You are on page 1of 9

Journal of Psychosomatic Research 53 (2002) 601 609

Gender differences in insomnia a study in the Hong Kong


Chinese population
R.H.Y. Lia, Y.K. Winga,*, S.C. Hob, S.Y.Y. Fonga
a

Department of Psychiatry, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong, China
Department of Community and Family Medicine, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong, China

Abstract
Objective: To study the epidemiology of insomnia in the adult
Chinese population in Hong Kong and to examine the potential
gender-related demographic and lifestyle factors in insomnia.
Methods: A population study via random telephone survey with a
structured questionnaire was carried out for noninstitutionalized
Chinese adults aged 18 65 by trained lay interviewers. The
questionnaire included demographic data, sleep habits and
problems, insomnia symptoms and lifestyle questions. Results:
A total of 9851 subjects (46.4% male; 53.6% female) were
included in the final analysis. The overall prevalence of Hong
Kong Chinese as suffering from insomnia during the preceding
month (with a frequency of sleep disturbance of at least three times
per week) was 11.9% (95% CI 11.2 12.6), including difficulty in
initiating sleep (DIS) (4.5%; 95% CI 4.1 5.0), difficulty in
maintaining sleep (DMS) (6.9%; 95% CI 6.4 7.5) and early
morning awakening (EMA) (4.0%; 95% CI 3.6 4.4). Females
were about 1.6 times at higher risk for insomnia than males. The
prevalence of insomnia was also shown to increase with age.

Multivariate analysis showed that unemployment, lower economic


status, alcohol consumption, regular medication and psychiatric
disturbance were all associated with higher risks of insomnia in
both sexes. Furthermore, lower education level and being retired
was associated with a higher risk of insomnia in males, but being a
housewife, divorced/widowed, and complaining of a nocturnal
noisy environment were associated with a higher risk of insomnia
in females. Among all these factors, psychiatric disturbance was
the most influential risk factor for insomnia in both sexes. The
reasons for gender differences of insomnia may include their
differences in the prevalence of psychiatric morbidities, symptom
endorsement, gonadal steroids, sociocultural factors and coping
strategies. Conclusions: Overall, 11.9% of the Hong Kong
Chinese adult population complained of frequent insomnia in the
preceding month. There was a higher prevalence of insomnia in
females. Although there were common risk factors for insomnia in
both sexes, there existed gender-specific risk factors. D 2002
Elsevier Science Inc. All rights reserved.

Keywords: Chinese; Gender differences; Insomnia; Prevalence

Introduction
Gender differences have been reported across a number
of medical, psychiatric and sleep disorders [1 3]. The
search for any gender differences has a significant impact
not only on the etiological and pathophysiological implications, but also on the health care utilization, treatment and
prognosis of the disorders [1,2].
Studies, in general, demonstrated that there were gender
differences for insomnia with a higher prevalence in females
[4 8]. However, the factors contributing to these differ-

* Corresponding author. Tel.: +852-2632-3634/47; fax: +852-26377884.


E-mail address: ykwing@cuhk.edu.hk (Y.K. Wing).

ences were not clear. Lindberg et al. [9] investigated the


relationship between gender differences of insomnia and
their sex-related psychological status and reported a higher
prevalence of anxiety among females. Another study on
self-evaluations of factors disturbing sleep found that
females perceived psychological factors as the most important causes of impeding sleep and males ranked work-related
causes as the most important [10].
Definition of insomnia
Insomnia is mostly a symptom and sometimes a syndrome [11]. It is one of the symptoms of a number of
psychiatric (e.g., depression) and medical disorders. Patients
with insomnia report difficulty in initiating sleep (DIS), i.e.,
not easily falling asleep, difficulty in maintaining sleep

0022-3999/02/$ see front matter D 2002 Elsevier Science Inc. All rights reserved.
PII: S 0 0 2 2 - 3 9 9 9 ( 0 2 ) 0 0 4 3 7 - 3

602

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

(DMS), i.e., having intermittent awakening during the night,


or early morning awakening (EMA), i.e., awakening occurs
in the early morning and unable to sleep again [12]. The
duration of insomnia was also an important diagnostic
implication. Insomnia lasting days is often a result of acute
and transient stress but insomnia lasting longer than a few
weeks is considered chronic and significant [13].
Chinese studies
Insomnia is very common in the general population in
most western and Japanese studies [14,15], but there are
relatively few studies about the epidemiology of insomnia in
Chinese. Our recent study [16] investigated insomnia in
1034 noninstitutionalized elderly (age > 70) in Hong Kong.
We found that there was a high proportion of elderly
suffering from insomnia (38.2%). The risk factors included
being a female, older age, poorer physical and mental
health. Our current study aimed to investigate whether
demographic and psychosocial factors were associated with
the sex differences for insomnia in the Chinese adult
population in Hong Kong.

Materials and methods


Study design
Wing et al. has described detailed methods related to the
study [17]. Briefly, the study was conducted in the Hong
Kong Chinese noninstitutionalized adult population aged
18 65 via random telephone interviews by trained interviewers in mid-1998. There were 6.5 million people in
Hong Kong in the mid-1998 population estimation with the
18 65 age range constituting about 4.5 million (68.2%).
Ninety-nine percent of the households in Hong Kong have
telephones. An eligible household telephone number was
randomly selected from the 1995 residential telephone
directories for the whole Hong Kong, followed by random
selection of a subject according to the Kish selection method
[18]. Proxy respondent was not allowed. Each telephone
number was dialed at least three times, each on different
hours and on different days of the week, including weekends. If there were no answers from the households after
three calls, the number would be dropped and replaced with
another number. Replacement would also be done under the
following situations: non-Chinese, nonresidential number,
fax line, nonsubject rejection and no eligible subjects. All
subjects with verbal consent to participate in the study were
then interviewed with a structured questionnaire (paper
pencil version). Subjects who refused to participate either
verbally or by their absence of reply despite three consecutive calls were classified as refusals. The university ethical
committee approved the study.
One trained research assistant monitored the interviews
continuously. The structured questionnaire contained 60

questions with demographic information, sleep habits, questions about sleep disorders including Chinese Ullanlinna
Narcolepsy Scale, and questions specific for detecting
insomnia and other sleep disorders, health-related questions
and self-reporting questionnaire [19].
Instrument and measurement
The sleep questionnaire covered questions regarding
different sleep disturbances. The structured questions about
insomnia employed by the current study were as follows:
Do you have difficulty in falling asleep? (DIS)
(0) No
(1) Seldom
(2) Sometimes
(3) Always
For those who answered Sometimes or Always, they
would be asked about the following subquestions. For the
No or Seldom answers, they would not be considered
as suffering from significant insomnia and would proceed to
the next main item question.
When do you start to have difficulty in falling asleep?
How long do you take to fall asleep?
How often did you have difficulty to fall asleep in the
previous months?
(0) No
(1) Occasionally (2) 1 2 times/week
(3) 3 4 times/week (4) >4 times/week
Similar sets of structured questions for DMS and EMA
were asked.
In this paper, we would only concentrate on the insomnia
data about the preceding month. Insomnia was defined as
having either sometimes or always difficulty in initiating sleep (DIS), difficulty in maintaining sleep (DMS),
or early morning awakening (EMA) [14] and with at least
three times per week of such events within the last month.
The detailed analysis of the lifetime prevalence and other
sleep problems will be reported in other papers.
Potential risk factors
Demographic factors included sex, age, marital status,
education level, housing type, living area, and occupation.
The housing type was classified as public housing provided by government, private housing provided by private property developers and other housing types
including wooden house, village and other unclassified
housing. Living area was the size of the house measured
in square feet. Lifestyle factors included tea/coffee drinking,
smoking, alcohol consumption, and regular medication. The
questions regarding tea/coffee drinking, smoking and alcohol consumption were given choices as follows: no, only
previously, seldom, occasional, and frequent (defined as at
least three times or above a week). Those who answered no,
only previously or seldom were recoded as never or
seldom in the final analysis. The respondents were also
asked whether they regularly consumed Chinese herbal or
Western medication in the past year. A question regarding
environmental noise at night was included. Self-Reporting

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

Questionnaire-20 (SRQ-20) was used as the screening scale


for minor psychiatric disturbance e.g., depression and anxiety. The Chinese version of SRQ-20 was validated previously [19]. An optimal cut-off of 4/5 was found to have a
sensitivity of 68% and a specificity of 78%. Presence of
psychiatric disturbance was defined as scoring 5 points or
above (out of 20) on the SRQ-20.

Statistical analysis
Data were entered into Epi-Info ver 6 (Centres for Disease
Control and Prevention, Atlanta, GA), and statistical analysis
was performed using SPSS for Windows version 10.0
(SPSS, Chicago, IL). Descriptive results were presented as
mean (S.D.) or percentage. The data were weighed to adjust
for age and sex disparities between the sample and the mid1998 population by direct standardization method. Statistical
analysis was started with univariate analysis. Association of
the potential risk factors and insomnia were determined by
chi-squared test of association and chi-squared test of trend.
The age-adjusted OR of insomnia was estimated by logistic
regression method stratified by sex. Multivariate analysis
was then repeated using multiple logistic regression with
backward stepwise procedure. Sex-specific models for
estimating the likelihood of suffering insomnia as a function
of demographic and psychosocial factors were then
developed (Model 1). To assure that sex is a significant
factor of insomnia, we used a combined model including sex
and other common variables of insomnia for both sexes
(Model 2). Furthermore, to minimize the effects of interactions, the interaction terms between sex and other factors for
those that differed between sexes were added as independent
variables and built the third model (Model 3). The cut-off
point of entry of multiple logistic regression was fixed at
0.05 and the cut-off point of exclusion at 0.10. A two-sided
5% level of significance is considered significant for all

603

statistical tests; exact probability values were reported down


to P < .001. Ninety five percent confidence intervals were
provided as appropriate.

Results
A total of 43,921 telephone numbers were used in this
study. After excluding the no-answer numbers, nonsubject
rejection numbers (including those who hung up before/
after introduction), fax numbers, commercial numbers and
non-Chinese resident numbers, a total number of 13,738
telephone numbers were considered as eligible sample. In
this targeted sample, 10,072 subjects were interviewed
(72.9%) while 2354 targeted subjects were unable to contact
at home after three attempts and 1312 subjects refused. Of
the 10,072 interviewed subjects, 221 questionnaires were
rejected due to more than 30% of the items being incomplete. This resulted in a total of 9851 subjects (46.4% male;
53.6% female) who were included in the final analysis. The
study sample had a slightly higher female to male ratio than
the population of Hong Kong in 1998 (51% vs. 49%). The
mean age of the sample was 37 12. The age distribution
was similar except having more people in the youngest
group and fewer people in the middle-age group.
Prevalence of insomnia during the past month
The commonest type of sleep disturbance was DMS
(6.9%; 95% CI 6.4 7.5), followed by DIS (4.5%; 95% CI
4.1 5.0) and EMA (4.0%; 95% CI 3.6 4.4) (Fig. 1).
Females consistently had a higher prevalence of these sleep
disturbances than males. The preceding month prevalence of
insomnia, as defined by having DIS, DMS, or EMA and
with at least three times per week of such events within the
last month, was estimated to be 11.9% (95% CI 11.2 12.6).
Insomnia was more common in females than males (14.0%

Figure 1. Prevalence of difficulty in initiating sleep, difficulty in maintaining sleep, early morning awakening, and overall insomnia in male and female (DIS,
difficulty in initiating sleep; DMS, difficulty in maintaining sleep, EMA, early morning awakening).

604

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

Table 1
Prevalence of insomnia by gender and age groups
Male

Female

Total Prevalence
Age group N
(%)
95% CI

Total Prevalence
N
(%)
95% CI

18 29
30 44
45 65
Overall

1416
2714
1107
5237

1367 7.4
1971 8.9
1201 12.1
4539 9.3

6.0 8.8
7.6 10.2
10.2 14.0
8.4 10.1

11.1
13.8
18.1
14.0

9.4 12.8
12.5 15.1
15.8 20.4
13.0 14.9

Chi-squared test for trend for both sexes: P < .001.


Crude OR (women: men): 1.59 (1.40 1.80).
Age-adjusted OR (women: men): 1.61 (1.42, 1.83).

vs. 9.3%). The crude odds ratio was 1.59 (95% CI: 1.40
1.80) (Table 1). The prevalence of insomnia was shown to
increase with age and was higher in females in each age
group. A steady trend with increase in age was observed in
both groups ( P < .001; c2 test for trend). The age-adjusted
odds ratio was similar to crude odds ratio (1.61; 95% CI
1.42 1.83).
When sleep duration and satisfaction characteristics of
the male and female subjects were compared (Table 2),
females generally slept longer than males but they also
expected more sleep (i.e., longer sleeping hours) than males.
The insomniacs slept lesser than the non-insomniacs in both
sexes (male 6.53 vs. 7.03; female 6.69 vs. 7.25). More
insomniacs would report feeling inadequate rest after sleep
than the non-insomniacs (male 46.5% vs. 23.5%;
female 45% vs. 23.6%).
Effect of demographic and lifestyle factors on insomnia
(Table 3)
Education level
In Hong Kong, people normally spend 6 years in primary
school, 5 7 years in secondary school and 3 4 years in
tertiary institutions. There was a trend of increasing risk of
insomnia with lower educational levels in males. In com-

paring with tertiary-level education, secondary (ageadjusted OR 1.68; 95% CI 1.24 2.27) and primary school
level (age-adjusted OR 2.76; 95% CI 1.94 9.93) were
associated with higher prevalence of insomnia. Among
females, the pattern was less clear. Only the primary school
level was significantly associated with insomnia (ageadjusted OR 1.44; 95% CI 1.09 1.92) (Table 3).
Marital status
No significant association was found between marital
status and insomnia in males. However, divorced/widowed
females were more susceptible to insomnia (age-adjusted
OR 1.78; 95% CI 1.20 2.63).
Occupation
Unemployment was associated with a higher risk of
insomnia for both sexes (male: age-adjusted OR 2.44;
95% CI 1.75 3.41; female: age-adjusted OR 1.81; 95%
CI 1.21 2.69). Being a housewife (age-adjusted OR 1.59;
95% CI 1.33 1.90) and retiree (age-adjusted OR1.65; 95%
CI 1.06 2.56) were independently associated with higher
prevalence of insomnia in females and males, respectively.
Types of residence and living area
Compared to those living in private housing, men who
lived in public housing were more likely to have insomnia
(age-adjusted OR 1.75; 1.39 2.20). However, the association was not significant for those women who lived in
public housing. Those who lived in other housing types also
had higher risk for insomnia when compared to those who
lived in private houses in both sexes. There was a higher
risk of insomnia for males who lived in houses with
intermediate size (age-adjusted OR 1.40; 95% CI 1.04
1.89) and small size (age-adjusted OR 1.89; 95% CI 1.42
2.52). However, the effect was not significant in female. In
other words, the lower economic status as reflected by the
housing types was associated with insomnia in both sexes
but more pronounced in males.

Table 2
Characteristic of (a) sleep duration and (b) sleep satisfaction among insomniacs and non-insomniacs in both sex
(a) Sleep duration
Male

Female

Insomnia

Self-reported sleeping time (hours)


Expected sleeping time (hours)

No insomnia

Overall

Insomnia

No Insomnia

Overall

Mean

S.D.

Mean

S.D.

Mean

S.D.

Mean

S.D.

Mean

S.D.

Mean

S.D.

6.53
7.75

1.43
1.45

7.03
7.76

1.15
1.48

6.98
7.76

1.19
1.48

6.69
7.83

1.54
1.57

7.25
8.02

1.16
1.34

7.17
8.00

1.24
1.38

(b) Sleep satisfaction


Male

Female

Insomnia

Enough rest
General rest
Not enough rest

No insomnia

Overall

Insomnia

No insomnia

Overall

Cases

Cases

Cases

Cases

Cases

Cases

123
102
196

29.2
24.2
46.6

1835
1321
969

44.5
32.0
23.5

1958
1423
1165

43.1
31.3
25.6

175
228
330

23.9
31.1
45.0

1985
1474
1067

43.9
32.6
23.6

2160
1702
1397

41.1
32.4
26.6

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

605

adjusted for age, smoking became a significant factor


with insomnia. Effect of frequent smoking in females
(age-adjusted OR 2.04; 95% CI 1.47 2.82) was also
stronger than in males (age-adjusted OR 1.27; 95%CI
1.02 1.58). Regular medication (including both prescribed drugs and sometimes herbal medication) had more
than twofold risk for insomnia than those non-regular
drug users in both sexes (male: age-adjusted OR 2.46;

Lifestyle factors
Habits of tea/coffee drinking was not a significant
factor for insomnia in both sexes in the univariate
analysis. In our sample, men more often reported frequent
use of alcohol than did women. However, frequent
alcohol intake in women was more strongly associated
with insomnia than it was in men (age-adjusted OR 3.16;
95% CI 1.87 5.36). After stratification by sex and

Table 3
Association of demographic and lifestyle factors with insomnia, by gender
(a)
Male

Female
Age-adjusted

Education level
Tertiary level or above
Secondary level
Primary level/below
Marital status
Married/cohabiting
Single
Divorced/Widow
Occupation
Employed
Housewife
Student
Unemployed
Retired
Type of residence
Private
Public
Others
Living area (sq ft)
>600
401 600
 400

Age-adjusted
2

P value

Insomnia (%)

Total

OR

95% CI

c2

Insomnia (%)

Total

OR

95% CI

5.5
9.0
15.0

1006
2753
771

1.00
1.68
2.76

1.24 2.27
1.94 3.93

48.354

< .001

11.3
13.2
17.9

817
3278
1146

1.00
1.15
1.44

0.90 1.46
1.09 1.92

21.253

< .001

9.8
8.4
9.8

2699
1752
82

1.00
1.19
0.93

0.89 1.58
0.44 1.95

2.392

.302

14.7
10.7
24.8

3657
1445
145

1.00
0.80
1.78

0.62 1.02
1.20 2.63

28.754

< .001

8.6

5.6
19.0
15.7

3657

373
263
191

1.00

0.75
2.44
1.65

0.46 1.23
1.75 3.41
1.06 2.56

46.798

< .001

11.4
17.9
9.4
19.0
15.5

2727
1822
373
174
116

1.00
1.59
0.87
1.81
1.11

1.33 1.90
0.58 1.29
1.21 2.69
0.65 1.91

49.026

< 0.001

6.8
10.9
12.3

1900
2501
138

1.00
1.75
1.61

1.39 2.20
1.06 2.43

< .001

14.0
13.8
15.7

2340
2786
115

1.00
1.05
1.65

0.89 1.24
1.23 2.23

0.317

0.853

6.4
8.6
11.8

1060
1705
1718

1.00
1.40
1.89

1.04 1.89
1.42 2.52

24.223

< .001

13.7
13.6
14.4

1316
2021
1845

1.00
1.02
1.04

0.83 1.25
0.85 1.28

0.53

0.767

3659
883

1.00
1.42

1.13 1.80

10.577

.001

12.2
20.7

4182
1072

1.00
1.85

1.55 2.20

51.58

< .001

977
864
2705

1.00
0.96
0.81

0.58 1.13
0.74 1.24

3.023

.221

15.2
12.7
13.7

1680
1096
2479

1.00
0.83
0.86

0.67 1.04
0.72 1.03

3.736

.154

2984
1104
457

1.00
1.10
1.68

0.86 1.40
1.25 2.26

14.051

.001

13.7
14.0
31.8

4641
549
66

1.00
1.08
3.16

0.84 1.40
1.87 5.36

17.838

< .001

3125
214
1205

1.00
0.83
1.27

0.48 1.42
1.02 1.58

6.838

.033

13.6
12.5
21.9

4942
80
237

1.00
1.01
2.04

0.51 1.97
1.47 2.82

13.328

.001

4096
447

1.00
2.46

1.89 3.20

56.411

< .001

12.3
24.6

4546
711

1.00
2.22

1.83 2.70

78.01

< .001

3147
1384

1.00
4.01

3.25 4.93

183.235

< .001

7.9
24.1

3290
1953

1.00
3.80

3.22 4.48

267.13

< .001

22.585

P value

(b)
Environmental noise at night
No
8.6
Yes
12.1
Tea/coffee drinking
Never or seldom
9.3
Occasional
7.8
Frequent
9.7
Alcohol consumption
Never or seldom
8.5
Occasional
9.2
Frequent
14.0
Smoking
Never or seldom
8.7
Occasional
7.0
Frequent
11.0
Regular medication
No
8.2
Yes
19.0
Psychiatric disturbance
No (SRQ < 5)
5.4
Yes (SRQ  5)
18.1

606

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

95%CI 1.89 3.20; female: age-adjusted OR 2.22; 95%CI


1.83 2.70).
Noisy environment at night
Compared to those who regarded their environment as
not noisy at night, those residing in the noisy area were
associated with a higher risk of insomnia. Effect of noise
was stronger in females (age-adjusted OR 1.85; 95% CI
1.55 2.20) than males (age-adjusted OR 1.42; 95%CI
1.13 1.80).

Table 4
Model 1: Demographic and psychosocial predictors of insomnia in male
and female (multiple logistic regression)
Male

Control factor
Age
18 29
30 44
45 65
Adjusted factors
Education level
Tertiary level or above
Secondary level
Primary level or below
Marital status
Married
Single
Divorced/Widow
Occupation
Employed
Student
Housewife
Unemployed
Retired
Types of residence
Private
Others
Public
Environmental noise at night
No
Yes
Tea/coffee drinking
Never/seldom
Occasional
Frequent
Alcohol consumption
Never/seldom
Occasional
Frequent
Regular medication
No
Yes
Psychiatric disturbance
No (SRQ < 5)
Yes (SRQ  5)
* P < .05.

Female

OR

95.0%
CI

OR

95.0% CI
for OR

1.00
1.11
1.26

0.83 1.49
0.90 1.76

1.00
1.08
1.36

0.82 1.42
0.99 1.86

1.00
1.37*
1.84*

1.00 1.88
1.26 2.70

Table 5
Model 2: Demographic and psychosocial predictors of insomnia including
common risk factors and sex as a covariate

Control factor
Age
18 29
30 44
45 65
Adjusted factor
Sex
Male
Female
Occupation
Employed
Student
Housewife
Unemployed
Retired
Types of residence
Private
Others
Public
Alcohol consumption
Never/seldom
Occasional
Frequent
Regular medication
No
Yes
Psychiatric disturbance
No (SRQ < 5)
Yes (SRQ  5)

OR

95.0% CI

1.00
1.16
1.47*

0.96 1.39
1.20 1.81

1.00
1.34*

1.14 1.57

1.00
0.87
1.65*
1.86*
1.35

0.63 1.19
1.38 1.97
1.43 2.43
0.94 1.92

1.00
1.45*
1.16*

1.13 1.86
1.01 1.34

1.00
1.02
1.92*

0.85 1.23
1.47 2.49

1.00
1.71*

1.44 2.02

1.00
3.55*

3.11 4.06

* P < .05.
1.00
0.91
1.57*

0.68 1.21
1.02 2.40

1.00
0.78

1.91*
1.39*

0.46 1.30

1.34 2.73
0.86 2.23

1.00
1.00
1.65*
1.62*
1.01

0.65 1.52
1.35 2.01
1.06 2.49
0.57 1.80

1.00
1.26
1.47*

0.81 1.95
1.15 1.88

1.00
1.56*
0.98

1.13 2.15
0.82 1.17

1.00
1.56*

1.29 1.88

1.00
0.82
0.81*

0.64 1.04
0.66 0.97

1.00
1.02
1.69*

0.79 1.32
1.23 2.31

1.00
1.09
2.84*

0.83 1.44
1.61 4.98

1.00
1.93*

1.46 2.57

1.00
1.65*

1.34 2.04

1.00
3.41*

2.75 4.24

1.00
3.41*

2.86 4.06

Psychiatric disturbance
SRQ-20 measured minor psychiatric disturbances and
they were correlated with higher prevalence of insomnia in
both male (OR 4.01; 95% CI 3.25 4.93) and female (OR
3.80; 95% CI 3.22 4.48).
Analysis of sex differences in risk factors for insomnia
multivariate analysis
Table 4 shows the effects of each variable after adjusting
for other variables in the models for males and females
separately. The common risk factors retained in the models
included occupation, types of residence, alcohol consumption, regular medication and psychiatric disturbance. Education level and being retired were retained in the model for
males, while marital status and noisy environment at night
were factors retained in the model for females. Among the
variables in the models, psychiatric disturbance was found
to have the strongest effect on insomnia (male: age-adjusted
OR 3.41; 95%CI 2.75 4.24; female: age-adjusted OR 3.41;
95%CI 2.86 4.06)
A combined model was built to test if sex would be
retained in the model taking into account the factors
observed to have significant effect on insomnia. (Table 5)

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609


Table 6
Model 3: Combined model with interaction effects related to sex differences

Control factor
Age
18 29
30 44
45 65
Adjusted factor
Sex
Male
Female
Education level
Tertiary level or above
Secondary level
Primary level or below
Occupation
Employed
Student
Housewife
Unemployed
Retired
Types of residence
Private
Others
Public
Tea/Coffee drinking
Never/seldom
Occasional
Frequent
Alcohol consumption
Never/seldom
Occasional
Frequent
Regular medication
No
Yes
Psychiatric disturbance
No (SRQ < 5)
Yes (SRQ  5)
Education  Sex
Primary by female
Secondary by female
Noisy by female
Housing  Sex
Other housing by female
Public housing by female

OR

95.0% CI

1.00
1.12
1.35*

0.93 1.35
1.08 1.68

1.00
2.23*

1.51 3.29

1.00
1.37*
1.84*

1.00 1.87
1.28 2.67

1.00
0.87
1.70*
1.79*
1.26

0.63 1.20
1.41 2.04
1.37 2.35
0.88 1.80

1.00
1.28
1.45*

0.83 1.97
1.14 1.85

1.00
0.82*
0.85*

0.67 0.99
0.73 0.99

1.00
1.07
1.94*

0.88 1.28
1.48 2.54

1.00
1.72*

1.45 2.03

1.00
3.42*

2.99 3.92

0.52*
0.64*
1.55*

0.32 0.83
0.42 0.96
1.28 1.86

1.22
0.67*

0.71 2.08
0.49 0.90

* P < .05.

The combined model shows that both sex and other common factors were retained in the logistic regression. The
effect of sex was only slightly reduced after adjustment.
Females were 1.3 times more likely than males to have
insomnia (age-adjusted OR 1.34; 95% CI 1.14 1.57).
The final model combined the two sexes and also
included the interaction terms for variables that differed
between sexes. (Table 6) Three interaction effects were
statistically significant at a level of P < .05 for education,
environmental noise at night and types of residence, and
thus they were retained in the model after the stepwise
procedure of logistic regression. The effect of sex was
enhanced (age-adjusted OR 2.23; 95% CI 1.51 3.29).

607

Discussion
We observed the first large-scale cross-sectional study on
insomnia among Chinese adults in Hong Kong. Overall,
11.9% of Hong Kong Chinese adults reported suffering from
frequent insomnia (at least  3 times/week) for the preceding month with more female insomniacs than male (14.0%
vs. 9.3%). People with insomnia were characterized with less
nocturnal sleeping time and feeling inadequate rest in the
morning. The most frequent sleep disturbance was DMS,
followed by DIS and EMA. DMS being the most common
insomniac complaint were similarly reported in other studies
as well [9]. Similar with other studies [4,8], there was an
increase in the risk of insomnia in older subjects. In view of
some negative association between age and insomnia, further
studies were required to delineate the age effect on insomnia
[15,20]. The gender difference in the prevalence rate of
insomnia was still significant even after adjusting for age
and other potential confounding risk factors.
In parallel to other studies, insomniacs in Hong Kong
Chinese were found to share similar risk factors such as
unemployment, lower economic status (as reflected by types
of residence), alcohol consumption, poor health (regular
medication) and psychiatric disturbances [21,22]. Although
these factors were common for both sexes, unemployment,
retirement and regular medication had a stronger effect on
males than females. In addition, lower education level was
more related to insomnia in men but not in women. Those
women who experienced divorce or widowhood or being a
housewife were associated with a higher prevalence of
insomnia [21,22]. Environmental noise at night had a
significant effect only in women [21,22].
Albeit, men more often reported frequent use of alcohol
than did women; women seemed to be more vulnerable to
the deleterious effect of frequent alcohol intake to sleep than
men [23]. Interestingly, although tea or coffee drinking had
been suggested as a risk factor for insomnia [21,22], we
observed that in women, frequent tea or coffee drinking had
a paradoxically lower risk of insomnia. One possibility was
that people who had sleep problems might avoid consuming
these caffeine-containing beverages, which are notorious
stimulants [24]. In our study, psychiatric disturbances had
the strongest effect on insomnia among all the factors
studied [25]. The risk for potential psychiatric problems,
however, has similar strength of effect on insomnia in both
sexes in our study.
In our study, smoking was not associated with an
increase in the risk of insomnia after adjusting for other
variables. This was consistent with another study [26].
Although some studies [27,28] suggested that smoking
was associated with a higher prevalence of insomnia, they
did not control for other confounding factors such as
caffeine or alcohol consumption.
In short, although there were common risk factors for
insomnia in both sexes, there existed gender-specific risk
factors. Overall, lower education level and being retired

608

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

were more associated with a higher risk of insomnia in


males while marital status (divorced/widowed), being
housewife, and noisy sleeping environment were more
associated with a higher risk of insomnia in females [10].
Another crucial question that remains to be answered is
why females are more susceptible to insomnia than males.
The higher prevalence of insomnia in females seemed to be
a robust finding across different epidemiological studies [4
8,15]. The fact that similar gender differences do occur in
depression, anxiety and migraine may suggest that they
share some gender-related etiologies and pathogenesis
[15,29,30]. The possible reasons for gender differences in
various neuropsychiatric and sleep disorders are likely
multifactorial and multilevel. These could vary from artefactual reasons such as measurement and recall bias to
biological, psychological and social factors [2].
In general, females tend to report more intense and more
frequent somatic symptoms than males [31 33], but few
gender differences exist in the recall patterns, illness behavior and psychometric measurement scales that could fully
account for the gender differences in the prevalence of
neuropsychiatric disorders [2,34,35]. On the other hand,
females were consistently reported to have a higher prevalence of psychiatric morbidities, especially depression and
anxiety than males [2,9]. The higher prevalence of psychiatric morbidities may help to account for a higher prevalence of insomnia in females, albeit the presence of
psychiatric problems has similar strength in predisposing
to insomnia in both sexes in our study. Biological differences in sex steroids have been postulated to account for
some of the gender-related sleep changes and abnormalities
[3]. The reports of menstrual-related changes with more
subjective and objective sleep disturbances, especially in the
late luteal phase, of healthy female subjects supported the
role of gonadal hormones in mediating the sleep changes
[3], albeit direct evidence of menstrual-related changes in
primary insomnia was lacking [36]. On the other hand,
consistent findings of greater sleep dysregulation in
depressed females than males further implicated the role
of gonadal hormones [3,37].
Other social and psychological factors may also help to
explain the higher prevalence of insomnia in female. In
explaining the gender differences in depression, marriages
have been suggested as disadvantageous to females because
of gender-specific demand and burden as well as restricted
roles available to females in the marriage [2,38]. The
importance of family and marriage in female insomniacs
was again supported in our study, and being a housewife or
being widowed and divorced were most affected. The
differential effect of sociocultural environment towards
different gender was further suggested by work-related
factors. While unemployment was the risk factor for both
sexes, it was in males that the effect was more pronounced
and retirement was the risk factor for insomnia in males
only. Lack of social support or social isolation has been
reported to be the significant correlate or risk factor for

depression [39]. The importance of social support has been


indirectly reflected, in our study, by the higher risks of
insomnia in divorced/widowed ladies and retired men.
However, owing to the lack of detailed measurement of
the social support system in our study, further study on the
relationship between social support and insomnia is indicated. In addition, different gender-specific stress-coping
strategies have been suggested to explain the increased
vulnerabilities in depression in females [2,40].
Like all epidemiological studies, there were limitations in
our study. The major limitation was that this study was
based on self-reported data only without the second-phase
clinical and laboratory confirmation. Nevertheless, in order
to minimize the variation in subjective reporting, we defined
insomnia (outcome) with objective criteria, i.e., at least three
times per week in the last month. Secondly, the measurements of tea/coffee and alcohol consumption were based on
crude subjective estimates. Thirdly, as suggested by our
study, there seemed to be a significant contribution of
psychiatric comorbidities like depression and anxiety
towards insomnia; further studies are required to determine
whether these gender-specific factors would be applicable to
primary insomnia per se.

Acknowledgments
This research was funded by Research Grants Council of
Hong Kong, RGC CUHK 4282/97M.

References
[1] Reckelhoff JF. Gender differences in the regulation of blood pressure.
Hypertension 2001;37(5):1199 208.
[2] Piccinelli M, Wilkinson G. Gender differences in depression critical
review. Br J Psychiatry 2000;177:486 92.
[3] Manber R, Armitage R. Sex, steroids, and sleep: a review. Sleep
1999;22(5):540 55.
[4] Bixler EO, Kales A, Soldatos CR, Kales JD, Healey S. Prevalence of
sleep disorders in the Los Angeles metropolitan area. Am J Psychiatry
1979;136(10):1257 62.
[5] Mellinger GD, Balter MB, Uhlenhuth EH. Insomnia and its treatment.
Prevalence and correlates. Arch Gen Psychiatry 1985;42(3):225 32.
[6] Klink ME, Quan SF, Kaltenborn WT, Lebowitz MD. Risk factors
associated with complaints of insomnia in a general adult population.
Influence of previous complaints of insomnia. Arch Intern Med 1992;
152(8):1634 7.
[7] Hohagen F, Rink K, Kappler C, Schramm E, Riemann D, Weyerer S,
Berger M. Prevalence and treatment of insomnia in general practice. A
longitudinal study. Eur Arch Psychiatry Clin Neurosci 1993;242(6):
329 36.
[8] Reyner LA, Horne JA, Reyner A. Gender- and age-related differences
in sleep determined by home-recorded sleep logs and actimetry from
400 adults. Sleep 1995;18(2):127 34.
[9] Lindberg E, Janson C, Gislason T, Bjornsson E, Hetta J, Boman G.
Sleep disturbances in a young adult population: can gender differences
be explained by differences in psychological status? Sleep 1997;
20(6):381 7.
[10] Urponen H, Vuori I, Hasan J, Partinen M. Self-evaluations of factors

R.H.Y. Li et al. / Journal of Psychosomatic Research 53 (2002) 601609

[11]

[12]

[13]
[14]

[15]

[16]

[17]

[18]
[19]

[20]

[21]

[22]

[23]
[24]

[25]

promoting and disturbing sleep: an epidemiological survey in Finland.


Soc Sci Med 1988;26(4):443 50.
Partinen M. Epidemiology of sleep disorders. In: Kryger MH, Roth T,
Dement WC, editors. Principles and practice of sleep medicine. Philadelphia (PA): Saunders, 1994. pp. 437 52.
Ohayon MM, Caulet M, Priest RG, Guilleminault C. DSM-IV and
ICSD-90 insomnia symptoms and sleep dissatisfaction. Br J Psychiatry 1997;171:382 8.
Kupfer DJ, Reynolds CF. Management of insomnia. N Engl J Med
1997;336(5):341 6.
Kim K, Uchiyama M, Okawa M, Liu X, Ogihara R. An epidemiological study of insomnia among the Japanese general population.
Sleep 2000;23(1):41 7.
Ohayon MM, Zulley J, Guilleminault C, Smirne S, Priest RG. How
age and daytime activities are related to insomnia in the general population: consequences for older people. J Am Geriatr Soc 2001;49(4):
360 6.
Chiu HF, Leung T, Lam LC, Wing YK, Chung DW, Li SW, Chi I,
Law WT, Boey KW. Sleep problems in Chinese elderly in Hong
Kong. Sleep 1999;22(6):717 26.
Wing YK, Li RHY, Lam CW, Ho CKW, Fong SYY, Leung T. Prevalence of narcolepsy in Hong Kong Chinese. Ann Neurol 2002;51:
578 84
Kish L. Survey sampling. New York: Wiley, 1965.
Chen CN, Wong J, Lee N, Chan-Ho MW, Lau JT, Fung M. The Shatin
community mental health survey in Hong Kong: II. Major findings.
Arch Gen Psychiatry 1993;50(2):125 33.
Foley DJ, Monjan A, Simonsick EM, Wallace RB, Blazer DG. Incidence and remission of insomnia among elderly adults: an epidemiologic study of 6,800 persons over three years. Sleep 1999;22(Suppl 2):
S366 72.
Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances
and psychiatric disorders. An opportunity for prevention? JAMA
1989;262(11):1479 84.
Weissman MM, Greenwald S, Nino-Murcia G, Dement WC. The
morbidity of insomnia uncomplicated by psychiatric disorders. Gen
Hosp Psychiatry 1997;19(4):245 50.
Johnson JE. Insomnia, alcohol, and over-the-counter drug use in oldold urban women. J Community Health Nurs 1997;14(3):181 8.
Curless R, French JM, Williams GV, James OF. Colorectal carcinoma: do elderly patients present differently? Age Ageing 1994;23(2):
102 7.
Benca RM, Obermeyer WH, Thisted RA, Gillin JC. Sleep and psy-

[26]

[27]
[28]
[29]

[30]

[31]
[32]

[33]

[34]

[35]

[36]

[37]

[38]
[39]

[40]

609

chiatric disorders. A meta-analysis. Arch Gen Psychiatry 1992;49(8):


651 68.
Karacan I, Thornby JI, Williams RL. Sleep disturbance: a community
survey. In: Guilleminault C, Lugaresi E, editors. Sleep/wake disorders:
natural history, epidemiology, and long-term evolution. New York:
Raven Press, 1983. pp. 37 60.
Lexcen FJ, Hicks RA. Does cigarette smoking increase sleep problems. Percept Mot Skills 1993;77(1):16 8.
Phillips BA, Danner FJ. Cigarette smoking and sleep disturbance.
Arch Intern Med 1995;155(7):734 7.
Breslau N, Rasmussen BK. The impact of migraine: epidemiology,
risk factors, and co-morbidities. Neurology 2001;56(6 Suppl 1):
S4 12.
Foley DJ, Monjan AA, Izmirlian G, Hays JC, Blazer DG. Incidence
and remission of insomnia among elderly adults in a biracial cohort.
Sleep 1999;22(Suppl 2):S373 8.
Barsky AJ, Peekna HM, Borus JF. Somatic symptom reporting in
women and men. J Gen Intern Med 2001;16(4):266 75.
Silverstein B. Gender difference in the prevalence of clinical depression: the role played by depression associated with somatic symptoms.
Am J Psychiatry 1999;156(3):480 2.
Young MA, Scheftner WA, Fawcett J, Klerman GL. Gender differences in the clinical features of unipolar major depressive disorder.
J Nerv Ment Dis 1990;178(3):200 3.
Steer RA, Beck AT, Brown G. Sex differences on the revised Beck
Depression Inventory for outpatients with affective disorders. J Pers
Assess 1989;53(4):693 702.
Spinhoven P, Kooiman CG. Defense style in depressed and anxious
psychiatric outpatients: an explorative study. J Nerv Ment Dis 1997;
185(2):87 94.
Manber R, Bootzin RR, Acebo C, Carskadon MA. The effects of
regularizing sleep wake schedules on daytime sleepiness. Sleep
1996;19(5):432 41.
Steiger A, von Bardeleben U, Wiedemann K, Holsboer F. Sleep EEG
and nocturnal secretion of testosterone and cortisol in patients with
major endogenous depression during acute phase and after remission.
J Psychiatr Res 1991;25(4):169 77.
Bebbington PE. Sex and depression. Psychol Med 1998;28(1):1 8.
Roberts RE, Shema SJ, Kaplan GA, Strawbridge WJ. Sleep complaints and depression in an aging cohort: a prospective perspective.
Am J Psychiatry 2000;157(1):81 8.
Nolen-Hoeksema S. Sex differences in unipolar depression: evidence
and theory. Psychol Bull 1987;101(2):259 82.