Airing Dirty Laundry: the ComAir Crash

In the hope that maybe a little fresh air will clean it out



"Feds Say Air Controller Slept 2 Hours"

LEXINGTON, Ky. (AP) — In the day leading up to the crash of Comair Flight 5191, a federal investigator says the air traffic controller on duty had worked for almost 15 hours and slept for two.



National Transportation Safety Board member Debbie Hersman, the lead investigator in the crash that killed 49 people, said in her final briefing before leaving Lexington Thursday that the controller had only nine hours off between work shifts Saturday.



That was just enough to meet federal rules, which require a minimum of eight hours off between shifts, Hersman said.



"He advised our team that he got approximately two hours of sleep," Hersman said.

I worked as a controller for twelve years. In that time, the "backward rotation" or "phase advancing" was de riguer for controllers. This one evidently had nine hours off. More frequently, it's eight, the bare legal minimum.



At Los Angeles Center, where I worked for part of that time, the schedule is typical for the twenty-four hour facilities. Your first shift of the week started at 2:30pm, and you worked until 10:30pm. Then your next one started at 1:30, and ended at 9:30. Due to staffing considerations, lunch was subject to interruption for traffic, and you were not allowed to leave the facility, therefore the straight eight hour shift.



The next shift after that could be anything. Sometimes it started at noon, sometimes at 10, sometimes at 8, and perhaps at 6:30 am, leaving you only nine hours off between shifts. The fourth shift always started at 6:30 am and went to 2:30 pm. And finally the fifth shift could be another 6:30 start time, or it could be a 10:30 pm start time that same night. This left an individual controller with the following number of hours between shifts: 15, anywhere from 9 to 14.5, then between 16 and 10.5 depending upon the previous gap, and finally either 16 or 8. So whereas a worker with a regular shift has an aggregate of four 15 or 15.5 hour gaps, for a total of 60 to 62 hours of downtime during their work week, the controller has a total of either 56 or 48. It doesn't seem like much of a difference, between eight and fourteen hours. But the only place for that difference to come out of is sleep. They still have to drive home. They still have to unwind. Most folks need to eat. They still have to get ready for work, and they still have to drive back. Particularly during the rapid rotation shift that you would have every week, and especially if you had a mid shift so that you faced two rapid shifts forward, if you failed to get to sleep quickly during that period, you were a zombie during your next shift, as you were starting from a maximum of about five and a half hours of sleep, and losing sleep off that. I lived right in Palmdale, but many controllers there lived in places 45 minutes to an hour and a half drive each way (to be fair, some of them made a habit of using motels for the "quick turn" shifts.)



The FAA, for its part, always said (publicly) that it supported controllers who didn't feel sufficiently rested to work safely. However, in practice, if a controller called in with something like that as a reason, they could expect plenty of grief over it, ranging from a supervisor who let you know in no uncertain terms that they were unhappy, to putting records in your file, to using it as justification for a less satisfactory performance review than you might otherwise have gotten. The dichotomy between official position and what really would happen was clear, and you were in for no small amount of grief if you actually did call in because you felt you weren't rested enough to be alert. Nor could you count on your peers to be supportive, as it meant they had to work harder. Finally, the shifts where you were guaranteed to have the most accumulated sleep-debt were the day shifts, which were the busiest time, towards the end of the week.



Why is this done? Well, it was in place before the 1981 strike, but after the strike it sure made staffing a lot easier to manage. And of course if you offer people three day weekends every week, that's kind of attractive to the average person.



Now try this search: Effects on alertness of rotating shiftwork. Add the phrase "Air Traffic" in there, and you get all of this wonderful research, and more:



Effects on alertness of rotating shiftwork

advancing continuous systems seemed to be associated with marginally steeper declines in alertness across the shift (F (3,1080)=2.87, p<0.05). They were also associated with shorter sleeps between morning shifts (F (1,404)=4.01, p<0.05), but longer sleeps between afternoons (F (1,424)=4.16, p<0.05).

and

The absence of negative effects of advancing shifts upon the chronic outcome measures accorded with previous evidence that advancing shifts may not be as harmful as early research indicated. However, this interpretation is tempered by the possibility that difficult shift systems self select those workers most able to cope with their deleterious effects. The presence of quick returns in advancing continuous systems seemed to impact upon some of the acute measures such as duration of sleep, although the associated effects on alertness seemed to be marginal.

Shiftwork in the Practice of Emergency Medicine

Clockwise shift rotation (phase delaying) places less strain on the adaptive ability of the human internal clock than phase advancement. Such rotation has been shown to result in greater worker satisfaction manifested by fewer complaints about ill health and work schedules. Studies suggest a 20% increase in productivity with phase-delayed shifts compared with phase-advanced shifts.



With phase-advanced scheduling, for every hour of advancement, a full day is needed for entrainment. For example, phase advancement to an earlier shift every 7 days necessitates a week or longer to adapt after each rotation. Up to 25% of workers may not adapt. Studies suggest that the rotation most consistent with human circadian rhythm is 1 rotation every 21 days. Groups rotated at this rate have 70% fewer complaints than groups rotated every 7 days.

Into the Night: Coping with the Effects of Shiftwork

Sleep disorders can also occur, making it difficult during off duty to go to sleep, stay asleep, or experience a high quality of sleep. About 63% of nightworkers complain of sleep disturbance. Sleep length of night workers may be only 4-6 hours compared to day and afternoon workers who average 7-9 hours. This loss of sleep can become a "sleep debt" that robs an officer of energy and alertness. Evidence of sleep deficit can be seen in as short a time as 2 days of inadequate sleep, and with as few hours of sleep loss. Significant sleep deficit can accumulate after more than 3-4 nights worked in a row, and some researchers believe that a schedule of 6 nights in a row may be too exhausting (Scott, 1994). When this occurs, an officer can experience brief periods of "microsleep" in which normal activities are engaged in, but the person slips into light sleep for periods of 1-10 seconds. In one study of microsleep, participants were asked to press a button when a bright strobe light was flashed directly in their eyes every few seconds. During microsleep they did not notice the light or even that they had been asleep (Dement, 1974). Microsleep does not give a warning.



The psychological and behavioral effects of shiftwork can be equally troubling. When keeping a vigil, as in surveillance work, as early as 20-35 minutes after starting, concentration and attention can begin to lag (Krueger, 1989). Higher thinking skills are also affected. When fatigued, memory and recall are slower, logical and arithmetic reasoning have more errors, decision making is slower, and report writing and comprehension are not as good (Balkin & Badia, 1988). There is often a temptation to take shortcuts that can result in not following procedures, mishandling evidence, and safety violations.

and

Although there is no single work schedule that is optimal for all tasks (Krueger, 1989), many industries and emergency services have moved from three eight-hour shifts over several weeks, to two 12-hour shifts over about four days. Moving from 8-12 hours does not seem to significantly interfere with performance, and four days is about the limit of doing nightwork without making circadian changes (Folkand, 1992; Walker & Eisenberg, 1995; Williamson, Gower, & Clarke, 1994) . Studies examining such schedules have reported few ill effects and several improvements: increased productivity, higher morale, lessened fatigue, better time blocks for free time, improved family relations, reduced commuting, improved health, and more job satisfaction. Nonetheless, it is not possible at this time to recommend that all police departments convert to such a schedule. It is more advisable to design a schedule that fits the demands, risks, and personal needs of each organization.

NIOSH Update: (emphasis mine)

* As appropriate, employers might consider changes in shiftwork schedules — such as considering alternatives to permanent night shifts, avoiding quick shift changes, and adjusting shift length to the workload. Whether a particular change is useful depends on the specific work situation. When changing employees' work schedules, all aspects of the worker's job and home life should be considered, the publication suggests.



* Other potentially useful steps include scheduling heavy or demanding work at times when workers are most alert or at peak performance, providing training or awareness programs for new shiftworkers and their families, and ensuring that health care and counseling services are available to employees who work non-traditional schedules.



* Employees may consider various ways for coping with shiftwork, such as increasing their awareness of the need to get good sleep, establishing the sleep routine that works best for the individual, and looking at the utility of exercise, diet, and relaxation techniques for helping resist stress.

Shiftwork Annotated Bibliography: (emphasis mine)

This relatively comprehensive review article of research on shift schedules begins by stating, "All shift systems have advantages and drawbacks. There is no single 'optimum shift system' which can be used in industry or commerce at all work places. However, there are shift systems which are more favorable, and others which are less favorable, in the context of physiological, psychological, and social recommendations for the design of shift systems." (p.15)



Based on the research reviewed in the article, the author makes the following recommendations for shift design:



1. Night work should be reduced as much as possible, or rapid rotation should be used when needed. Slow rotation and permanent night shifts are not advised.



2. Extended workdays of 9 to 12 hours should only be used if the nature of the work is suitable. Further, where extended workdays are used, accumulated fatigue should be minimized through limited days-in-a-row.



3. An early start for the morning shift should be avoided.



4. Quick changeovers between shifts must also be avoided.



5. Consecutive days should be limited to 5 to 7 days. Additionally, schedules should include some free weekends.



6. The forward or clockwise rotation of shifts appears to be the recommended approach for continuous shift schedules.

this one has some counter-evidence:

In the UF study, 19 of the air traffic controllers worked two or three night shifts, from 4 p.m. to midnight, followed by two or three day shifts, from 8 a.m. to 4 p.m. The other 18 worked a rapid rotation of these two shifts, followed by a third one from midnight to 8 a.m.



Over seven days of computerized testing, the controllers working the rapidly rotating three-shift schedule demonstrated a quicker reaction time during a series of spatial visualization and tracking tasks on the computer, McAdaragh said. They also improved their attention skills when learning a new cognitive task, while the other group did not, he said

(But I found no evidence they considered shiftwork rotating in the opposite direction)



but hits the nail on the head for the reason:

Air traffic controllers tend to prefer these counterclockwise rapidly rotating work schedules over weekly rotating schedules anyway because of the greater breaks they provide between work weeks, he said.

Now the research appears to be strong, if not unanimous, that rotating shifts in this fashion has detrimental effects upon alertness, alertness which is the controller's primary stock in trade. It doesn't matter how great my Plan For The Situation is if some pilot does something unexpected and I don't notice right away. The sooner I see it, the sooner I have the possibility to fix it. The longer the situation goes, the worse it gets. And in this situation, the controller who got two hours sleep didn't notice at all. Perhaps he wouldn't have noticed anyway, and it is supposed to be the pilot's responsibility to fly the aircraft in accordance with directions. Nonetheless, in practice, particularly for itinerant pilots at commercial airports, you do have to keep an eye on them to make certain they don't do stuff like take off on the wrong runway, as happened here, or just taxi across the active at the wrong time. Nor is this the first time such concerns have been cited as a contributing factor. A more alert controller would have a higher likelihood of spotting such an incident before it gets to an unavoidable stage.



Maybe he wouldn't have noticed anyway. But is seems fairly likely to me that another 49 people just died so that controllers can have longer weekends.



UPDATE: I should also link to this over at Argghhh! again to place it in context with a pilot's eye view