t·,
Sleep, 19(10):763-766
© 1996 American Sleep Disorders Association and Sleep Research Society
Sleepiness and Safety
Effect of Sleep Deprivation on Driving Safety In Housestaff
Carole L. Marcus and Gerald M, Loughlin
The Eudowood Division of Pediatric Respiratory Sciences, fohns Hopkins University, Baltimore, Maryland, U,S,A,
Summary: Sleep deprivation is known to affect driving safety. Housestaff (HS) are routinely sleep-deprived when
on call. We hypothesized that this would affect their driving, We therefore administered questionnaires regarding
driving to 70 pediatric HS, who were on call every fourth night, and to 85 faculty members (FAC), who were rarely
disturbed at night. HS were questioned about events during their residency, and FAC were questioned about events
during the preceding three years. There was an 87% response rate for each group. HS slept 2.7 :t 0.9 (SD) hours
when on call vs 7.2 :t 0.8 hours when not on call (p < 0.001). 44% of HS had fallen asleep when stopped at a
light, vs 12.5% FAC (p < 0.001). 23% of HS had fallen asleep while driving vs. 8% FAC (ns). A total of 49% of
HS had fallen asleep at the wheel; 90% of these events occurred post-call. In contrast, only 13% of FAC had fallen
asleep at the wheel (p < 0.001). HS had received a total of 25 traffic citations for moving violations vs. 15 for
FAC and were involved in 20 motor vehicle accidents vs. 11 for FAC. One traffic citation clearly resulted from HS
falling asleep at the wheel vs. none for FAC. We conclude that HS frequently fall asleep when driving post-call.
We speculate that current HS work schedules may place some HS at risk for injury to themselves and others. Further
study, using prospectively objective measures is indicated. Key Words: Sleep deprivation-Residency-Automobile-Motor vehicle accident.
In recent years, there has been a dramatically increased awareness of the relationship between sleepiness and driving risk (1). Patients with obstructive sleep
apnea (1), narcolepsy (2), and other causes of excessive
daytime sleepiness (3) are all at increased risk for motor
vehicle accidents. Estimates of the effects of sleepiness
on driving vary. However, a recent British study, based
on police databases and on-the-spot interviews, suggested that as'many as 16% of motor vehicle accidents
(MVAs) were due to drivers falling asleep at the wheel
(4). In 1994, MVAs resulted in 42,170 deaths in the
U.S.A. (5). Thus, more than 6,000 deaths may have
been sleep-related. In addition to the human toll, MVAs
result in a financial toll. In 1988, MVAs cost the United
States $70.2 billion per annum (6). This suggests that
sleep-related MVAs may have cost the United States as
much as $11 billion per annum.
One population group at risk for excessive daytime
sleepiness are medical housestaff (HS). HS frequently
suffer from both sleep deprivation and sleep fragmentation (5) when on call. Both sl~.ep deprivation and
sleep disruption affect vigilance, cognitive functions,
reaction time, and mood (7-11), all of which could be
expected to have an impact upon driving skills.
Accepted for publication September 1996.
Address correspondence and reprint requests to Carole L. Marcus,
M.D., Division of Pediatric Pulmonology, Park 316, Johns Hopkins
Hospital, 600 N. Wolfe St., Baltimore, MD 21287-2533, U.S.A.
We hypothesized that sleep-deprived HS were at increased risk of hazards when driving. In order to evaluate this, we administered a questionnaire regarding
driving habits to HS, as well as a control group of
pediatric faculty (FAC). This was a pilot study prompted by our interest in factors that influence housestaff
performance.
METHODS
A self-administered questionnaire was mailed to pediatric residents and full-time faculty at our institution.
Faculty routinely doing in-house call (i.e. emergency
room and neonatology attendings) were not included.
The questionnaire was anonymous. However, in order
to optimize the response rate, subjects were also sent
a numbered coverletter, to be returned separately. A
second questionnaire was sent to those who did not
respond. General reminders were also issued during
pediatric' conferences. The questionnaire included
structured questions related to being on call, participation in other types of nocturnal work, falling asleep
while driving, traffic citations for moving violations,
and MVAs. In addition, space was available for comments regarding citations and MVAs. The time period
covered by the questionnaire was during residency (for
HS) and during the past three years (for FAC). Post-
763
C. L. MARCUS AND G. M. LOUGHLIN
764
TABLE 1.
Study group
HS
FAC
shift. HS averaged 2.7 ::±:: 0.9 hours of sleep when on
call and 7.2 ::±:: 0.8 hours of sleep when not on call (p
< 0.001). FAC slept 6.5 ::±:: 0.8 hours a night, which
was less than HS not on call (p < 0.00l).
61
72
Age (years)
range:
29 :!: 2
26-35
46:!: 10
29-73
Female
44 (72)
30 (42)
Falling asleep at the wheel
Marital status
Single
Married
Divorced
32 (53)
27 (44)
2 (3)
10 (14)
58 (81)
4 (6)
Postgraduate year
1
2
3
4 (chief resident)
15 (25)
18 (30)
27 (44)
1 (2)
Forty-four percent of HS reported having fallen
asleep while stopped at a red light. All of these episodes occurred when the HS were driving home postcall. One HS member wrote that she routinely used
her emergency brakes when stopped at a light because
of her sleepiness post-call. In contrast, only 12.5% of
FAC had fallen asleep while stopped at a light (p <
0.001). Twenty-three percent of HS had fallen asleep
while driving, with 71 % of the episodes occurring
when the HS were driving home post call, as compared
to 8% of FAC (p = 0.07; ns). A total of 49% of HS
had fallen asleep at the wheel (during residency) vs.
13% FAC (during the past 3 years) (p < 0.001).
n
All data displayed as n (%), except for age, which is displayed as
mean :!: SD. HS were significantly younger, more likely to be female, and more likely to be single than FAC (p < 0.001 for all
parameters).
call was defined as the calendar day (until midnight)
following the on-call night. Most HS drove home in
the late afternoon or early evening post-call.
Traffic citations
Sixty-one (87%) of the 70 pediatric HS and 74
(87%) of the 85 FAC responded. Two FAC were excluded from analysis, one because he moonlighted at
least one night a week and one because he stated that
he underwent polysomnography every other month
and, therefore, presumably had a sleep disorder (further information was unavailable). Thus, 72 FAC were
evaluated. Three FAC moonlighted 1-2 nights a
month; they were not excluded due to the infrequent
nature of the nighttime work. No faculty performed
other types of night work.
Fifteen (25%) of HS had received traffic citations
for moving violations. A total of 25 citations (range
0-3 per person) had been received. Thirteen (18%) of
FAC had received traffic citations for moving violations, with a total of 15 citations (range 0-2 per person). HS received significantly more citations than
FAC (p < 0.005). The majority of the citations were
for speeding (84% of HS and 60% of FAC). Twenty
percent of citations were issued when HS were postcall, 32% on other days, and 4% when on vacation.
However, the timing of the ticket was not specified in
44% (because of the layout of the questionnaire, some
HS who received multiple citations only recorded the
time for one event). All HS and FAC stated that they
had not consumed alcohol prior to the incident, and
they had not used other drugs that could affect their
driving ability. Forty-seven percent of HS and 18% of
FAC who had received citations had a history of falling asleep at the wheel. One HS citation was clearly
related to sleepiness-a resident reported receiving a
ticket when he fell asleep and went through a red light
while driving home at 7 p.m. post-call.
Study group
Motor vehicle accidents
Details of the study group are shown in Table 1. As
expected, FAC were significantly older, more likely to
be male, and more likely to be married than HS. All
residents were on call every fourth night (except during one month of elective during the residency), during
which they would work for approximately 33 hours a
Fifteen (25%) of HS had been involved in MV As. A
total of 20 MVAs had occurred (range 0-3 per person).
Thirty percent of MVAs occurred when HS were driving to work, 45% driving from work, 10% while on
vacation, and 15% unspecified. Forty percent of MVAs
occurred when HS were post-call, 30% on other days,
Data analysis
All results are expressed as mean ::±:: SD unless otherwise specified. Comparison of parameters for the
same individual were compared using the paired t test.
Parameters were compared between HS and FAC using Chi-squared analysis or independent t test. A p
value of 0.05 was considered significant.
RESULTS
Sleep, Vol. 19, No. 10, 1996
.
\.
SLEEP DEPRIVATION AND DRIVING SAFETY
15% when on vacation or elective without call, and
15% unspecified. There were no injuries or fatalities.
All HS stated that they had not consumed alcohol prior
to the MVA, and they had not used other drugs that
could affect their driving ability. Sixteen MVAs involved two-car collisions, three involved single-vehicle
roadway departures (in one case, it was raining), and
details were not provided for one MVA. Forty-seven
percent of the HS who had been involved in MVAs had
a history of having fallen asleep at the wheel. No MVA
was clearly related to sleepiness, although one was suspicious. In that case, the resident who had mentioned
the need to use her parking brakes because she fell
asleep at red lights, and who also stated that she fell
asleep while driving, reported that she drove into a
roadside barrier, then ricocheted and hit a parked car.
Ten (fourteen percent) of FAC had been involved in
an MVA, with a total of 11 MVAs (range 0-2). FAC
were involved in significantly fewer MVAs than HS (p
< 0.02). Fifty-five percent of MVAs occurred when
FAC were driving to work, 27% driving from work,
and 11 % on weekends. There were no fatalities. One
pedestrian sustained a fractured ankle when hit by a car
driven by a FAC member. All FAC denied alcohol consumption prior to the MVA. One FAC was receiving
atenolol and thought that this impaired his reaction
time; no other FAC thought that drugs contributed to
the MVA. Nine MVAs involved two-car collisions, one
involved a vehicle/pedestrian collision, and one consisted of a single-vehicle roadway departure in bad
weather (rain). Thirty percent of the FAC who were
involved in MVAs had a history of having fallen asleep
at the wheel. No FAC MVA was clearly related to
sleepiness.
DISCUSSION
This study has shown that HS frequently fall asleep
at the wheel when driving post-call. HS received more
traffic citations and were involved in more MVAs on
their post-call days than on other work days. One traffic citation, and possibly one MVA, occurred when HS
fell asleep at the wheel post-call.
Because of the retrospective nature of this study, it
was not possible to determine whether traffic citations
and MVAs were directly related to sleepiness. HS were
involved in more MVAs than FAC, but this may be due
to the fact that they were younger than FAC and, therefore, at increased risk for MVAs (12). Nevertheless,
several factors suggest that sleepiness post-call did play
a role in HS traffic citations and MVAs. As HS were
on call every fourth night, the predicted rate of citations
and MVAs post-call would be 25%. Excluding events
that were unspecified or occurred when HS were on
vacation, 39% of HS citations and 57% of HS MVAs
765
occurred when HS were post-call, exceeding the predicted rate (p < 0.05 and 0.001, respectively). Vacation
time was excluded from this analysis, as driving patterns on long vacation trips may differ from routine
driving habits. One HS citation, but no FAC citations,
was clearly related to sleepiness. It is probable that one
MVA (involving two collisions) was also due to a HS
member falling asleep at the wheel. That incident involved a HS member with a history of falling asleep
frequently when driving and involved the driver going
off the road. Sleep-related MVAs are more likely to
involve drivers drifting off the road (4,13).
On average, HS slept less than three hours when on
call, which is far less than the five hours of sleep
thought necessary to preserve cognitive and psychomotor skills (10). Not only do HS sleep less when on
call, but the sleep that they do have is fragmented due
to frequent pages and telephone calls. Frequent sleep
disruption can be expected to result in decreased slow
wave and rapid eye movement sleep, resulting in impaired cognitive function and increased sleepiness similar to the effects of total sleep deprivation (7). It should
be noted that the results of sleep deprivation found in
our study may be even more marked in those programs
which have call every third or every second night.
There were some differences between the HS group
and the control group. We chose faculty members as
our control group because they were of similar socioeconomic class to the HS and had many of the same
work-related stresses. As expected, the FAC were older than the HS, and middle-aged drivers are less likely
than young adults to be involved in MVAs (12). We
cannot exclude the possibility that some FAC had
sleep apnea, although statistically it is unlikely that a
substantial number would have had this condition (14).
On the other hand, the HS group had a greater proportion of females, who are less likely than males to
be involved in MVAs (6). The incidence of falling
asleep at the wheel for FAC was 12.5%, which is similar to the reported incidence of less than 20% (for
lifetime occurrences) in the literature (2,6,15,16), suggesting that they were a representative control group.
The period of assessment varied for different respondents. Faculty were questioned about the preceding three years, whereas housestaff were questioned
about their residency period. Thus, the period for housestaff varied from one to three years. This may have
resulted in an under-representation of HS events.
This study has the limitations of a retrospective
questionnaire study. Respondents completed an anonymous questionnaire, and it is possible that HS provided biased responses. We consider this unlikely as
HS were not informed of the precise objectives of the
study and were unlikely to obtain any immediate benefit from misrepresenting information. However, furSleep, Va!. 19, No. 10, 1996
766
C. L. MARCUS AND G. M. LOUGHLIN
ther prospective studies, using objective measures rath- jective measurements of sleep and of driving skill, be
performed.
er than self-report, are indicated.
Many studies have shown that call-related sleep deAcknowledgements: Dr. Marcus was supported in part
privation in HS results in impaired function, such as by CAP grant number RR-00052, Pediatric Clinical Reimpaired cognitive and psychomotor skills, increased search Center, The Johns Hopkins Hospital, Baltimore, MD
sleepiness, delayed reaction time, and an impaired and the American Lung Association of Maryland. Dr.
ability to interpret electrocardiograms (8-1l). Sleep Loughlin is supported, in part, by grant number RR-00052,
deprivation in this group has also been shown to affect Pediatric Clinical Research Center. The authors wish to
mood, resulting in increased fatigue, sadness, hostility, thank the pediatric housestaff and faculty at Johns Hopkins
anxiety, and confusion post-call (8,9). In addition, it Hospital for their participation in this study. We thank Kathhas been speculated that the sleep deprivation in HS leen Bukowski for secretarial assistance and Steven Goodman, M.D., Ph.D., for statistical assistance.
contributes to the increased incidence of depression,
marital discord, and even suicidal ideation that have
REFERENCES
been demonstrated in this population (8,17-19). In this
study we have demonstrated that driving hazards are
1. American Thoracic Society. Consensus statement on "Sleep apyet another risk that current training schedules impose
nea, sleepiness, and driving risk. Am J Respir Crit Care Med
1994;150: 1463-473.
on young physicians in training.
2. Broughton R, Ghanem Q, Hishikawa Y, Sugita Y, Nevsimalova
Some countries, such as Britain, have limited junior
S, Roth B. Life effects of narcolepsy in 180 patients from North
doctors' hours because of the effects of sleep depriAmerica, Asia and Europe compared to matched controls. Can
J Neurol Sci 1981;8:299-304.
vation (20). Following the Libby Zion case, in which
3. Aldrich MS. Automobile accidents in patients with sleep disHS sleep deprivation was thought to contribute to the
orders. Sleep 1989;12:487-494.
4. Horne JA, Reyner LA. Sleep related vehicle accidents. BMJ
death of a patient, New York City imposed limits on
1995;4:565-7.
the hours of doctors in training (21). However, in most
5. Guyer B, Strobino DM, Ventura SJ, Singh GK. Annual sumcases there are no limits on doctors hours, even though
mary of vital statistics-1994. Pediatrics 1995;96:1029-39.
6. Leger D. The cost of sleep-related accidents: a report for the
limits are placed on other professionals, such as pilots
National Commision on Sleep Disorders Research. Sleep 1994;
and truck drivers, in whom vigilance is important.
17:84-93.
What is the reason for this? Stated reasons include
7. Bonnet MH. Effcct of sleep disruption on sleep, performance,
and mood. Sleep 1985;8:11-9.
maintaining continuity in patient care, instilling a
8. Asken MJ, Raham DC. Resident performance and sleep deprisense of responsibility in residents, and increasing the
vation: a review. J Med Educ 1983;58:382-8.
learning opportunities for residents. Others have com9. Friedman Re, Kornfeld DS, Bigger TJ. Psychological problems
associated with sleep deprivation in interns. J Med Educ 1973;
pared the system to a rite of passage (10,22). However,
48:436-441.
the probable reason for the persistence of the system 10. Hawkins MR, Vichick DA, Sisby HD, Kruzich DJ, Butler R.
Sleep and nutritional deprivation and performance of house ofis economic. In New York City, it was estimated that
ficers. J Med Educ 1985;60:530-5.
limiting housestaff to 12-16 hour shifts (in association
11. Lingenfelser Th, Kaschel R, Weber A, Zaiser-Kaschel H, Jawith increased attending physician coverage) would
kober B, Kuper J. Young hospital doctors after night duty: their
task-specific cognitive status and emotional condition. Med
cost the city $204,000,000 (21).
Educ 1994;28:566-72.
Ideally, HS hours should be limited for several rea- 12. Williams
AF, Carsten O. Driver age and crash involvement. Am
sons: to improve patient care, avoid the detrimental
J Public Health 1989;79:326-7.
effects of sleep deprivation on HS personality and af- 13. Pack AM, Cucchiara A,'Schwab W, Rodgman E, Pack AI. Characteristics of accidents attributed to the driver having fallen
fect, and, now, to prevent driving injuries related to
asleep. Sleep Res 1994;23:141 (abstract).
sleep deprivation. In the interim, it is incumbent upon 14. Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The
occurrence of sleep-disordered breathing among middle-aged
sleep practitioners to educate staff about the risks of
adults. N Engl J Med 1993;328:1230-5.
driving when sleep deprived so that alternative mea- 15. Martikainen K, Urponen H, Partinen M, Hasan J, Vuori I. Daytime sleepiness: a risk factor in community life. Acta Neural
sures (such as using public transportation or taking a
Scand 1992;86:337-41.
nap before driving home) can be entertained (1).
16. Broughton RJ, Guberman A, Roberts J. Comparison of the psyIn conclusion, we have demonstrated that HS have
chosocial effects of epilepsy and narcolepsy/cataplexy: a controlled study. Epi/epsia 1984;25:423-433.
an increased incidence of falling asleep at the wheel
17. Editorial. The dangers of not going to bed. Lancet 1989;i:138-9.
when driving home post-call, and this probably re- 18. McCall TB. The impact of long working hours on resident physulted in increased traffic citations and motor vehicle
sicians. New Engl J Med 1988;12:775-8.
accidents. We recommend that training programs begin 19. McCue JD. The distress of internship. New Engl J Med 1985;
312:449-452.
to review HS scheduling with an eye towards reducing 20. Editorial. Junior doctors' hours. Lancet 1991;337:1538.
extended periods of sleep deprivation and fragmenta- 21. Asch DA, Parker RM. The Libby Zion case. New Engl J Med
1988;18:771-5.
tion. We recommend that prospective studies, using a 22. Cousins
N. Internship: preparation or hazing. JAMA 1981;245:
larger study group from several institutions, with ob377.
Sleep. Vol. 19. No. 10. 1996
/