it is important to note that it appears that the majority of the injuries in the database that involved a vehicle did not actually involve direct contact between the vehicle and the worker.
Another variable that was investigated was the major causes of injuries. These are presented in Table 3-4 below. The data are sorted by percentage and indicate that the most common causes of injuries are general falls, slips, and trips (27.8 percent); being struck by an object (11.1 percent); and pulling, lifting, and pushing an object (9.1 percent).
|Cause||Number of Injuries||% of Injuries|
|Fall/Slip/Trip - General||600||27.80%|
|Struck by Object||240||11.10%|
|Exposure (chemicals, fumes, smoke, weather, etc.)||159||7.40%|
|Hit body part on object||154||7.10%|
|Fall/Slip/Trip - Slick surface (ice, oil, etc.)||120||5.60%|
|Physical interaction with driver||56||2.60%|
|Opening/Closing Toll Booth Door/Window||55||2.50%|
|Fall/Slip/Trip - Construction Area||3||0.10%|
The study team also analyzed injury data by the type of injury as shown in Table 3-5 . Based on the data analyzed, the most common injury types are cuts, scrapes, and abrasions (22.0 percent), strains (17.9 percent), pains (11.0 percent), and sprains (10.7 percent).
|Injury||Number of Injuries||% of Injuries|
The study team also analyzed the data to determine what part of the body is most commonly injured as shown in Table 3-6. The most common body parts injured include the knee (11.9 percent), back (10.8 percent), head (8.3 percent), and hand (8.1 percent).
|Body Part Injured||Number of Injuries||% of Injuries|
Based on an analysis of records for 2,662 worker injuries representing 15 agencies, it can be observed that:
- The most common causes of injuries are general falls, slips, and trips (28 percent); being struck by an object (11 percent); and pulling, lifting, and pushing an object (9 percent).
- The most common injury types are cuts, scrapes, and abrasions (22 percent); strains (18 percent); pains (11 percent); and sprains (11 percent).
- The most common body parts injured include the knee (12 percent), back (11 percent), head (8 percent), and hand (8 percent).
In terms of what else can be seen from the data, it is interesting to note that 12 percent of the workplace injuries were designated as "involving a vehicle," although it appears that the majority of these injuries did not involve direct contact between the vehicle and the worker. Instead they may have involved a motorist pulling the collector's hand while passing through the plaza. This does indicate, however, that the interaction between vehicles and workers plays a critical role when it comes to worker safety.
While it was possible to examine these trends in the injury data, these findings can only be said to be general observations. The data obtained during this study was not broad enough or consistent enough to draw significant industry-wide conclusions or to fully examine trends. Therefore the findings of this analysis cannot necessarily be said to be representative for the Nation's toll facilities as a whole.
3.2 Vehicular Crashes
Crash data collected during the study represented 10,322 vehicular crashes occurring between 1994 and 2006. The majority of the data represents crashes occurring between 2001 and 2006, although one agency provided data going back to 1994. As different agencies provided data for different time periods, it is not possible to discern meaningful yearly trends from the data.
Although the data supplied were not in a standard format, many agencies collected similar data elements. The most common fields included in the crash data were:
- Accident location.
- Date and time.
- Crash type (at a minimum indicating whether the accident caused property damage only, or an injury or fatality).
Data fields that were less common but present in some of the datasets included:
- Weather conditions at the time of the accident.
- Number of vehicles involved.
- Type of vehicles involved.
- Cause of the accident.
- Accident type (e.g., sideswipe, rear-end, etc.).
- Violation committed.
- Number of injuries.
- Number of fatalities.
- Police crash report number.
- General comments.
About half of the crash records included time of day of the crash; an analysis of this by hour is shown in Table 3-7 below. Not surprisingly, the greatest number of crashes in the database occurred between 7:00 AM and 6:00 PM, which is likely the period of highest traffic volume at these facilities.
|Hour||Number of Crashes||% of Injuries|
|12:00 AM - 1:00 AM||81||1.70%|
|1:00 AM - 2:00 AM||45||1.00%|
|2:00 AM - 3:00 AM||47||1.00%|
|3:00 AM - 4:00 AM||45||1.00%|
|4:00 AM - 5:00 AM||53||1.10%|
|5:00 AM - 6:00 AM||60||1.30%|
|6:00 AM - 7:00 AM||152||3.30%|
|7:00 AM - 8:00 AM||267||5.80%|
|8:00 AM - 9:00 AM||325||7.00%|
|9:00 AM - 10:00 AM||268||5.80%|
|10:00 AM - 11:00 AM||218||4.70%|
|11:00 AM - 12:00 PM||283||6.10%|
|12:00 PM - 1:00 PM||235||5.10%|
|1:00 PM - 2:00 PM||272||5.90%|
|2:00 PM - 3:00 PM||299||6.40%|
|3:00 PM - 4:00 PM||307||6.60%|
|4:00 PM - 5:00 PM||340||7.30%|
|5:00 PM - 6:00 PM||350||7.50%|
|6:00 PM - 7:00 PM||288||6.20%|
|7:00 PM - 8:00 PM||178||3.80%|
|8:00 PM - 9:00 PM||134||2.90%|
|9:00 PM - 10:00 PM||144||3.10%|
|10:00 PM - 11:00 PM||129||2.80%|
|11:00 PM - 12:00 AM||116||2.50%|
It was also possible to explore the frequency of crashes based on the number of vehicles involved. Table 3-8 shows that the majority of crashes in the database (75.4 percent) involved 2 vehicles. Only 21.4 percent involved a single vehicle and very few (3.1 percent) involved 3 or more vehicles.
|Number of Vehicles Involved||Number of Crashes||% of Injuries|
Looking at the data for a single agency (Agency 2), it is possible to analyze the locations of crashes with respect to the toll plaza. Of 406 crash records that reported the crash location, approximately half occurred at the plaza itself (212 or 52.2 percent). Of the remaining crashes, 151 (or 37.2 percent) occurred upstream of the plaza, and 43 (or 10.6 percent) occurred downstream of the plaza.
Looking at the data for another single agency (Agency 7) it is possible to gain insight into the types of crashes occurring at their facilities (representing two large plazas). These are summarized in Table 3-9. Nearly all of the crashes were sideswipes (75.6 percent) or rear end crashes (16.4 percent).
|Crash Type||Number of Crashes||% of Crashes|
|Damaged While Parked||1||0.00%|
|Head On (Two Vehicles)||5||0.10%|
|Object Lying in Road||4||0.10%|
|Ran Off Road||2||0.00%|
|Turned Over in Road||3||0.10%|
Based on analysis of 10,322 crash records obtained from 7 toll agencies, it can be seen that multi vehicle accidents are much more common than single vehicle accidents (75 percent of crashes involved 2 vehicles, while only 21 percent involved a single vehicle). In terms of where accidents most often occur in the vicinity of a toll plaza, only one agency provided this level of detail in their record-keeping, but the data from this agency showed that approximately half of all accidents occurring in the vicinity of a toll plaza (52 percent) occur at the plaza itself while 37 percent occur upstream of the plaza, and 11 percent occur downstream of the plaza. In terms of the most common crash types, again only a single agency provided this level of detail; for this agency a majority of the crashes were sideswipes (76 percent), with the next most-common crash-type being rear-end collisions (16 percent).
While the team examined several trends in the data as just presented, the crash data was not broad enough or consistent enough to allow significant industry-wide conclusions to be drawn or to fully examine trends. Therefore, the findings of this analysis cannot be said to be representative for the Nation's toll facilities as a whole.
3.3 Study Limitations and Recommendations for Archiving of Data in the Future
This section presents both limitations of the data analysis portion of the study as well as suggestions for future data archiving of accidents and injuries at toll plazas.
3.3.1 Study Limitations
As a whole, there simply was not enough data available in a consistent format to develop concrete conclusions. While the team examined several trends in the data, the data obtained was not broad enough or consistent enough to allow significant industry-wide conclusions to be drawn or to fully examine trends. Analyses were performed on an agency-by-agency basis for those agencies with strong data collection and archiving, but the findings of these analyses cannot necessarily be said to be representative for the Nation's toll facilities as a whole.
An example of the limitations of the data is that reporting thresholds (i.e., what level of accident or injury severity is reported in their database) varied significantly by agency, so the number of incidents documented varied independent of safety factors or facility characteristics, making it difficult to make comparisons across agencies and facilities to determine contributing factors.
A limitation specific to the crash records is that while the fields in the injury records were typically short descriptions providing enough detail to allow for categorization and comparison across agencies, the crash data were primarily defined by discrete variables, and these variables differed from agency to agency. As a result, comparing data across agencies was not desirable as it would have required interpretation of these discrete data fields which could lead to misinterpretation of the results.
Another issue with the crash records was that the amount of data provided by each agency for vehicular crashes varied significantly, with one agency providing as few as 15 records and with another providing as many as 5,114 records. As a result, data from only two agencies comprises 88 percent of the vehicular crash records in the database, making it impossible to draw industry-wide conclusions.
These are just some of the many challenges that the study team faced as a result of the limited electronic data archiving currently in practice across the country. The team recommends that agencies consider more consistent reporting in the future. This is discussed in further detail in the following section.
3.3.2 Recommendations for Data Archiving Moving Forward
In order to facilitate comparison of data across toll facilities to make industry-wide observations and conclusions in the future, the study team recommends that standardized reporting procedures be implemented for both accident and injury data, and that a centralized database be created and maintained to store and organize this data in a searchable format. This would allow data to be compared across toll facilities to make industry-wide observations and conclusions. The team recommends that OSHA record-keeping requirements and the Model Minimum Uniform Crash Criteria (MMUCC) Guideline be considered in developing any standards.
The MMUCC Guideline was developed through a partnership between the Governors Highway Safety Association, the National Highway Traffic Safety Administration, the Federal Motor Carrier Safety Administration, and FHWA in response to a similar challenge. Nationwide analysis of crash data was being hindered due to the lack of uniformity between and within States. The purpose of the MMUCC is to provide a data set for describing motor vehicle crashes that will generate the information necessary to improve highway safety within each State and nationally. The MMUCC is a voluntary and collaborative effort to generate uniform crash data that are accurate, reliable, and credible for data-driven highway safety decisions.4
For consistency, the study team recommends that standardized crash data reporting procedures for toll facilities, if implemented, should follow the MMUCC guideline. The MMUCC guideline is very extensive and may take time to implement. Therefore, the study team recommends that data collected for a national toll facility crash database for vehicular crashes include the following MMUCC elements as a minimum (MMUCC data references are in parentheses next to each item):
- Location (C5).
- Date (C2).
- Time (C2).
- Crash type (C8).
- Weather (C11).
- Number of vehicles involved (CD2).
- Type of vehicles involved (V7).
- Apparent crash cause (C6).
- Violation committed (P13/P15).
- Number of injuries / fatalities (CD1/CD3/CD4/CD5/CD6).
- Annualized Average Daily Traffic (RL6).
In addition, in order to perform more detailed analyses of what the possible causes of crashes might be, the geometric and roadway/traffic characteristics (i.e., number of toll lanes, volume, lane widths, sign placement) would be required, and would need to be in an accessible, consistent manner. The FHWA is currently working on an initiative, called the Model Minimum Inventory of Roadway Elements (MMIRE), that will identify the importance of roadway inventory and traffic data for safety programs and will define what critical roadway data variables are required in order to take advantage of current and future cutting-edge analytical tools and resources. 5 The study team recommends that toll operators voluntarily participate in MMIRE data collection activities to have this data available for their own safety analyses, as well as national safety analyses.
As for the injury data, although the written descriptions included with the injury data made it easier for the study team to determine the various characteristics of the injury, the study team recommends that similar, consistent data be collected across agencies if national trends and comparisons are of interest. The study team recommends that the following data should be collected as a minimum for all workplace injuries occurring in the vicinity of toll plazas:
- Involvement of a vehicle.
- Body parts affected.
5 For more information about MMIRE, contact the FHWA Office of Safety Research and Development or the FHWA Office of Safety. [ Return to note 5. ]