2. LITERATURE SEARCH AND DATA GATHERING

2.1 Overview

The literature search for this project focused on gathering data and sources related to brake condition, maintenance, training, and certification of technicians and inspection of brake systems by both technicians and drivers, as contributing factors related to CMV brake problems including crashes and OOS conditions. The literature search was also designed to review books, documents, and data on the need for driver/technician brake training, specific training programs, and techniques for economic analysis. The information provided in this report has been organized into four sections, based on each information source's primary focus. Topical areas covered in the literature search included the following:

• Data and studies on crashes and OOS violations that could be prevented or mitigated,
• The need for better driver/brake technician training and inspection,
• Existing brake training and certification programs, and
• Economic analysis techniques.

Each topical area was examined to uncover those sources that were directly applicable to the brake project. These references were then summarized and analyzed to evaluate their usefulness in assessing the status of brake-related crashes; OOS violation rates; training programs for drivers performing brake inspections, repair, and maintenance; and their suitability for use in the benefit-cost analysis. Finally, the references were examined for those most applicable for benefit-cost analysis of the brake project related cost data.

Two key uses of the material surveyed in the literature review were to:

1. Form the basis for estimating the number of crashes that could potentially be prevented through improved brake maintenance and inspection practices reflecting the implementation of the NTSB recommendations described in Section 1.1 of this report.
2. Estimate the number of OOS violations that could be prevented by implementation of the enhanced maintenance and inspection practices

The most important references used in the preparation of the project analyses are described in this section. Other sources of data and information are summarized and included in the annotated bibliography found in Appendix B.

The project's methodology to identify the percentage of truck crashes that could be prevented or mitigated by improved brake inspection practices was to examine each of the relevant sources and select the most effective CMV crash and inspection data for determining which crashes could be prevented or mitigated and which negative brake inspection results could be avoided from (1) improved inspection practices and training for CMV drivers and (2) improved training and a certification system for those maintaining CMV brake systems. Some of these data would also feed into the calculation of the fraction of vehicles that experienced a certain type of brake-related crash that could be prevented or mitigated through the implementation of an enhanced brake system. Not only were the type of brake problems associated with a crash valuable, information concerning the likelihood that the crash was caused by the brake problem was vital. That is, where the crash type involving a CMV could identify the crash as having been brake related or not brake related. For example, if a CMV that was experiencing brake problems were hit by another vehicle from behind, it is unlikely that the brake problems would have contributed to the crash.

2.2 Estimating Prevention and Mitigation of Crashes and OOS Violations

During the literature review, several databases were examined in detail to estimate the number of CMVs where crashes would potentially be caused by brake-related problems. These included the Trucks Involved in Fatal Accidents (TIFA), Hazmat Accident, Fatal Accident Complaint Team (FACT), and LTCC databases.

The TIFA data were received from UMTRI for 1993 through 2000 and covered more than 30,000 fatal truck crashes (University of Michigan, 2005). TIFA uses the Fatality Analysis Reporting System (FARS) as the source for identifying the fatal crashes. TIFA reports data included in FARS but then supplements FARS with their own data that are gathered from police accident reports and telephone calls to carriers, police, and other involved parties such as tow truck drivers. In most cases, these data provide additional detail for the user. For example, FARS provides data on collision type and although "rear-end crash" is one type of collision, there is no means to differentiate between a crash where the truck crashed into another vehicle or another vehicle crashed into the truck. In the TIFA data, the inclusion of "crash type" variables enables the identification of those trucks that struck the rear of another vehicle.

These data can be used to help determine the type of truck crashes that could have been avoided if brakes were functioning properly. Clearly, these data must be used in conjunction with other data such as FMCSA's Motor Carrier Management Information System (MCMIS) inspection data that show the percentage of trucks on the road with serious brake-malfunction problems.

TIFA data also include an identification of truck crashes known to be associated with various mechanical defects. These include brakes but unfortunately due to sporadic and inconsistent inspections at crash scenes, the data were not very thorough. For the purposes of this brake project, the utility of the TIFA data demonstrated that brake problems as a contributor to fatal truck crashes were strongly linked to accident type. That is, if the truck crashed into the rear of another vehicle, identified brake problems were more likely to be a causative factor in the crash than if another vehicle crashed into the back of the truck. However, the utility of these data for the brake project was somewhat limited because the post-crash inspections lacked the consistency required.

Another source of crash data is the FACT file. The FACT file was analyzed by UMTRI and investigated by the project team as a possible source of data (Blower, 2002). The program investigated every fatal accident in Michigan since 1996 that involved a commercial motor vehicle and at least one death. The investigation includes analysis of the driver, motor carrier, accident scene, sequence of events, role of the truck, and an inspection to determine the compliance of the driver and vehicle with motor vehicle regulations.

At the time of Blower's analysis of the FACT data, 442 trucks were included in the analysis representing crashes from 1996 to February 2001. Of these, data from a North American Standard (NAS) Level 1 inspection were available for 354 of them. A violation was found for either the truck or the driver in 65.8 percent of these inspections. Over 34 percent (121) of the vehicles had at least one brake violation. Mechanical defects including those related to brakes and other vehicle components were common. Fifty-five percent of the trucks had at least one mechanical violation and 28.5 percent had at least one OOS condition.

This rate of violations and OOS for all causes is consistent with trucks routinely inspected under Michigan's Motor Carrier Safety Assistance Program (MCSAP) program. In the three-year period ending in June 2001, 21,322 trucks were subject to a Level 1 inspection. Of these, 31.8 percent were placed out of service.

The FACT data show that accident type (i.e., crash type) combined with the presence of brake problems can indicate the likelihood that a particular crash was caused by brake problems. For example, in a rear-end accident in which a CMV crashes into the rear or side of another vehicle and a brake problem already exists in the CMV, there is a good chance that the brake problem contributed to the crash. On the other hand, when another vehicle crashes into the rear of a CMV, even if the truck's brakes have an identified mechanical flaw, there is little likelihood that the CMV brake problem contributed to the crash since even brakes in perfect condition would not have helped to avoid the crash. Fifty-nine of the trucks were involved in rear-end collisions. Of these, 22 (37.3 percent) had a brake-related violation. This analysis indicates that 5.0 percent (22 of 442) of trucks involved in fatal crashes had brake-related problems that may have contributed to the crash. Note that this is close to the violation rate for brakes for all FACT trucks. For those trucks that struck another vehicle, the overall brake violation rate was 50 percent. Only 27.3 percent of the trucks struck in the rear had a brake-related violation. As expected, a far higher percentage of trucks that struck another vehicle from behind had brake problems when compared with those trucks that were struck from behind. Those truck crashes with demonstrated truck brake problems and where the truck struck another vehicle from behind have a high likelihood that faulty brakes were a contributing factor. Attempts to obtain raw FACT data from the Michigan State Police were unsuccessful. However, even if the data could have been obtained, their applicability for this project was limited because of the small population size, only involving fatal accidents within one State.

The project team also investigated utilizing Battelle's Hazmat Accident Database that was created by a Battelle team under a project for FMCSA (Battelle, 2005). The project's purpose was to obtain additional data and information on hazardous materials (hazmat) crashes that could be used to provide added information for safety analyses. The database uses the MCMIS crash file for 2002 and selects all trucks involved in hazmat crashes. From the approximately 2,000 crashes, half were selected to be enhanced through use of Police Accident Reports (PAR) and contact with motor carriers. The data for each of the approximately 1,000 hazmat crashes for which additional data were collected provide such information as accident type and cause. These data could be used to help identify those hazmat trucks that could have crashed because of a brake malfunction. In-depth data analysis showed that the identification of brake-related problems, although providing some valuable information, was insufficient for the brake project because inspections were often not thorough enough. Furthermore, the population surveyed included only hazardous material trucks and the project team believed that they would not be representative of all CMVs.

The literature survey data also included two very important sources of data that produced estimates used to feed into the benefit-cost analysis. These were the FMCSA's Large Truck Crash Causation Study (LTCCS) and the MCMIS OOS inspection data. The LTCCS is a compilation of inspection data gathered through intensive field inspection and data collection at post-accident scenes (FMCSA, 2005). There are currently data for 1,070 truck crashes (includes some data collected for trucks during a pilot study). This study has produced a considerable quantity of valuable data for people interested in large truck crashes. The LTCCS is a three-year study conducted by the National Highway Traffic Safety Administration (NHTSA) and FMCSA. The study was conducted within the National Automotive Sampling System (NASS), a nationally representative sample involving the General Estimation System (GES). GES is an annual representation of crashes from 60 sites - called Primary Sampling Units (PSUs) -throughout the United States. NASS was chosen because it provides nationally representative data chosen from police accident reports. The data were collected through response by NASS truck researchers and CVSA Level 1-certified State truck inspectors. Inspections were conducted on large trucks involved in the selected crashes.

The LTCCS data files include data for four types of truck brake problems related to a particular crash. These four types are brake failure, brake inoperable, brakes OOA, and brakes deficient. The LTCCS show that brake problems exist in about 30 percent (321 out of 1,070) of the inspections conducted for the study. The brake problems are described in more detail in Section 5.2.

The LTCCS includes detailed descriptions of accident type. These data enabled the project team to correlate those CMVs with identified brake problems with the accident types that could potentially be caused by brake problems. Therefore, the data facilitated the estimation of the percentage of crashes that could have been avoided or mitigated if brake malfunctions were prevented through more efficient inspection and preventive maintenance programs. The Battelle Team concluded that the percentage of trucks with brake violations would best be determined from LTCCS data. Similarly, the accident types that had associated brake problems and would have been avoided or mitigated with an improved brake program were also derived from the LTCCS data. Analysis of the data indicated that approximately seven percent of all CMV crashes could either be avoided or mitigated if the enhanced brake program recommended by NTSB (see Section 1.1) were in place. Section 5 of this report explains how the estimated number of crashes that would be avoided or mitigated was calculated.

The FMCSA Inspection File contains data from State and Federal inspection actions involving motor carriers, shippers of hazardous materials, and transporters of hazardous materials operating in the United States. State personnel under the Motor Carrier Safety Assistance Program (MCSAP) conducted the majority of the inspections at the roadside. Federal and State field enforcement staff perform inspections on interstate and intrastate motor carriers and shippers and transporters of hazardous materials. Violations of the Federal Motor Carrier Safety Regulations may result in a vehicle and/or driver being placed OOS. Inspectors use CVSA-developed criteria and CMVs are placed OOS if a defect or deficiency is so severe that it presents an imminent hazard. The data collected from inspection activity are collected and stored in the FMCSA MCMIS Inspection Data File. The file also contains a limited quantity of "post-crash" inspection data derived from selective inspections at crash scenes. The roadside inspection data also include the type of OOS violation attributed to a particular CMV, including brake problems. Therefore, the data could be used as an effective baseline for estimating the number of brake problems that could be prevented if the enhanced brake training, certification, and inspection programs were implemented. The OOS data for brake violations in the FMCSA Inspection File also provide a reliable baseline for estimating the number of OOS violations that would be avoided once a new program is implemented. Therefore, this data feeds directly into the benefit-cost analysis.

The number of truck crashes was obtained from the 2002 data included in the 2004 edition of FMCSA's Large Truck Crash Facts (FMCSA, 2004). The data in this reference have been derived from several data sources including FARS, GES, MCMIS Crash File, and Highway Statistics. These data were used to provide a baseline for estimating the percentage of annual CMV crashes that might have been prevented if more effective brake maintenance, training, certification, and inspection procedures were in place. The Large Truck Crash Facts shows that approximately 416,000 crashes occurred in 2002.

OOS brake violation statistics were essential for the brake project in order to be able to estimate the number of OOS violations that could be avoided by the implementation of the more effective brake training, certification, and inspection program. Operation Air Brake provides a detailed summary of data for the years 1998 to 2004 that tabulates the number of OOS brake violations found in inspections of a sample of CMVs (CVSA, 2004). Both the CVSA and the Canadian Council of Motor Transport Administrators sponsor the annual Operation Air Brake campaign, which collects data in both countries. The program emphasizes reducing the number of brake-related violations. During 1998 to 2004, Operation Air Brake checked 66,941 units. Of these, 11,360 (16.97 percent) have had an OOS brake condition. Seventy-six motor coaches were checked in 2004. Of these, 20 (26 percent) had an OOS brake condition. These percentages of OOS brake violations were not utilized to estimate OOS violations that could potentially be avoided with improved training, certification, and inspection because the smaller, targeted and intensive sampling resulted in a violation rate considerably higher than that in the FMCSA Inspection File (FMCSA, 2005a), which shows a brake problem-related OOS rate of 9.8 percent. The purpose and implementation of the Operation Air Brake program differs systematically from the regular roadside inspection program, thus, the results of the two types of inspections are not directly comparable. Section 5 of this report includes additional analysis concerning OOS brake violations.

2.3 Need for Better Driver/Brake Technician Training and Inspection

The following subsection describes data sources, papers, reports, and books that were consulted during Battelle's analysis of driver/technician training protocols for the preparation of the brake project to describe the need for enhanced driver/technician training, certification, and inspection programs.

A survey was conducted in 2003 as part of Operation Air Brake, which is intended to reduce the number of brake-related violations, and the results are described in Report of Driver Survey: Air Brake Adjustment Knowledge (VanderZwaag, 2003). The survey results emphasize the gap that exists between the amount of knowledge drivers should have about air brake adjustment and their actual knowledge. The 4,055 driver surveys collected indicated widespread misunderstanding of the importance of brake adjustment and the correct method of brake adjustment for their vehicles. The survey asked drivers to identify the four conditions necessary to properly inspect brake adjustment.^2-1 Fewer than five percent (192 of 4,055 drivers) were able to correctly identify all four of the conditions. The survey results indicate that professional drivers have a poor understanding of the fundamentals of brake adjustment and how to inspect brakes to ensure that brakes are in proper adjustment. Driver lack of knowledge concerning air brake adjustment concepts and inspection would seem to contribute to brake malfunctions during truck operation. While the VanderZwaag report cannot be used to estimate the number or percentage of crashes attributable to correctable brake defects, it does underscore the need for improved driver training in the area of brake inspection and maintenance.

Research results have clearly demonstrated that improved safety practices result in reduced crashes for trucking companies. A 1992 paper investigated crash results for 233 trucking firms given a safety rating of 'unsatisfactory' in earlier Federal inspections (Moses and Savage, 1992). Faced with re-inspection, these 233 firms tended to improve their safety practices and consequently operated more safely. The estimated effect of the second inspection for these carriers was to reduce their overall crash rate by 10.4 percent. It is clear that enhanced brake inspections would likely result in measurable safety improvements. However, since the inspections only affected a small portion of the industry, the improvement in safety could not be quantified in terms of a lower number of crashes or lower crash rates for the total truck population. Furthermore, the research is somewhat dated because the compliance review process has changed. In addition, unfit carriers are shut down if they do not improve their safety performance within a limited time period (45 days for passenger and hazmat carriers; 60 days for non-hazmat property carriers).

The impact of two FMCSA safety programs on CMV safety reinforces the conventional wisdom that the implementation of vehicle inspection and safety compliance programs can result in safety improvements and prevent both fatalities and injuries. The report on FMCSA safety program effectiveness documents the methodology and results from an improved model to measure the effectiveness of two safety programs sponsored by FMCSA (Volpe, 2004). These programs, roadside inspections and on-site compliance reviews were evaluated with respect to their impact on reducing crashes, injuries, and fatalities. The improved model was applied to three years of inspections and reviews—2001, 2002, and 2003—and calculated the reduction in crashes, injuries, and fatalities by State for each year. Results indicated that the programs resulted in significant reductions in crashes, fatalities, and injuries. For example, in 2003, the State roadside inspection program alone resulted in reducing crashes by more than 12,600, and prevented more than 9,600 injuries and 530 fatalities. This report showed that, overall, increased inspections resulted in improved safety. These results are likely to apply to increased inspections that would be implemented if the NTSB brake-related recommendations were followed.

2.4 Training and Certification Programs

The following subsection describes references that were used during Battelle's analysis of training programs and materials.

One important source for training programs is the manual Practical Airbrakes (VanderZwaag 2001). This is a comprehensive manual designed to give drivers the knowledge and understanding needed to operate airbrakes in a CMV. The manual includes overall descriptions of the braking system as well as details about the major subsystem components such as the supply subsystem, the service brake subsystem, and the trailer brake subsystem. Separate chapters are included on inspecting airbrake system components and testing the airbrake system. A highly valuable chapter describes inspecting airbrake adjustment. The project team used the manual as a useful reference when considering the cost and effort associated with developing brake training materials.

The Technician Certification Program Guidelines was another source used to provide recommended practice for the maintenance shop manager, training manager, and/or human resources manager for use in planning, designing, and implementing a technician development program that encourages professional development through technical certification (TMC/ATA, 2004). A certification program may be developed internally or may be based on existing national certification or local licensing programs. The paper recommends that the components described in the guideline should be considered during the development of such a program. The primary goal of a motor carrier technician certification program is professional development of the technical skills and knowledge required for competent performance of the motor carrier technician.

If adopted by the industry, the Battelle Team believes the guidelines should result in improved preventive maintenance and repair programs that, because they include brakes, could help to reduce the number of crashes resulting from brake malfunctions.

The Air Brake Handbook proved to be another valuable source of information for developing brake-related training programs (Bendix, 2000). The handbook provides instruction and data sheets on most air brake system components. A valuable section of the handbook discusses the "fundamentals of brakes" and includes discussion of the physics of braking, how various components of air brakes operate, and a section on brake maintenance. The handbook would be a valuable resource for providing background information for inspectors and drivers and, for technicians, as a guide for troubleshooting problems. The brake maintenance section lists specific precautions designed to protect air brakes that should be followed when working around a vehicle.

Finally, another valuable resource was a report^2-2 that documented the findings of a peer exchange team that examined best practices for CMV safety (U.S. DOT, 1998). The report reflects these best practices in seven States and several trucking companies. The report contains some recommendations that are relevant to the brake project. For example, in the section on handbooks and manuals, five major recommendations are made as to how these sources should be designed, prepared, and distributed. Some of these recommendations are including publications written in easily understood language focusing on the most relevant materials for the driver, ensuring the material's design provides a well organized and useable format, providing for a mechanism that publications reach all who need them, and using the Internet as an alternate to supplying printed materials. A section on training provides descriptions of the diverse training programs available in the various States surveyed. Some of the information was found to be dated, and it would be valuable to update this peer-exchange reference guide.

2.5 Economic Analysis

The following subsection briefly describes the three references that were most important for the economic analysis of the cost and benefits of the proposed brake training, certification, and inspection program. These cost data were used to estimate the savings achieved from avoiding or mitigating crashes. Although the data provide a reliable overview of crash costs, they lack the specificity to differentiate costs related to certain crash types such as cases in which a truck strikes another vehicle or the truck is struck by a vehicle from behind.

The report entitled Economic Cost of Motor Vehicle Crashes constitutes one of the major sources of crash cost information in the United States. The report estimated the economic cost of all motor vehicle crashes in the United States in 2000 at $230.6 billion (Blincoe, L. et al, 2000). This study monetized the costs associated with 41,821 fatalities, 5.3 million non-fatal injuries, and 28 million damaged vehicles. The study also included a number of cost elements: (a) productivity losses, (b) property damage, (c) medical costs, (d) rehabilitation costs, (e) travel delay, (f) legal and court costs, (g) emergency services, (h) insurance administration costs, and (i) costs to employers. The costs included those associated with both police-reported and unreported crashes. The crash costs are stratified by severity according to the Abbreviated Injury Scale. This study examines crash costs associated with all vehicles, including both automobiles and heavy trucks. The average crash cost when all vehicles are included is $14,102 (2002 dollars).

The report entitled Revised Costs of Large Truck- and Bus-Involved Crashes, prepared by the Pacific Research Institute for FMCSA, documents the costs associated with large truck- and bus-involved crashes and provides a good source for estimating the comprehensive costs of truck and bus crashes (FMCSA, 2002). More specifically, the report estimates the dollar value cost per crash, victim injured, or fatality incurred because of large truck- and bus-involved crashes. The cost elements examined in the study include the following: medical-related (e.g., hospital, rehabilitation), emergency services, property damage, lost productivity (e.g., crash investigation, lost wages, recruiting and training replacement workers), and lost quality of life (e.g., pain and suffering). The costs associated with crashes are differentiated based on crash severity, ranging from no injury to fatality, and by vehicle class, including straight trucks, truck-tractor combinations, and buses. The cost of police-reported crashes involving large trucks with a gross weight rating in excess of 10,000 pounds averaged $59,153 (2000 dollars). The cost of police-reported crashes involving buses averaged $32,548. These costs are reported from a societal perspective.

The report entitled Economic Burden of Traffic Crashes on Employers found that traffic crashes occurring on the job resulted in 2,100 fatalities and 353,000 injuries annually during the 1998-2000 timeframe (NHTSA 2003). Further, job-related vehicle crashes accounted for nearly 6.5 percent of all crash injuries. The study estimated the economic burden of traffic crashes on employers, including health-related fringe benefits costs, employer health care costs, sick leave, life and disability insurance for employees involved in crashes, and wage-risk premiums. These cost elements amounted to nearly $60 billion in annual costs to employers in the United States during the 1998-2000 time horizon. This study estimated costs to employers by State. The costs to employers in California and New York were highest, topping $3.5 billion in each State. It also examined the costs associated with crashes involving an alcohol-impaired driver and passengers that are not using restraining devices. Finally, the study broke down cost by industry, estimating the highest costs in the land transportation, construction, mining, and agriculture sectors. The average cost per crash to employers was estimated at $16,471 in direct costs (2000 dollars) and $24,536 in total costs, including the wage-risk premium (a premium paid in the workplace to compensate employees accepting risk due to potentially dangerous working conditions).

From these reports, the Pacific Institute study was deemed most relevant because it comprehensively examines a broad range of costs associated with crashes involving buses and trucks weighing in excess of 10,000 pounds. The Pacific Institute report examines costs from a societal perspective. For example, this study includes costs associated with emergency response and the decline in quality of life experienced by drivers and their dependents following crashes involving serious injuries and fatalities. Both of these cost items are not borne directly by industry. Therefore, these costs can be used to examine the economic value of the brake inspection program on society but cannot be used to determine how the program would affect the industry's bottom line.