Sleep and Biological Rhythms 2007; 5: 226–230 doi:10.1111/j.1479-8425.2007.00278.

COMMENTS AND OPINIONS

Do medical students in Hong Kong have enough sleep?

Lok-Lam E HUEN, Tze-Wai G CHAN, Wai-Man M YU and Yun-Kwok WING

Department of Psychiatry, Prince of Wales and Shatin Hospital; Chinese University of Hong Kong, Hong Kong SAR

Abstract
This was a cross-sectional study of sleep patterns among Hong Kong medical students. A represen-
tative number of 419 medical students with a mean age of 21.2 (SD = 2.1, range 17–32) reported a
mean nocturnal sleep of 6.6 h (SD = 1.2 h). About 70% reported having sleep deprivation without
significant gender and age differences. Self-perceived stress and long hours of study were found to
be the major risk factors for sleep deprivation whereas longer nocturnal sleep duration was its
protective factor. Sleep deprivation was common among medical students in this cosmopolitan Asian
city. Effective stress management and sleep education should be included in the training of medical
doctors.

Key words: education, medical students, sleep deprivation.

INTRODUCTION students who are future health-care providers. However,

there are few sleep studies on medical students. The
Studies show that sleep deprivation is common among current study aimed to investigate the sleep and daytime
university students in both the East and the West. A activity pattern, in terms of hours and the risk factors for
substantial number of university students (24% in the sleep deprivation, among medical students in Hong
UK, 30% in Korea, and 49% in Taiwan) sleep less than Kong, a cosmopolitan city embedded in both western
7 h a night, especially on weekdays.1–3 Poor subjective and Asian cultures.
health status,2,4 poor lifestyles, such as using visual
media for long stretches of time,4,5 abuse of alcohol1 and
coffee,2 lack of exercise,4 eveningness chronotype5 and
perceived study stress4 are found to affect sleep quantity METHODS
and quality in students. A recruitment email was sent from the investigators to
Sleep problems may be intensified in medical stu- all medical students of the Chinese University of Hong
dents as they have longer years of training before gradu- Kong between April and May 2004. This university is
ation than non-medical students and require more one of the major medical training centers in Hong Kong
rigorous training that includes overnight internships.6 (there are two medical schools with a similar number of
Since sleep deprivation leads to poor outcomes such as medical students in Hong Kong) with 764 students
impaired psychological status and cognitive perfor- currently admitted. The email stated clearly the purpose
mance, it is necessary to improve the sleep of medical of the study and was attached with a set of self-reported
questionnaires. As a formal consent to the study, the
students were requested to return electronically the
Correspondence: Dr Yun-Kwok Wing, Department of completed questionnaires with their name and student
Psychiatry, Prince of Wales and Shatin Hospital, Chinese identity number, after the confidentiality of the study
University of Hong Kong, Shatin, Hong Kong SAR. Email: and voluntariness of their participation were stressed. In
[email protected] addition, we randomly selected 25 students to wear
Received: 18 August 2006 / Accepted: 14 February 2007. an actiwatch (Mini Mitter, Bend, OR, USA,) for 3

226 © 2007 The Authors

Journal compilation © 2007 Japanese Society of Sleep Research
 The sleep pattern of medical students in Hong Kong

consecutive days. This would enable us to verify the Table 1 Demographic data, sleep patterns and daily activity
self-reported sleep duration with objective measure- of medical students
ments. The study was approved by the university ethical No. of students
committee. (n = 419) (%)
Four hundred and nineteen medical students (166
men vs 253 women, mean age 21.2 ⫾ 2.1) consented to Demographic data
Sex
the study. The response rate was 54.8%. The gender and
Male 166 (39.6)
age ratios of the participating students did not differ Female 253 (60.4)
significantly from that of local medical students in Years
general. 1 120 (28.6)
2 83 (19.8)
3 46 (11.0)
Instruments 4 98 (23.4)
5 72 (17.2)
The set of questionnaires contains two Chinese version Age (years: mean ⫾ S.D.) 21.2 ⫾ 2.1
instruments for self-completion, including the sleep Sleep parameters (hours: mean ⫾ S.D.)
questionnaire7 and the morningness–eveningness ques- Weekdays sleep duration 6.6 ⫾ 1.2
tionnaire (MEQ).8 The former questionnaire is com- Weekend sleep duration 8.7 ⫾ 1.6
posed of structured questions designed by one of the Daily activities (hours)
investigators (YK Wing) for epidemiological sleep stud- Study 9.3 ⫾ 2.7
Entertainment 2.4 ⫾ 1.7
ies.7 These questions included demographic data,
Using the Internet 1.9 ⫾ 1.7
bedtime duration during weekdays and the weekend, Talking on the phone 0.6 ⫾ 0.6
self-perceived sleep deprivation, time allocated for daily Chronotypes
activities such as study (class and revision), use of visual Neither 280 (66.8)
media (the Internet and electronic games), doing physi- Eveningness 77 (18.4)
cal exercise, engaging in non-physical leisure activities Morningness 49 (11.7)
and chatting on the phone with friends, together with Additional questions
questions on self-perceived stress over study and the Perceived higher stress than other 334 (79.7)
adequacy of their sleep-related knowledge. students
The MEQ is a 19-item self-assessment questionnaire Adequate sleep hygiene knowledge 191 (45.6)
determining morningness and eveningness in human
circadian rhythms.8 It classifies individual sleep patterns
into an eveningness type, morningness type or neither
types. The questionnaire was translated into Chinese RESULTS
and then back-translated by the investigator.9 The average sleep duration of the medical students
during weekdays and weekend, and their demographic
data including their age, gender and academic year are
STATISTICAL ANALYSIS shown in Table 1. These students reported a mean noc-
turnal sleep of less than 7 h during weekdays and 66%
Descriptive statistics were obtained for the sleep pattern, reported self-perceived sleep deprivation (Table 1). The
weekday bedtime duration and self-perceived sleep validity of the self-report sleep duration was further
deprivation. c2 analysis was used for the categorical supported by the actiwatch measurement. The average
responses while a t-test or ANOVA was used for continu- actual sleep time as measured by the actiwatch was
ous responses. A post hoc Bonferroni t-test was used to 6 h 2 min ⫾ 59 min. It correlated strongly with self-
delineate group differences in academic years. The asso- reported sleep duration (6 h 5 min ⫾ 55 min, r = 0.68,
ciation of potential risk factors was determined by mul- P = 0.001).
tiple logistic regressions with a backward stepwise Those students who perceived themselves suffering
procedure. A two-tailed 5% level of significance was from sleep deprivation reported significantly shorter
used for all the statistical tests. All statistical procedures sleep duration than medical students who were not
were conducted with Statistical Package System Soft- deprived of sleep during weekdays but were on
ware version 14.0 (SPSS, Chicago, IL, USA). weekends (Table 2). No significant age or gender

© 2007 The Authors 227

Journal compilation © 2007 Japanese Society of Sleep Research
LLE Huen et al.

Table 2 Comparison between sleep deprivation and non-sleep deprivation group

Normal sleeper Sleep deprived
(n = 138) (%) (n = 277) (%) P-value
Demographic data
Age (year: Mean ⫾ SD) 21.5 ⫾ 2.1 21.1 ⫾ 2.0 NS
Sex
Male 58 (42.0) 108 (39.0) NS
Female 80 (58.0) 169 (61.0)
Year
1 30 (21.7) 89 (32.1)
2 32 (23.2) 48 (17.3)
3 11 (8.0) 35 (12.6) 0.038
4 41 (29.7) 57 (20.6)
5 24 (17.4) 48 (17.3)
Sleep parameters (hour: mean ⫾ SD)
Weekdays sleep duration 7.2 ⫾ 1.4 6.3 ⫾ 1.0 <0.001
Weekend sleep duration 8.6 ⫾ 1.8 8.8 ⫾ 1.5 NS
Chronotypes
Eveningness 15 (11.3) 61 (22.7) 0.006
Morningness 24 (18.0) 24 (8.9) 0.008
Additional questions
Perceived higher stress as a medical student 91 (66.9) 239 (88.5) <0.001
Perceived that sleep might affect academic performance 113 (83.1) 252 (94.0) <0.001
Adequate sleep hygiene knowledge 73 (53.7) 116 (43.0) 0.041
Daily activities (hours: mean ⫾ SD)
Study 8.6 ⫾ 2.7 9.7 ⫾ 2.7 <0.001
Entertainment 2.8 ⫾ 1.8 2.2 ⫾ 1.6 <0.001
Using the Internet 1.6 ⫾ 1.3 2.0 ⫾ 1.9 0.013
Talking on the phone 0.7 ⫾ 0.6 0.6 ⫾ 0.6 NS

difference was found in the self-reported sleep depriva- (0.43 ⫾ 0.47 h vs 0.32 ⫾ 0.43 h; t = 2.65; P = 0.008)
tion or the duration of sleep during weekdays. However, than the female students. Students in different academic
academic years were significantly associated with self- years also reported significant differences in the length
reported sleep deprivation (Table 2) and sleep duration of time spent on studying (F = 10.44, d.f. = 4,
(F = 2.67, d.f. = 4, P = 0.032). Post hoc comparisons P < 0.0001), engaging in non-physical leisure activities
showed that first-year students reported shorter noctur- (F = 2.94, d.f. = 4, P = 0.02) and chatting on the tele-
nal sleep duration than second- and fourth-year stu- phone (F = 2.61, d.f. = 4, P = 0.035). Post hoc compari-
dents. sons revealed that students in the fifth year significantly
For daytime activities, students spent most of their reported spending the longest amount of time in study,
time studying, followed by using visual media, engaging while third year students reported spending signifi-
in non-physical leisure activities, chatting with friends cantly longer time than fourth year students on non-
on the phone and doing physical exercise (Table 1). The physical leisure activities. But no significant post hoc
self-perceived sleep-deprived students reported signifi- academic year difference was found between them in
cantly longer hours of study and use of visual media, chatting on the phone.
respectively. They also reported a marginally significant About 80% of them perceived themselves as experi-
shorter exercise time (Table 2). encing more stress than non-medical students and
Male students reported spending a significantly 45.6% as having an adequate knowledge of sleep
shorter time in studying (8.95 ⫾ 2.84 h vs hygiene. Significantly more self-perceived sleep-
9.54 ⫾ 2.64 h, t = -2.14, P = 0.033) but longer hour deprived students reported having more stress than
on visual media usage (2.60 ⫾ 2.28 h vs 1.87 ⫾ 1.61 h; other students. In addition, significantly less of them
t = 3.60; P < 0.0001) and on physical activities perceived themselves as having adequate sleep hygiene

228 © 2007 The Authors

Journal compilation © 2007 Japanese Society of Sleep Research
 The sleep pattern of medical students in Hong Kong

knowledge (Table 2). In addition, significantly more of Hours sleeping were sacrificed for several reasons.
the fifth-year students perceived themselves as experi- In this study, the medical students spent an average of
encing more stress than non-medical students half a day (9.3 ⫾ 2.7 h) on study. Similar to previous
(c2 = 9.74, d.f. = 4, P = 0.045) and more fourth year studies, more sleep-deprived students reported having
students perceived themselves as having adequate sleep unhealthy lifestyles, including longer hours of study and
hygiene knowledge (c2 = 29.57, d.f. = 4, P < 0.0001). use of visual media, and shorter exercise time. Long
Those who reported having more stress had a signifi- study hours may have generated a heavy cognitive
cant longer duration of studying (9.53 ⫾ 2.82 h vs burden, whereas long visual media usage, including
8.42 ⫾ 2.05 h, t = 3.87, P < 0.0001), shorter sleep using the Internet and playing electronic games, though
duration (6.55 ⫾ 1.13 h vs 6.93 ⫾ 1.64 h, t = -2.38, entertaining, may have increased the tiredness of stu-
P = 0.018) and shorter time spent in non-physical dents, and the enhanced immobility of students might
activities (0.74 ⫾ 0.77 h vs 1.16 ⫾ 1.62 h, t = -2.14, worsen their sleep. However, physical exercise may
P = 0.035) than those who reported less stress. Those improve physical well-being and nocturnal sleep.
who reported having adequate sleep hygiene knowledge Most medical students believed that they experienced
did not differ in the duration of various daytime activi- more stress than other students. The medical curricu-
ties compared with those who reported not having an lum is tight and medical knowledge is enormously
adequate knowledge of sleep hygiene. changing.6 More of the stressed students reported sleep
About 70% of students were classified as being deprivation. Those who felt more stressed studied
neither morningness nor eveningness types but com- longer and had shorter hours of sleep and of non-
paratively more students belonged to the eveningness physical leisure activities. More of the fifth-year students
than the morningness type (Table 1). Significantly more reported feeling more stressed than the others and their
self-perceived sleep-deprived students belonged to the study hours were the longest among the five academic
eveningness type and significantly less of them belonged years. This may be related to the stress at the pending
to morningness type (Table 2). Significantly more of the graduation and final examination at the fifth year. They
fifth-year students (c2 = 14.56, d.f. = 4, P = 0.006) and may also start to adapt to the tradition of medical
more men (c2 = 6.00, d.f. = 1, P = 0.014) belonged to doctors, namely, long working hour and having little
the eveningness type. sleep. This has inbred in medical students as well as
Significant risk factors for sleep deprivation were physicians who regard sleep deprivation as a badge of
having more perceived stress than other students (odds honor or demonstration of professional dedication.10
ratio = 3.62, 95% CI = 2.01–6.52; P < 0.0001) and However, sleep hygiene knowledge may alleviate
longer study hours (odds ratio = 1.13, 95% CI = 1.02– sleep deprivation. Those who perceive themselves as
1.24; P = 0.017) while sleep duration was its protective having inadequate sleep hygiene knowledge reported
factor (odds ratio = 0.49, 95% CI = 0.38–0.63; more sleep deprivation. As the psychiatry training
P < 0.0001). module, which includes teaching sleep hygiene, was
arranged in the fourth year, it is not surprising that more
of the fourth-year students perceived themselves as
DISCUSSION having adequate sleep knowledge.
Similar to previous studies, our medical students also
Medical students in this study, similar to the university confirmed that the eveningness chronotype is associated
students in other studies in the Western and Asian with sleep deprivation. More of the fifth-year students
cultures,1–5 reported an average sleep of less than 7 h were classified as eveningness types. This indicated that
and the majority perceived themselves as suffering from they might have sacrificed their sleep at night for their
sleep deprivation. The sleep-deprived students slept heavy workload or their overnight shift, but were unable
much less than their counterparts during weekdays but to replenish it the next morning as they had immedi-
had similar hours of sleep on the weekend, so their ately afterwards resumed their work.6
heavier weekday sleep debt seemed not to be fully Sleep deprivation might create a vicious cycle,
replenished and might accumulate chronically. It was because its subsequent impaired daytime performance
not surprising that more first-year students felt that they might lead to heightened stress and perpetuate the
were sleep-deprived and reported a tendency to have problem. Long study hours seem inevitable for medical
shorter sleep, as they might be adapting to the university students, as they have rigorous training with endless
life.6 assessments, a large volume of material to learn and

© 2007 The Authors 229

Journal compilation © 2007 Japanese Society of Sleep Research
LLE Huen et al.

strong peer competition. However, their self-perceived REFERENCES

stress was a much stronger predictor than hours study-
ing for sleep deprivation in this study. Therefore, effec- 1 Webb E, Ashton CH, Kelly P, Kamali F. Alcohol and
tive stress management should be included in the drug use in UK university students. Lancet 1996; 348:
922–5.
medical training, apart from systemic education in sleep
2 Ban DJ, Lee TJ. Sleep duration, subjective sleep distur-
knowledge for these future clinicians. Longer sleeping bances and associated factors among university students
hours, which are the protective factor of sleep depriva- in Korea. J. Korean Med. Sci. 2001; 16: 475–80.
tion, and healthy lifestyles should also be promoted. 3 Tsai LL, Li SP. Sleep patterns in college students: gender
Special care of students with the eveningness chrono- and grade differences. J. Psychosom. Res. 2004; 56:
type may be needed. 231–7.
This study had some limitations. The accuracy of the 4 Pallos H, Gergely V, Yamada N, Miyazaki S, Okawa M.
findings of this self-reported study relied on the stu- Exploring the quality of sleep on long-term sojourn:
dents’ subjective accounts. In view of this, we con- international graduate students in Japan. Sleep Biol.
ducted a small-scale study to validate this self-reported Rhythms 2005; 3: 142–8.
sleep duration with objective actitwatch measurement. 5 Gaina A, Sekine M, Kanayama H, Sengoku K, Yamagami
T, Kagamimor S. Short-long sleep latency and associated
The result showed that the objective measurements
factors in Japanese junior high school children. Sleep
were highly correlated with self-reported sleep duration. Biol. Rhythms 2005; 3: 162–5.
In assessing their self-perceived stress compared to non- 6 Wong JGWS, Patil NG, Beh SL, Cheung EPT, Wong V,
medical students and adequate sleep hygiene knowl- Chan LC. Cultivating psychological well-being in Hong
edge, categorical scales were used. Further studies with Kong’s future doctors. Med. Teach. 2005; 27: 715–19.
quantifiable measures are recommended to delineate the 7 Wing YK, Li R, Lam CHC, Fong SY, Leung T. The
extent of stress and sleep knowledge. Longitudinal out- prevalence of narcolepsy among Chinese in Hong Kong.
comes of sleep deprivation among medical students Ann. Neurol. 2002; 51: 578–84.
regarding their academic performance and mental 8 Horne JA, Ostberg O. A self-assessment questionnaire to
health status will warrant further investigation. determine morningness–eveningness in human circa-
dian rhythms. Int. J. Chronobiol. 1976; 4: 97–110.9
9 Zhang B. Habitual short and long sleeper in middle-aged
ACKNOWLEDGMENTS Hong Kong Chinese: epidemiological, clinical, and poly-
somnographic study (Doctor of Philosophy Thesis).
This study was a selected study module project of Chinese University of Hong Kong, Hong Kong, 2005.
medical students of the Chinese University of Hong 10 Daugherty SR, Baldwin DC Jr. Sleep deprivation in
Kong. Special thanks to Professor Jim Horne for allow- senior medical students and first-year residents. Acad.
ing us translate his MEQ questionnaire. We wish to Med. 1996; 71: S93–5.
express our gratitude to the participants in our study.

230 © 2007 The Authors

Journal compilation © 2007 Japanese Society of Sleep Research