The Human Factor: A Perfect Storm of Fatigue

A look back at fatigue's role in the 2006 Comair crash.

The critical role of fatigue in aviation accidents.

The critical role of fatigue in aviation accidents.

(April 2011) AFTER WRITING THREE articles on the subject of fatigue I figured I had pretty much covered that topic. Then I started receiving e-mails and phone calls from Flying readers, more than I have ever received before. It is obvious that fatigue is a critical issue for many people.

Paul Reeves described his own battle with micro naps that developed into micro sleep intrusions in which he was “still awake but under severely degraded motor function and a zombielike cognitive state — eyes open, brain off.” He said that he could not tell when this occurred, and could hear everything in real time, but he had dreamlike experiences that were close to hallucinations. Paul stopped driving and lost his medical for eight years while he got help with this problem. He pleaded with other pilots with sleep disorders to seek help, because he has “no doubt if I had continued on my own, you would not be reading this letter.”

Dr. L.C. Carmichael, who works at a sleep disorder center, echoed that plea, saying that studies indicate “there is a sizable number of pilots with a sleep disorder. One should not duck and run at the thought of being a member of that group as there are a number of successful strategies available for many sleep problems and one has only to seek the advice of a sleep physician.”

The most compelling and eye-opening responses came from air traffic controllers. It turns out that many air traffic facilities use a rotating shift, with controllers working evening, afternoon, morning and night rotations in the span of one week. While this allows an 80-hour break between work periods, one retired controller said that “it certainly took a toll on my body and had me working fatigued on many occasions.” He said it was “hard to get a good night’s rest when your shift starts just eight hours after the last one ended.”

Another controller said, “This schedule is exhausting because it is very hard to adjust sleeping times and get a solid eight hours of sleep. Even the weekends are not restful because you are trying to transition back to staying up late for the first late shift when you resume your workweek. Sometimes you get unscheduled overtime which could be any shift, and that can really throw your schedule off.” He said the fatigue makes it hard for him to concentrate and do the job properly.

I had always wondered about accident reports that described a controller with very little traffic or other workload making critical mistakes or even watching an airplane enter severe weather without providing any warning to the pilot. Working the night shift is difficult enough, but the first night on the night shift is the worst. On controllers’ rotating shift schedules, they are changing shifts throughout the week and working one night shift after working that same morning. Research shows that is about the worst kind of schedule from the standpoint of causing fatigue. As one controller said, “You know you should do something, but you can’t force yourself to do it. It becomes hard to do your work.”

It seems like we have a “perfect storm” of fatigue in aviation — tired pilots working with tired controllers. The system of keeping an eye on each other breaks down when everyone involved is having a hard time keeping his or her eyes open. One example of this is the accident at Lexington, Kentucky, in August 2006, when a Comair regional jet crashed after attempting to take off on the wrong runway. This is a particularly interesting case because the pilots and the controller were all respected by their peers and the NTSB did not find any history of past performance issues. There also were no extraordinary circumstances that would have led anyone to be excessively fatigued. The captain had arrived in Lexington at 3:15 p.m. Saturday and had eaten dinner with his family before going to bed. The copilot had arrived at 1:40 a.m. Saturday but then had had no other duties the rest of the day.

Sunday morning, the pilot and copilot woke up at 4:15 a.m. and arrived at the airport at 5:15 a.m. This was earlier than they had been getting up, and a switch for the copilot from getting in very late and sleeping late to getting up very early. Thus it could have caused some disruption to the crew and especially to the copilot. The NTSB report states, “A captain and friend of the first officer described him as an evening person and recalled conversations they had regarding the difficulty of going to bed on nights that preceded an early report time.” The captain also said they had discussed how hard it was to go to bed at 9 p.m. to try to get enough sleep for an early departure, and that going to bed at 11 p.m. “resulted in less but better quality sleep.”

The crew, and especially the copilot, gave lots of indications of not being very sharp that morning, starting with getting in the wrong airplane. They had already started the APU before they were alerted by a ramp agent to their mistake. The Comair standard operating procedure required the captain to conduct a full taxi briefing for their first flight together; however, he merely stated, “Comair standard.” A minute later, the copilot started working through the “before starting engines” checklist until the captain pointed out to the surprised copilot that it had already been completed. They also engaged in nonoperational conversation while taxiing to the runway, and they obviously failed to check that their heading matched the departure runway before initiating the takeoff in the dark on a runway with no lights.

The controller had been working the rotating shift discussed above:

Wednesday 5 p.m. to 1 a.m.
Thursday 3 to 9 p.m.
Friday 8:15 a.m. to 4:15 p.m.
Saturday 6:30 a.m. to 2:30 p.m.
Sunday midnight to 8 a.m.

He stated that he took a two-hour nap from 3:30 to 5:30 p.m. Saturday but didn’t sleep very well. Thus, when the accident occurred at just after 6 a.m. Sunday, he had been awake for 24 hours with only a two-hour nap, and had worked 14 hours during that period. While the NTSB noted that his recorded communications were “prompt and professional,” the controller said he was tired but felt “fine and alert.” However, he failed to record information about taxiway closures required to be included in the ATIS broadcast, and he decided to start an administrative task rather than pause for a few seconds to observe the regional jet as it took off.

One of the problems with fatigue is that, other than in situations in which someone actually falls asleep, it is hard to quantify its effects. While the NTSB did not list fatigue as a probable cause or contributing factor in this accident, it did reiterate several previous recommendations concerning controller fatigue, and in her concurring statement, board member Deborah Hersman referred to the “difficult and dangerous work schedules” that are sanctioned by the FAA and the National Air Traffic Controllers Association (NATCO) as a possible reason that the controller put a higher priority on performing an administrative function than on monitoring the takeoff.

I spoke with Vladimir Nacev, a doctor of clinical psychology and professor at the University of Maryland University College, who emphasized that decision-making is impaired while fatigued, and decision-making is the core of aviation safety. It is very possible that, if it had not been so early in the morning, the Comair captain might have conducted a complete taxi briefing as required, or the crew might not have allowed themselves to be distracted by nonoperational conversation when they were supposed to be sterile cockpit, or they might have questioned why there were no runway lights, or they might have done a final lineup check to ensure the heading matched the assigned runway. And if the controller had not been looking forward to the end of a 24-hour “day,” he might have decided it was more important to observe the takeoff of the regional jet rather than to start counting how many operations had occurred during his shift.

Nacev emphasized that it is critical to not cheat on sleep. In our modern society, there will always be a need for pilots and controllers to work through the night. Even under the best of circumstances those individuals will not be as sharp as they would be on a normal daytime schedule. It seems that, at the very least, the FAA, NATCO and all other organizations that work in the aviation field should pay attention to the wealth of research that proves the devastating impact of rotating shifts, and develop a schedule for controllers, pilots and others that minimizes rather than maximizes that impact. All of us who work in aviation in any capacity need to take circadian rhythms and fatigue seriously and to do everything we can to assure we get the rest we need to operate at a professional level. On those days when circumstances conspire to make that impossible, we also need to be ready to just say "no." Zombies don't belong in the cockpit, in the control tower, in the hangar or anywhere else in aviation.