Skip Ribbon Commands
Skip to main content
Bookmark and Share this page


Testimony before the Committee on Transportation and Infrastructure, Subcommittee on Railroads, U.S. House of Representatives Regarding Federal Railroad Administration Reauthorization: Human Factors Issues
Jim Hall
U.S. House of Representatives, Committee on Transportation and Infrastructure, Subcommittee on Railroads, Regarding Federal Railroad Administration Reauthorization: Human Factors Issues, Washington, DC

Good morning Chairman Franks and Members of the Subcommittee. I appreciate the opportunity to represent the National Transportation Safety Board (NTSB) before your Subcommittee to discuss the important matter of railroad safety, particularly the important issue of human factors.

Before I begin my testimony, I would like to introduce Mr. Robert Lauby, Director of the Office of Railroad Safety, and Mr. Barry Sweedler, Director of the Office of Safety Recommendations and Accomplishments.

According to data published by the Federal Railroad Administration (FRA) for Calendar Year 1996, approximately 78 percent of train collisions were caused by human factors, and data for 1995 show a remarkably similar 79 percent of collisions caused by human factors.

In the course of our railroad accident investigations, we have identified a number of human factors issues, such as: training of operating employees, substance abuse, supervisory oversight, and judgment and decision-making issues. But an issue that we encounter with disturbing frequency is human fatigue.

Human Fatigue in Transportation Operations

Human fatigue in transportation operations is probably the most widespread safety issue in the transportation industry. It has been an item on the Safety Board’s "Most Wanted" list of transportation safety improvements since its inception in 1990. All of us are subject to fatigue irrespective of age, gender, or occupation, and the Board has encountered fatigue issues in all modes of transportation.

Fatigue as a Causal Factor

In preparing for today’s testimony, I had occasion to review several of the Safety Board’s railroad accident reports in which the Board determined that fatigue was causal to the accident. Although the accidents varied in their particular circumstances, I want to briefly relate one of them to you as an example of how fatigue erodes rail safety.

About 3:13 a.m. Eastern Daylight Time, on August 9, 1990, northbound Norfolk Southern freight train 188 collided with southbound Norfolk Southern freight train G-38 near Sugar Valley, Georgia. The conductor on train 188 and the conductor and engineer on train G-38 were fatally injured. The trainmen on both trains and the engineer on train 188 received minor injuries.

The investigation determined that southbound train G-38 had been given clear signals to operate on the mainline and on to Sugar Valley. The northbound train 188 had entered the siding, but instead of waiting for train G-38 to pass it proceeded through the siding, beyond a stop signal, and back out to the mainline when the collision took place.

The Safety Board determined that the probable cause of the accident was the failure of the engineer of train 188 to stop at the stop signal because he was asleep, distracted, or inattentive.

While reading the Sugar Valley report, I came across a sentence that my predecessors at the Board had included in their final report issued in July 1991. The sentence reads "The Safety Board is hopeful that the FRA will soon provide guidelines to help the railroad industry reduce or eliminate accidents caused by fatigue."

Mr. Chairman, we are aware that the FRA in 1992 submitted a legislative proposal to Congress that would have amended the Hours of Service Act. However, that proposal was rejected by Congress, and seven years have passed since the Sugar Valley report was adopted. We are pleased that the Administration’s recently-submitted railroad reauthorization bill includes fatigue guidelines. In this regard, I would like to discuss three issues: 1) education and training, 2) irregular and unpredictable schedules, and 3) abrupt changes in rest cycles – and how each is relevant to the railroad accidents investigated by the Safety Board.

Education and Training

Education and training was the focus of the Board’s report of the collision involving two New York City subway trains on the Williamsburg Bridge in Brooklyn, New York, that occurred June 5, 1995.

About 6:12 a.m., a New York City Transit (NYCT) southbound J subway train collided with the rear car of a stopped NYCT M subway train. The collision occurred on the Brooklyn side of the Williamsburg Bridge, and the operator of the J train was fatally injured. Sixty four passengers were treated at area hospitals for serious or minor injuries. The Safety Board determined in part that the probable cause of this accident was the failure of the J train operator to comply with the stop indication because he was asleep.

As a result of this accident, the Safety Board issued several recommendations (R-96-20, 21, 22) regarding the development of fatigue awareness and educational programs to the Federal Transit Administration (FTA), the American Public Transit Association (APTA), and New York City Transit (NYCT). In response to these recommendations, a joint symposium by the FTA, APTA and the FRA was convened in Washington, D.C. on February 18 - 20 of this year for the purpose of educating the transit industry on the issues and concerns about transit operator fatigue as an important element of the fitness for duty equation. Approximately 100 people representing large transit operations, academia, and government attended the symposium. On March 26, 1998, the FTA hosted a fatigue awareness roundtable with participants from government, industry, and labor who further discussed the issue of fatigue. Although the group will continue to refine their objectives, they have proposed that the Department of Transportation develop and issue a department-wide policy on fatigue awareness.

The NYCT also responded to the Board’s recommendation by developing a training and education module to inform train operators and other employees involved in safety-sensitive positions about the hazards of performing their duties while fatigued. The Board commends NYCT for being the first transit system in the United States to develop such a program.

The Board also focused on training and education in its report on the head-end collision of Consolidated Rail Corporation (Conrail) freight trains near Thompsontown, Pennsylvania, that occurred on January 14, 1988. About 7:54 a.m., two Conrail freight trains – one travelling west and the other east – collided. The engineers and brakemen on both trains were fatally injured, and the conductors on both trains received minor injuries. Damage to the trains was estimated at over $6 million. The Safety Board determined that the probable cause of the accident was the sleep-deprived condition of the engineer and other crewmembers of one of the trains, which resulted in their inability to stay awake and alert, and their consequent failure to comply with restrictive signal aspects.

One of the issues discussed in the report was the train crew’s fitness for duty. As a result, the Board recommended that Conrail provide education and counseling to employees and their families on proper health and diet regimens, as well as the avoidance of sleep deprivation (R-89-11). The Board also recommended that the Brotherhood of Locomotive Engineers and the United Transportation Union cooperate with Conrail and the other railroads in the implementation of voluntary education and counseling programs designed to improve train crewmembers’ knowledge of proper health and diet regimens, as well as the necessity to avoid sleep deprivation. (R-89-19).

Irregular and Unpredictable Schedules

For many years the NTSB has been concerned about the unpredictable nature of train crew work assignments and its attendant effect on crew fatigue. Although there are some exceptions, the majority of train crews are subject to call with little notice. The Board pointed out in its 1985 report on Burlington Northern Railroad collisions in Wiggins, Colorado, and Newcastle, Wyoming (Railroad Accident Report-NTSB/RAR-85/04) that railroad crews are confronted by the most unpredictable work/rest cycles in the transportation industry.

The unpredictability of work assignments was also discussed in great detail in the Board’s report on the Thompsontown, Pennsylvania collision I referred to earlier. The report noted that the engineer and brakeman never adapted to their nonsystematic pattern of work times and that they probably were susceptible to sleep disorders and chronic sleep deprivation resulting in fatigue, frequent microsleeps or lapses, and napping. The report further concluded that the crewmembers were probably chronically sleep deprived because their work shifts and off-duty periods at home were unpredictable and irregular. As a result of the accident, the Safety Board issued recommendation R-89-012 to Conrail, recommending that they improve the current methods of utilizing train crews to reduce the irregularity and unpredictability of crewmembers’ work/rest cycles.

The Safety Board also focused on the lack of sleep and the irregular and unpredictable work schedule of train crews in its report on the head-on collision of the Atchison, Topeka and Santa Fe Railway Company (ATSF) trains in Corona, California, on November 7, 1990. Again, it was determined that the crewmembers were asleep and consequently failed to stop for a red signal. Contributing to the accident was the irregularity and unpredictability of the engineer’s work schedule. In an effort to mitigate the unpredictability of work assignments, the Board recommended that the ATSF improve its automated train and crew line-up telephone and computer monitor system by posting the times of the last update and the next projected update so that crewmembers would be better able to plan their rest (R-91-43). ATSF responded by posting the time of the last update, as well as the next update, at 4-hour intervals.

Abrupt Changes in Rest Cycles

The Board has also investigated the effect of abrupt changes in work/rest cycles of train crewmembers as related to causal or contributing factors in railroad accidents. During its investigation of the Sugar Valley, Georgia, accident previously mentioned, the Board determined that the engineer who failed to comply with a stop signal resulting in the accident had reverted to a routine of sleeping at night on his three days off duty before the accident after having been on his normal night-work and day-rest routine for over three weeks. Specifically, he had slept eight hours or more on the three nights previous to going on duty at 9:30 p.m. the night before the accident, which changed an established work/rest pattern. In addition, he had been awake for more than 17 hours at the time of the accident. The need for sleep typically recurs after about 15 or 16 hours of being awake, even for someone who is well rested.

The Board also determined that the train operator who was at the controls of a train involved in the accident I previously discussed on the Williamsburg Bridge in June 1995, had recently changed from a weekend schedule of sleeping at night to his weekday schedule of working at night. A significant problem in rotating shift work is the time required for an individual’s internal biological, or circadian clock, to change. When there is a change in a person’s work/rest cycle, the body does not adjust immediately. Adaptation to the new schedule requires several days to several weeks. As a result of that accident, the operator sustained fatal injuries, 64 passengers were injured, and damages and associated cleanup exceeded $2.3 million.

The circumstances of a collision between two freight trains last summer near Delia, Kansas, also implicate abrupt changes in the work/rest cycle of an engineer who was at the controls of one of the trains involved. About 2:15 a.m. on July 2, 1997, a westbound Union Pacific freight train failed to comply with a stop signal at a point where a main track and siding merge. As a result, it collided with an eastbound Union Pacific freight train that was traveling at about 70 miles per hour. The accident resulted in the death of the engineer of the westbound train and caused over $5 million in damages. The derailed cars released hazardous materials that forced the evacuation of more than 1,400 people from the nearby community -- including 76 residents of a nursing home. Some of the nursing home residents were hooked up to oxygen and intravenous tubes or were on medication, so their evacuation also entailed moving essential medical records and supplies.

The Board’s investigation has revealed that the engineer of the westbound train had been on vacation for about two and a half weeks before reporting back to work the night before the accident. During his vacation, he slept at night and obtained anywhere from seven to eight hours of sleep during those periods. The day before the accident he awoke at about 8:00 a.m., and remained awake until he reported for duty at 8:30 p.m., 12 ½ hours later. The accident occurred at about 2:15 the following morning. At the time of the accident, the engineer had been continuously awake for more than 18 hours. Furthermore, the accident occurred at a time when for the past two and a half weeks he had been asleep. It is likely that in this case the engineer did not have the necessary time for his body to match his new sleep/wake cycle. As a result, he was probably not prepared to remain awake all night and into the early morning hours. I would like to point out that this accident is still under investigation, and conclusions concerning its cause are still being developed.

These accidents clearly illustrate the devastating effects that abrupt changes in work/rest schedules can have on railroad safety. The results of such changes can have profound effects on human performance, including degraded alertness and increased reaction time. Studies have shown that shift workers who rotate schedules are especially prone to fatigue on both the first and second nights of their work week.

As noted above, in some of our investigations it was determined that the individuals involved had been awake for 18 or more hours at the time of the accident. Research conducted by Dr. Drew Dawson of the Centre for Sleep Research, at the University of South Australia, equates impaired human performance due to sustained wakefulness with that caused by alcohol intoxication. The results of this research indicate that after 18 hours without sleep, individuals performed as poorly on cognitive and motor skills tests as they had with a blood alcohol concentration (BAC) of 0.05%. After 24 hours without sleep, performance equated to that of a person with a BAC of 0.096%.

To put those research findings in context, consider that FRA regulations prohibit a railroad employee from going on duty or remaining on duty with a blood alcohol concentration of 0.04% or greater. But there are no regulations prohibiting an employee from going on duty or remaining on duty after a sustained period without sleep.

The discussion of the above-mentioned railroad accidents represents only a small sample of accidents stemming from fatigue, and the Safety Board is pleased that the FRA has incorporated the Board’s findings and recommendations from the Board’s reports into their proposed plan.

The Safety Board also notes that the successful development and implementation of any plan requires the establishment of some critical features. For instance, the preparation of a comprehensive fatigue management program requires technical expertise from personnel involved in development and evaluation of this plan. After its implementation, the evaluation of the program becomes a critical phase. Therefore, specific criteria must be developed in order to measure the program’s effectiveness. Moreover, an effective program must also be capable of making modifications to its original plans as necessary. Shortcomings in any phase of the program certainly could result in its failure. Therefore, the Safety Board strongly urges the FRA to include in its plan careful consideration of specific personnel and other resources needed to help ensure the success of this program.

Another Board concern about FRA’s proposed requirement for fatigue management plans is that it fails to squarely confront the root cause of fatigue: the provisions of the current Hours of Service Act. In fact, the amendments would even permit a waiver of any provision of the hours of service laws.

The Hours of Service Act

As I stated earlier, fatigue has been an issue of great concern to the Safety Board for many years. While it remains one of the more perplexing problems to substantiate in accident investigations, the body of scientific evidence collected over the past decade clearly reflects the critical need for adequate rest for those people operating the transportation system. I do not believe that anyone can honestly dispute that evidence today. However, it also appears to be one of the most challenging problems to fix.

The Department of Transportation has spent millions of federal dollars researching this human condition in most of the modes of transportation and has even embarked upon rulemaking efforts to remedy it. But little meaningful progress has been made, we believe, because the solution requires a fundamental change in habits and culture – and neither is easy to change. Labor has grown accustomed to the extra money earned and companies save money by employing fewer operators. This was made evident in testimony given at a rail hearing the Safety Board recently held. During that meeting, union representatives complained of long working hours, yet stated their objection to revisions to the hours of service regulations because of the pay available to them through the current system. We must all recognize that fatigue is equally debilitating, whether it is the result of voluntary or required scheduling, and fewer workers and more overtime are the fundamental ingredients for fatigue.

It is reality that the hours of service regulations form the parameters in which schedules are developed. While we applaud the work being done at some individual railroad companies, this problem is not unique to any one railroad. It is a national problem that is deserving of national attention. Reducing the hours of service parameters would prevent gross abuses of work hours and would provide a level playing field upon which all workers can be provided a healthier work environment.

Chairman Franks, as the Safety Board has testified at previous hearings before this Subcommittee, the Hours of Service Act is flawed. It has been flawed since it was first enacted in 1907, and has remained flawed throughout its substantial revision in 1969 and its amendments in 1976 and 1988.

We believe that the Hours of Service Act is flawed for 3 reasons: the first is the burdensome amount of work and the minimal amount of rest that the laws permit; the second reason is the findings of the Safety Board’s accident investigations; and the third reason is the lack of any scientific support for the work/rest provisions of the current law. Let me enlarge on those 3 points.

Burdensome Amount of Work

The current railroad hours of service laws permit, and many railroad carriers require, the most burdensome fatigue-inducing work schedule of any federally regulated transportation mode in this country. A comparison of the modes is revealing. The aviation, highway, marine, and rail modes all have federally imposed limits on the amount of work and rest in a 24-hour period. The aviation and highway modes also impose weekly limits. Only aviation has monthly and annual limits. To keep the comparison simple, consider the number of hours an employee of each mode is permitted to work in the course of a 30-day month:

• A commercial airline pilot can fly up to 100 hours per month;

• Shipboard personnel, at sea, cannot operate more than 240 hours per month;

• A truck driver can be on duty up to about 260 hours per month; and

• Locomotive engineers can operate a train up to 432 hours per month, which equates to more than 14 hours a day each of those 30 days.

We fail to understand why a locomotive engineer, or other train crew member, is permitted to work more than 4 times longer than an airline pilot, and 1.5 times longer than a truck driver. Beyond the self-images that pilots, engineers, or even subcommittee or agency chairmen may have of themselves, we are all alike in our humanness and our need for rest. Let me emphasize that we are not advocating reducing everybody’s hours to 100 hours a month. Our point is that allowing any transportation worker in a safety-sensitive position – operating powerful equipment through our Nation’s cities -- to work over 400 hours per month is excessive, if not downright unconscionable.

Safety Board Accident Investigation Findings

The second reason for my saying that the current hours of service laws are flawed is based on the Safety Board’s investigative experience. In all of the railroad accident investigations in which the Board determined fatigue to be a causal or contributing factor, the train crew members were all in compliance with the Hours of Service laws.

However, I should add that in a number of our investigations, some crew members did not avail themselves of the opportunity to get sleep during their off-duty period. Generally, it was because either they had an expectation that they would be called for duty at a later time, or their time off was during the day and they found it difficult to sleep.

The Safety Board also believes the Hours of Service laws have no scientific basis. In fairness to those who framed the laws in 1907, there was little more than anecdotal knowledge about fatigue at that time. But in the last two decades, the scientific and research communities have conducted an in-depth study of sleep and fatigue. We now know a great deal about the structure of sleep, the effects of human biological or circadian rhythms, and the debilitating effects that cumulative sleep loss has on alertness and health.

The railroad hours of service are too simplistic. They prescribe only maximum hours on duty and a minimum amount of rest in a 24-hour period. They do not take into account (1) how human circadian rhythms interact with the time of day when the work/rest periods take place, (2) the cumulative effects of working an unlimited number of successive days, or (3) the long-term health effects of various work/rest schedules. In short, it is time for a substantial, scientifically-based revision to the Hours of Service Act.

Chairman Franks, in closing, I would like to read a fax that one of my railroad investigators received last week from the widow of the engineer who was killed in the tragic collision last July near Delia, Kansas. I spoke earlier in my testimony about that accident and a related issue of the unpredictability of train lineups. Her fax provides us with a far better perspective than I can provide. I sincerely thank her for sharing her heartfelt thoughts with us.

"On July 1, 1997, My husband Mike called the recording to find out when he would supposedly be going to work. Afterwards, he left to run errands. Later I called the recording too, but of course it said the same thing all day. That he was on the line-up for 5:00 p.m. That afternoon he came home and called it again. After listening he said ‘well, as usual the lyin-ass line-up isn’t holding up.’ Mike and some of his co-workers called it the ‘lyin-ass line-up,’ being that it was so highly inaccurate. Several hours went by before they finally called him for 8:30 p.m. This was a common occurrence. He never knew when he would leave, how long he would be gone, or how long he would be home, either. He used to say our local weather forecaster was more reliable than the line-up. We could never plan anything. About the time he would decide to give up on the call and make other plans, they would call him to work. Sometimes he would be ready for work, 8 to 12 hours before he would finally receive a call. There is no other industry that I am familiar with, which is so unprofessional as to keep their employees uninformed about something as essential as their work schedule. The last time I called that recording, it was again inaccurate. It said Mike was still on duty. That was 8:30 a.m. July 2, 1997, approximately two hours after Mike was pronounced dead."

We as a government need to decide to what extent the status quo is acceptable. If we can agree that fatigue-caused accidents are unacceptable, then we must move to change the status quo.

Mr. Chairman, that completes my testimony, and I will be happy to respond to any questions the you or the Subcommittee Members may have.

Jim Hall's Speeches