Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 77
4
Sleep, Wakefulness,
Circadian Rhythms, and Fatigue
FATIGUE
Prevention of fatigue as a safety risk in commercial aviation operations
has focused on effective management of duty and rest scheduling (Dinges et
al., 1996). When considered in the context of work safety, fatigue has been
broadly defined as a biological drive for recuperative rest (Williamson et al.,
2011). Over the past several decades, the scientific knowledge base about
the causes of fatigue and its effects on performance has grown significantly.
The Aviation Fatigue Management Symposium: Partnerships for Solutions,
supported by the Federal Aviation Administration (FAA) (2008), included
several presentations summarizing the state of the science relevant to fa-
tigue in aviation (and other transportation modes).1 Additional work was
presented a year later in International Conference on Fatigue Management
in Transportation Operations: A Framework for Progress (U.S. Department
of Transportation, 2009).
It is well established that fatigue has multiple interactive sources. The
primary ones that may be relevant to pilots’ commutes include duration of
time awake prior to work, duration of time slept prior to work, restfulness
of sleep (i.e., sleep continuity) prior to work, and the biological time (i.e.,
circadian phase) at which sleep and/or commuting occur relative to the start
of work. The duration of time at work (i.e., time on task) is a regulated
factor for fatigue mitigation.
1 For more information on the Aviation Fatigue Management Symposium: Partnerships
for Solutions, see http://www.faa.gov/news/conferences_events/2008_aviation_fatigue/ [June
2011].
77
OCR for page 78
78 THE EFFECTS OF COMMUTING ON PILOT FATIGUE
In the aviation industry, commutes that involve travel across multiple
time zones have the potential to exacerbate the fatigue associated with
commuting, as can chronic restriction of sleep for multiple days prior to
commuting. It is important to recognize that these fatigue effects can be
mitigated to some extent by following good sleep hygiene practices2 in the
period between the end of the commute and the time of reporting for duty.
Due to a lack of relevant data, it is unknown to what extent good sleep
hygiene practices are followed by commuting pilots to ensure they are alert
during their postcommute flight and duty periods.3
Extensive scientific evidence exists on the negative effects of fatigue on
the performance of many cognitive tasks, including those essential for safely
operating a commercial aircraft. The adverse effects of fatigue induced by
sleep loss include maintaining wakefulness and alertness; vigilance and se-
lective attention; psychomotor and cognitive speed; accuracy of performing
a wide range of cognitive tasks; working and executive memory; and higher
cognitive functions, such as decision making, detection of safety threats,
and problem solving; and communication and mood (Harrison and Horne,
2000; Thomas et al., 2000; Durmer and Dinges, 2005; Philibert, 2005;
Banks and Dinges, 2007, 2011; Goel et al., 2009b; Lim and Dinges, 2010).
The Institute of Medicine (IOM) defines fatigue as “an unsafe condi-
tion that can occur relative to the timing and duration of work and sleep
opportunities” (Institute of Medicine, 2009, p. 218). It further states:
In healthy individuals, fatigue is a general term used to describe feelings of
tiredness, reduced energy, and the increased effort needed to perform tasks
effectively and avoid errors. It occurs as performance demands increase
because of work intensity and work duration, but it is also a product of
the quantity and quality of sleep and the time of day work occurs.
Pilot commuting practices and individual day-to-day experiences can be
quite variable, depending on many factors. The extent to which pilot com-
muting is contributing to fatigue at work—by reducing sleep time, extend-
ing wake time prior to duty, or interrupting a habitual nocturnal sleep
period—is not known.
2 Good sleep hygiene practices generally refer to those behaviors that effectively control
all behavioral and environmental factors that precede sleep and may interfere with sleep, to
ensure the sleep is as restful as possible, in order to promote daytime alertness or help treat
or avoid certain sleep disorders (see Thorpy, 2011).
3 The committee did not consider the use of sleeping medications by pilots during commutes
prior to duty because the FAA has restrictions on pilots’ use of Federal Drug Administration-
approved prescription sleep medications, over-the-counter drugs, and supplements for sleep.
OCR for page 79
79
SLEEP, WAKEFULNESS, CIRCADIAN RHYTHMS, AND FATIGUE
SLEEP AND CIRCADIAN RHYTHMS
A full understanding of the relationship between commuting and pilot
fatigue is complicated by the fact that there are inadequate data on the
timing, duration, and quality of pilots’ sleep before and during commutes.
Quality of sleep encompasses factors that can affect the recuperative value
of sleep, immediately prior to and during a commute period, such as noise,
light, body posture, sleep surface, and ambient temperature.
Time Awake, Sleep Time, and Circadian Time
Scientific understanding of the interaction of circadian biology and
homeostatic sleep need is fundamental to identifying how fatigue can oc-
cur relative to commuting. Circadian rhythms are daily (24-hour) rhythms,
reflected in microbiology, physiology and behavior, that control the timing
of the sleep/wake cycle and influence physical and cognitive performance,
activity, food consumption, body temperature, cardiovascular rhythms,
muscle tone, and aspects of hormone secretion and immune responses, as
well as many other physiological functions. When an individual is acutely
sleep-deprived by remaining awake into his or her habitual nocturnal sleep
period, elevated homeostatic pressure for sleep due to time awake extend-
ing beyond 16 hours develops as the internal circadian clock in the brain
is withdrawing the drive for wakefulness (Van Dongen and Dinges, 2005;
Institute of Medicine, 2009). Performance deficits not apparent up to 16
hours awake can suddenly become evident as a result of these two interac-
tive processes (i.e., increasing sleep pressure and decreasing wake drive).
These deficits can be similar to those observed when people are under the
influence of alcohol.4
The quantifications of fatigue-related performance noted above suggest
that pilots should not be awake beyond approximately 16 hours at the time
a duty period ends, unless there are unexpected reasons for this to occur
or adequate system mitigation (discussed in Chapter 3). To the extent that
commute time may lengthen a duty day beyond this threshold, such com-
mutes should be avoided.
The sleep homeostatic drive can produce fatigue during waking per-
4 For example, performance on an unpredictable tracking task after being awake more than
17 hours (after 3:00 a.m.) was equivalent to the effects of a 0.05 percent blood alcohol con-
centration (Dawson and Reid, 1997). Williamson and Feyer (2000) also reported that after
17-19 hours without sleep (corresponding to 10:30 p.m. and 1:00 a.m.), speed or accuracy
on some cognitive tests was equivalent or worse than those found at a 0.05 percent blood
alcohol concentration. Depending on the cognitive task measured, 20-25 hours of wakefulness
produced performance decrements equivalent to those observed at a blood alcohol concentra-
tion of 0.10 percent (Lamond and Dawson, 1999).
OCR for page 80
80 THE EFFECTS OF COMMUTING ON PILOT FATIGUE
formance at work as a result of inadequate sleep duration or poor quality
sleep in the day prior to work (e.g., due to environmental disturbances or
physical problems within the individual such as illness or a sleep disorder).
Whether due to being awake too long prior to work or to sleeping too little
prior to work, the elevated pressure for sleep in the human brain results
in subjective fatigue and sleepiness, and degradation of attention, work-
ing memory, mental speed, and other cognitive performance functions,
including higher-order functions involved in decision making (Harrison and
Horne, 2000; Killgore et al., 2006; McKenna et al., 2007; Venkatramen et
al., 2007). These cognitive changes can result in turn in adverse effects on
work performance (Mitler et al., 1988; Dinges, 1995). Fatigue as a risk to
individual pilot performance can result from (1) being awake continuously
for more than approximately 16 hours, or (2) sleeping too little (especially
less than 6 hours on the sleep opportunity prior to work), or (3) when
undertaking work at a time when the body is biologically programmed to
be asleep (i.e., an individual’s habitual nocturnal sleep period), which for
most people is between 10:00 p.m. and 7:00 a.m. (Van Dongen and Dinges,
2005; Basner and Dinges, 2009; Institute of Medicine, 2009).5
Fatigue-related performance deficits from inadequate sleep can vary
markedly across a day and night (without sleep). This variation in perfor-
mance is due to the fact that sleep and circadian drives in the brain interact
nonlinearly in the control of performance and alertness (Dijk et al., 1992;
Goel et al., 2011). For example, performance deficits after being awake at
night peak between 6:00 and 10:00 a.m. but are less severe by 6:00-10:00
p.m. (12 hours later) (Goel et al., 2011). The detrimental effects of fatigue
on performance may be exacerbated by a tendency for individuals to have
reduced awareness of their cognitive performance deficits that result, even
as these deficits increase in frequency with consecutive days of inadequate
sleep (Van Dongen et al., 2003a; Banks et al., 2010).
Although the effects of acute sleep deprivation on performance may be
transiently influenced by such factors as social and physical activity (Goel
et al., 2011), a recent meta-analysis of 70 articles that covered 147 cogni-
tive tests of several moderators identified time awake as the most significant
predictor of behavior during a period of acute sleep deprivation (Lim and
Dinges, 2010). This finding could be especially relevant to those instances in
which pilots may obtain little to no sleep within the 24 hours before a flight
5 The period of habitual sleep time at night has also been identified as encompassing the
“window of circadian low,” defined as the hours between 2:00 a.m. and 6:00 a.m. for in-
dividuals adapted to a usual day-wake/night-sleep schedule. This estimate of the window of
circadian low is calculated from extensive scientific data on the circadian low in performance,
alertness, subjective fatigue, and body temperature (see Dinges et al., 1996).
OCR for page 81
81
SLEEP, WAKEFULNESS, CIRCADIAN RHYTHMS, AND FATIGUE
and then undertake a lengthy duty day. In such an instance, sleep time is
reduced and time awake is increased, and both factors contribute to fatigue.
Much is known about the cognitive and functional deficits that result
when healthy adult volunteers remain awake for 24-40 hours (Harrison
and Horne, 2000; Philibert, 2005; Institute of Medicine, 2006, 2009; Goel
et al., 2009b). Scientific understanding of the effects of sleep deprivation
on cognitive functions has accumulated for more than a century (for re-
views of this extensive literature, see Kleitman, 1963; Dinges and Kribbs,
1991; Harrison and Horne, 2000; Durmer and Dinges, 2005; Institute of
Medicine, 2009).
The cognitive effects of sleep deprivation are due to changes in the
brain. Recent advances in neuroimaging technologies have provided further
insights into physiological changes in the brain and underlying performance
functions that manifest themselves when fatigue results from reduced sleep
(Portas et al., 1998b; Drummond et al., 1999, 2005; Drummond and Brown,
2000; Thomas et al., 2000; Bell-McGinty et al., 2004; Chee and Chieh,
2004; Habeck et al., 2004; Chee et al., 2006, 2008; Chuah et al., 2006; Wu
et al., 2006; Lim et al., 2007, 2010; Venkatramen et al., 2007; Institute of
Medicine, 2009).
It is now recognized that although most adults exposed to a night
without sleep experience fatigue-related declines in performance, the timing
and severity of the declines vary across individuals, including pilots (Doran
et al., 2001; Leproult et al., 2003; Van Dongen et al., 2004; Bliese et al.,
2006; Institute of Medicine, 2009). These differences in individual cogni-
tive vulnerability to sleep loss may have a basis in biological factors (e.g.,
normal genetic variation) regulating sleep and circadian rhythms (Goel et
al., 2009a, 2010; Institute of Medicine, 2009). People with untreated sleep
disorders are also subject to individual vulnerability and may experience
negative effects on their performance and safety beyond those experienced
by healthy individuals.
Chronic Partial Sleep Deprivation
In addition to acute sleep deprivation, fatigue can be exacerbated by
chronic partial sleep loss, also known as cumulative sleep debt, which oc-
curs when the sleep obtained over multiple days is too short in duration
to maintain behavioral alertness during the daytime (Van Dongen et al.,
2003b). There is scientific evidence that chronic sleep restriction results in
cumulative performance deficits across days and that the rate of the perfor-
mance decline is inversely proportional to the sleep obtained (Dinges et al.,
1997; Belenky et al., 2003; Van Dongen et al., 2003a). These conclusions
are also supported by data from an experiment in which daily sleep was
chronically obtained by supplementing various shortened nocturnal sleep
OCR for page 82
82 THE EFFECTS OF COMMUTING ON PILOT FATIGUE
periods with varying-duration daytime naps (Mollicone et al., 2007). In
all of these controlled laboratory studies, measures of behavioral alertness
decreased and cognitive performance deficits increased cumulatively across
consecutive days, at a rate inversely proportional to the amount of sleep
provided each day (for reviews, see Banks and Dinges, 2007, 2011).
Performance deficits from chronic sleep restriction can accumulate
across days to levels equivalent to those found after even one or two nights
without any sleep (Van Dongen et al., 2003a). They are also influenced by
the duration of habitual sleep prior to the sleep restriction period (Rupp et
al., 2009). Moreover, chronic sleep restriction that is followed by a night of
little to no sleep results in severe deficits in cognitive performance (Banks
et al., 2010). Recovery of behavioral alertness following chronic sleep loss
often requires extended periods of sleep that are 1-3 hours longer than
habitual sleep for one or more nights (Belenky et al., 2003; Lamond et al.,
2007; Banks et al., 2010).
The threshold at which chronic sleep restriction adversely affects behav-
ioral alertness and cognitive performance in the majority of healthy adults
is when time in bed for sleep is 7 hours or less per 24 hours for a number
of consecutive days (Belenky et al., 2003; Van Dongen et al., 2003a; Banks
and Dinges, 2007, 2011; Mollicone et al., 2007). Since physiological sleep
at night in healthy adults aged 25-65 years averages 90 percent (± 5 percent)
of time in bed (Ohayon et al., 2004)—this percentage is often less with in-
creasing age and when sleep is taken in the daytime—the duration of actual
physiological sleep time during a 7-hour time in bed in a healthy adult can
range from 6.0 hours to 6.7 hours of physiological sleep.
What people report as their usual sleep duration is almost always an
overestimate. For example, a population-based sample of N = 669 middle-
aged adults found that subjective reports of habitual sleep were only moder-
ately correlated (r = 0.47) with an objective measure of sleep time (i.e., wrist
actigraphy), and that there was a systematic over-reporting bias—those who
actually slept 5 hours overreported their sleep durations by 1.2 hours (i.e.,
6.7 hours), and those sleeping 7 hours overreported by their sleep durations
by 0.4 hours (Lauderdale et al., 2008).
Although scientific experiments indicate a minimum threshold of 7
hours’ time in bed for sleep is appropriate for at least 80 percent of adults
(many of whom will require more than 7 hours in bed to achieve the physi-
ological sleep duration necessary to prevent reductions in alertness and
cognitive functions), it is unknown whether this threshold should be ap-
plied to the estimated 20 percent of adults who report sleeping 6 hours or
less per night (Kripke et al., 2002). There is currently no consensus among
scientists and public health experts on (1) what proportion of adults are
naturally short sleepers (i.e., they require 6 or fewer hours of sleep a night
to be alert) and what proportion are simply chronically sleep deprived
OCR for page 83
83
SLEEP, WAKEFULNESS, CIRCADIAN RHYTHMS, AND FATIGUE
at this level of daily sleep; and (2) the extent to which caffeine and other
stimulant foods and supplements—none of which are chemical substitutes
for sleep—can prevent deficits in behavioral alertness and cognitive perfor-
mance under chronic sleep restriction. Although there are extensive studies
on the alertness-promoting effects of caffeine under acute sleep loss condi-
tions, there are few studies on its effects under conditions of chronic sleep
restriction, when tolerance can develop within a few days (Bonnet et al.,
2005). In addition, many people do not consume caffeine, and the many
who do so are not aware of the dose they are ingesting or whether tolerance
has developed from chronic use. Therefore it is difficult to estimate whether
caffeine is being used effectively to prevent the adverse effects of chronic
restriction of sleep (below 7 hours of time in bed) on behavioral alertness
and cognitive performance.
Evidence that cognitive performance is adversely affected when sleep
per 24 hours is cumulatively less than 7 hours (time in bed) suggests that
pilots should seek to obtain sufficient bed time to ensure they are fit for
duty. To the extent that time taken for sleep may be reduced by commuting,
such commutes should be avoided.
Napping
Naps are most easily defined as periods of sleep less than half the
duration of a typical nocturnal sleep period (Dinges, 1989; Dinges and
Broughton, 1989). They are also most often identified with sleep obtained
in locations other than a bed, and sleep when clothing is not removed.
Naps are one of the frequently used fatigue countermeasures (caffeine be-
ing the other), and there is substantial scientific evidence that nap sleep can
help reduce the severity of fatigue during prolonged duty periods of work
(Dinges et al., 1987; Institute of Medicine, 2009). Naps have been found
to be beneficial for fatigue mitigation in commercial pilots (Rosekind et al.,
1994), although some pilots report that some sleep obtained in sleeping
berths on long-haul aircraft—which is most often less than 3.5 hours in
duration—often affords less restoration than equivalent-duration sleep in
a bed (Rosekind et al., 2000; Roach et al., 2010).
Although nap sleep obtained while sitting has some benefit for reduc-
ing fatigue, studies have found that sleeping while sitting upright (with
the requirement of at least partial antigravity posture) results in less sleep
time and poorer sleep quality than sleeping in a semirecumbent position
(Nicholson and Stone, 1987). Sleeping in a semirecumbent position with
ambient noise results in more light sleep and less deep sleep than sleep-
ing supine without noise (Dinges et al., 1981). Since pilot commuting can
involve sleep opportunities while seated or semirecumbent (e.g., in a car,
bus, or plane during a commute), the recovery potential of such sleep may
OCR for page 84
84 THE EFFECTS OF COMMUTING ON PILOT FATIGUE
be of less value than an equivalent period of sleep in a bed. These findings
make it difficult to estimate the value of naps and longer sleeps obtained
by pilots during commutes to work.
Body posture and ambient noise level, as well as physical comfort, are
not the only factors that can influence the benefits of nap sleep. Although
napping has been shown to be an effective technique for restoring alertness
and performance during periods of continued wakefulness, it is the timing
and length of a nap, along with the timing of the nap in the 24-hour pe-
riod (i.e., when it occurs relative to circadian phase), that also moderates
the benefits of napping for performance (Dinges et al., 1987; Webb, 1987;
Rogers et al., 1989; Bonnet, 1991; Rosa, 1993; Matsumoto and Harada,
1994; Vgontzas et al., 2007; Caldwell et al., 2009). Naps taken following
a prolonged period of wakefulness or during the habitual sleep period (and
in the window of circadian low) can be associated with more severe and
prolonged sleep inertia, which is a period of grogginess and performance
deficits immediately after awakening (Jewett et al., 1999b; Tassi and Muzet,
2000). After the sleep inertia dissipates—in 20 minutes to 2 hours, depend-
ing on the degree of prenap sleep deprivation and the timing of the nap—
the fatigue-reducing benefits of the nap occur, but they will dissipate faster
than the benefits of a full nocturnal sleep period.
FATIGUE MANAGEMENT TECHNOLOGIES
Recognition of the complex nature of the multiple interacting factors
that influence the buildup and reduction of fatigue as a state that can affect
performance has been at the core of the development and application of
various fatigue management technologies. The science of fatigue manage-
ment has developed rapidly over the past decade in civilian transporta-
tion sectors, with much of the applied research sponsored originally by
the military, where sustained and continuous operations pose acute and
chronic fatigue-related challenges. There are now several well-documented
candidate technologies for managing fatigue and its negative effects on
performance. These fall into two broad categories: fatigue-detection tech-
nologies and mathematical models of fatigue risk. Substantial progress has
been made in each of these areas.
Development of technologies in the first category—fatigue-detection
technologies for management of the fatigue risk—has been of interest to
transportation modalities in particular for the past 10-15 years, especially
in motor vehicle operations (Balkin et al., 2011). These technologies include
development of relatively unobtrusive ways to determine an operator’s level
of alertness or performance during duty, as well as devices that predict fa-
tigue in advance of a work cycle or trip (Balkin et al., 2004, 2011; Basner
and Dinges, 2011). A recent review of fatigue-detection technologies orga-
OCR for page 85
85
SLEEP, WAKEFULNESS, CIRCADIAN RHYTHMS, AND FATIGUE
nized them into four categories (Balkin et al., 2011): (1) Fitness-for-duty
tests that are designed to assess whether operators have sufficient alert-
ness/performance capacity prior to a work cycle or duty period; (2) online
operator monitoring technologies that are designed to provide real-time
monitoring of an operator’s physiological or behavioral state during work;
(3) performance-based monitoring systems that are designed to continuously
track operational performance to detect operator conditions/behaviors that
can lead to reduced safety; and (4) embedded or secondary-task technologies
that are designed to monitor and/or enhance operator performance/alertness
by modulating the amount of stimulation provided through a secondary
task. Although some of the fatigue-detection technologies have been studied
in operational environments and shown promise—especially in combination
(e.g., Dinges et al., 2005), there remain important unresolved questions and
limitations regarding the validity and reliability of their use relative to actual
work performance, and their acceptance by operators and industries (Dinges
and Mallis, 1998; Balkin et al., 2011). There is also a need to determine how
such technologies could be used most effectively in fatigue management in
commercial aviation, especially as it relates to commuting.
The second major category of fatigue management technologies con-
sists of mathematical models of fatigue or the risk fatigue poses to safety.
The models make predictions about the likelihood of fatigue as a risk to
performances using information on duty time and scheduling, sleep quan-
tity and quality, circadian and time-zone information, and other variables
(for reviews see Mallis et al., 2004; Dawson et al., 2011). Over the past
decade, a number of federal agencies have supported the continued devel-
opment and evaluation of mathematical models of fatigue risk (Jewett et
al., 1999a; Neri, 2002). A workshop sponsored by the U.S. Department of
Defense, the U.S. Department of Transportation, and the National Aero-
nautics and Space Administration provided an opportunity to conduct an
initial evaluation and comparison of seven of these mathematical models
from the United States, Europe, and Australia (Mallis et al., 2004). Al-
though predictions of performance were promising, the evaluation showed
that further research was needed to demonstrate the models’ validity and
reliability using real-world data, and that the models could not make reli-
able predictions of group performance risks from fatigue over multiday
schedules (Dinges, 2004; Van Dongen, 2004). In recent years, some of the
models have undergone further improvements in the accuracy of their pre-
dictions of both the basic dynamics of chronic sleep restriction in relation
to fatigue (e.g., McCauley et al., 2009), and in relation to prediction of
accidents (e.g., Hursh et al., 2008).
The potential for practical application of the mathematical models
in the commercial aviation context—and particularly in relation to pilot
commuting—has not yet been determined. A recent review of eight of the
OCR for page 86
86 THE EFFECTS OF COMMUTING ON PILOT FATIGUE
BOX 4-1
Risk Factors for Fatigue-Related Errors and Accidents
Risks of fatigue-related errors and accidents stem from multiple interrelated
and interacting aspects of work, rest, and sleep. These include but are not limited
to:
1. duration of work periods within a single day and over time,
2. time of day at which work occurs,
3. variation in the timing of work within and between weeks,
4. duration of sleep obtained on work days and on nonwork days,
5. frequency and duration of days off from work,
6. different vulnerabilities of workers to fatigue from these factors, and
7. volume and intensity of work.
SOURCE: Institute of Medicine (2009, pp. 218-219) citing the works of Dinges (1995), Rosa
(2001), Drake et al. (2004), Folkard et al. (2005), and Van Dongen (2006).
mathematical models of fatigue in work settings concluded that although
the models are intended to provide quantitative information on the likely
average level of fatigue risk associated with a given pattern of work and
sleep, there is considerable individual variability attributable to personal
biology and task variables not included in current models (Dawson et al.,
2011; see also Van Dongen et al., 2004). The review also concluded that
given the current limitations of the fatigue models, they may be most useful
as one element in a fatigue risk management system (Dawson et al., 2011).
Considerable research is needed to address how to use these models, and
other knowledge in the design and implementation of staffing and work-
scheduling programs in order to minimize fatigue (see National Research
Council, 2007; Horrey et al., 2011).
The issue of fatigue in safety-sensitive work operations cuts across
many industries and has been addressed broadly in the scientific literature.
The combination of work demands, sleep restriction, and circadian factors
can negatively affect alertness, performance, speed, accuracy, and central
nervous system functioning (Cabon et al., 1993; Goel et al., 2009b): see
Box 4-1. The next chapter looks more closely at the potential for pilots’
commuting patterns to affect their risk of fatigue.
