Wzorzec-przegl d lekarski-XX-2001

ORIGINAL PAPERS – PRACE ORYGINALNE
Sara ARBER
Gender, marital status and sleep problems
in Britain
P³eæ, stan cywilny i problemy ze snem
w Wielkiej Brytanii
Department of Sociology, University of Surrey,
Guildford, Surrey
Co-Director, Centre for Research on Ageing
and Gender (CRAG): Professor Sara Arber
Additional key words:
sleep problems
gender
marital status
socio-economic status (SES)
Dodatkowe s³owa kluczowe:
problemy ze snem
p³eæ
status cywilny
status spo³eczno-ekonomiczny
Professor Sara Arber
Co-Director, Centre for Research on Ageing
and Gender (CRAG),
Department of Sociology, University of Surrey,
Guildford, Surrey GU2 7XH
Tel: 01483-686973; Fax: 0441483-689551
Email: S.Arber@surrey.ac.uk
Sleep is fundamental to health and
well-being, with women consistently
reporting greater sleep problems than
men, yet scant sociological research
has examined gender differences in
sleep quality. This paper analyses (i)
gender differences in sleep problems,
and (ii) how marital status differences
in sleep problems differ among women
and men. In both cases, the relative
contributions of socio-economic status
(SES), smoking, worries, health
and depression in explaining these
gender and marital status differences
are analysed. Logistic regression is
used to analyse the British Psychiatric
Morbidity Survey 2000, which interviewed
8578 men and women aged 16 to
74. Women reported significantly more
sleep problems than men, as did the
divorced and widowed compared with
married respondents. Gender differences
in sleep problems were halved following
adjustment for socio-economic
characteristics, suggesting that SES
inequalities play a major part in accounting
for gender differences. This study
casts doubt on the primacy of physiological
explanations underlying
gender differences in sleep. Marital
status differences in sleep are greater
among men than women, with previously
partnered men reporting particularly
poor quality sleep. However, this
is largely explained by the more disadvantaged
socio-economic circumstances
of the previously partnered,
especially for men. The paper concludes
that gender and marital status differences
are partly due to the lower
socio-economic status of women and
of the previously partnered.
Sen jest podstawowym wyznacznikiem
zdrowia i samopoczucia, kobiety
konsekwentnie podaj¹ wiêcej problemów
ze snem ni¿ mê¿czyŸni, dot¹d
niewiele badañ socjologicznych analizowa³o
genderowe ró¿nice w jakoœci
snu. Artyku³ ten analizuje 1/genderowe
ró¿nice w odniesieniu do problemów
ze snem, 2/ próbuje wskazaæ jak
stan cywilny ró¿nicuje problemy ze
snem kobiet i mê¿czyzn. W obu przypadkach
wp³yw statusu spo³ecznoekonomicznego
(SSE), palenie tytoniu,
zmartwienia, stan zdrowia i depresja
by³y analizowane dla wyjaœnienia tych
ró¿nic wynikaj¹cych z p³ci i stanu cywilnego.
Model regresji logistycznej
zosta³ zastosowany dla przeanalizowania
wyników British Psychiatric Morbidity
Survey 2000, w którym przeprowadzono
wywiady z 8578 mê¿czyzn i
kobiet w wieku 16 do 74 lat. Kobiety
zg³asza³y istotnie wiêcej problemów ze
snem ni¿ mê¿czyŸni, tak jak osoby rozwiedzione
i owdowia³e w porównaniu
do respondentów pozostaj¹cych w
zwi¹zkach ma³¿eñskich. Ró¿nice genderowe
w problemach ze snem wystandaryzowane
w odniesieniu do statusu
spo³eczno-ekonomicznego sugeruj¹,
¿e nierównoœci wynikaj¹ce z statusu
spo³eczno-ekonomicznego odgrywaj¹
g³ówn¹ rolê w okreœlaniu ró¿-
nic genderowych. To badanie zasia³o
w¹tpliwoœæ oddawania pierwszeñstwa
fizjologicznemu wyjaœnianiu tych ró¿-
nic genderowych. Ró¿nice w œnie zale¿ne
od stanu cywilnego s¹ wiêksze
wœród mê¿czyzn ni¿ wœród kobiet,
mê¿czyŸni ,którzy byli poprzednio w
zwi¹zkach szczególnie zwracali uwagê
na z³¹ jakoœæ snu. Jakkolwiek, by³y
one w wiêkszoœci wyjaœniane przez
wiêcej niekorzystnych uwarunkowañ
zwi¹zanych ze statusem spo³ecznoekonomicznym
mê¿czyzn poprzednio
przebywaj¹cych w zwi¹zkach. Wnioski
wskazuj¹ na fakt, ¿e ró¿nice genderowe
i wynikaj¹ce ze stanu cywilnego s¹
czêœciowo zale¿ne od niskiego statusu
spo³eczno-ekonomicznego kobiet i
od przebywania poprzednio w zwi¹zku.
54 Przegl¹d Lekarski 2012 / 69 / 2 S. Arber

Introduction
Social scientists have paid relatively little
attention to quality of sleep, nor have they
assessed whether social factors mediate
gender differences in sleep quality. This
paper examines the British Psychiatric Morbidity
Survey [28] to better understand gender
and marital status differences in sleep
problems.
It is well-known in sleep research that
women report higher levels of sleep complaints
than men [16, 26, 33]. A meta-analysis
of 29 published studies concluded that
the overall risk ratio of insomnia was 1.41
for women compared to men [33]. Biological
or physiological sex differences are often
identified as explanations for women's
higher levels of disturbed sleep [4,7,19], and
psychological explanations are also prevalent
[19]. Women have higher levels of depression
and anxiety, and research shows
that individuals suffering from psychiatric
disorders, such as depression and anxiety,
have poorer quality sleep [23,30]. However,
gender differences in sleep quality remain
after removing the effects of women's higher
rates of psychiatric morbidity [19,33].
Studies of gender differences in sleep
less frequently consider sociological explanations.
Chen et al.[4] conclude that 'In contrast
with explanations emphasising gender
differences in biology and prior psychiatric
illnesses, the sociological perspective has
not been well investigated in the existing
literature.' They found that the gender difference
in sleep disturbance was reduced
after controlling for women's social roles
(using marital status, employment status,
and number of children under 15 as a proxy
for childcare responsibilities). However,
women's sleep quality still remained significantly
poorer than men's, suggesting the
need for research on gender differences
using more extensive measures of social
roles and socio-economic status (SES).
Sekine et al. [26] found that the gender difference
in reported sleep quality among
Japanese civil servants could be entirely
explained by gender differences in work
characteristics, domestic roles and familywork
conflicts. However, it remains unclear
whether gender differences can be explained
by socio-economic, work and family
characteristics among nationally representative
population samples.
Other researchers have shown the
poorer sleep quality of the divorced/separated
and of widowed people. However, it
is also important to examine whether there
is a gender and marital status interaction in
relation to sleep quality, rather than simply
analysing gender and marital status as additive
variables within analyses. In particular,
whether there are greater marital status
differences among men than women,
and the reasons underlying any such differences.
Married men have been shown
to be particularly advantaged, regarding a
number of other dimensions of well-being
(including income, material circumstances,
health and social support), with smaller
marital status differences found among
women than among men [1]
A major reason for poorer sleep quality
is because chronic ill-health causes pain and
discomfort at night, resulting in sleep complaints
and disorders [5, 29, 32]. Since poor
physical and mental health is associated
with disrupted sleep it is important to assess
whether gender or marital status differences
in sleep quality are confounded by poorer
health among women and among those who
are previously married.
Materials and Methods
This paper addresses the following research questions:
(1) What is the relative contribution of socio-economic
characteristics, compared with smoking, worries,
poor health and depression, in explaining gender differences
in sleep problems
(2) How does the association between marital status
and sleep problems differ among women and men,
and what is the relative contribution of socio-economic
characteristics and other factors to marital status differences
in sleep problems
The paper analyses a British nationally representative
cross-sectional survey: the 2000 Psychiatric Morbidity
Survey [28]. Home interviews (averaging 1 1/2 hours)
were conducted with 8,580 people aged 16-74. A representative
sample of addresses was selected from the
Postcode Address File with one household member aged
16-74 randomly selected for interview from each sample
household. The response rate of 69.5% is high, considering
the length and complexity of the interview [28, 29].
The maximum age is 74 years, therefore our findings
cannot be generalised to older people above this age.
Sleep Problems
Sleep problems were measured as part of the revised
version of the Clinical Interview Schedule (CIS-R)
[18]. Respondents were asked: 'In the past month, have
you been having problems with trying to get to sleep or
with getting back to sleep if you woke up or were woken
up' Those answering 'Yes', were asked: 'On how many
of the past seven nights did you have problems with your
sleep' Three response categories were provided: None;
1 to 3 nights; 4 nights or more. The paper analyses a
dichotomous variable of reporting sleep problems on 4
or more nights per week (versus less often) as an indicator
of frequently experienced sleep difficulties.
Socio-demographic and socio-economic
characteristics
Age was categorised as 16-24, 25-34, 35-44, 45-
54, 55-64, 65-74.
Marital status - Married/cohabiting, Never married,
Widowed, Divorced/separated. Number of children living
in the household, coded as None, 1, 2, 3 or more.
Highest educational qualifications were coded into
4 ordinal categories: Degree level or higher qualifications;
Professional, teaching, or technical qualifications
and A levels (national examinations taken at 18 and required
for university entry); Lower qualifications, e.g.
GCSE/O levels, secretarial, trade and apprenticeship
qualifications; and No educational qualifications.
Employment status was self-reported as Full-time
employed, Part-time employed, Unemployed (looking for
work in last 4 weeks) or Economically inactive (coded
from main reason not working in last 4 weeks as retirement,
full-time student, household duties, or disability/
long-term sick).
Housing tenure was coded as Owns accommodation;
Rents from Public Housing (Local Authority or Housing
Association); or Rents from other sources, primarily
private renting.
Household income is measured by the sum of personal
gross income from all sources for each household
member, equivalised using the McClements Scale [6],
which adjusts for differences in number of adults and
children in the household. Income was coded into 5 ranges:
the lowest category (< £ 150 per week) comprises
20% of the whole sample, and the highest (> £ 750 per
week) comprises 12%.
Measures of other variables
Smoking was categorised as Never smoked, Exsmoker
and Current smoker.
Worries. Self-reported data about Worries (in general),
Worries about health, and Depression were obtained
using the CIS-R symptom scores [18]. For each
of these measures, interviewees were asked if they had
experienced 4 different symptoms in the last 7 days. In
each case, this was scored as None, Medium (1 symptom
reported), and High (2 or more symptoms reported).
Self-rated (or self-assessed) health is used extensively
in research on determinants of health, and is a
good predictor of mortality and general health [8,14]. Selfrated
health was measured by asking: 'How is your health
in general Would you say your health is Excellent, Very
Good, Good, Fair or Poor', recoded into 3 categories:
Very good (representing 'excellent' or 'very good'), Good,
and Poor (representing 'fair' or 'poor').
Number of chronic illnesses was measured by asking;
'Do you have any long-standing illnesses, disability
or infirmity By long-standing I mean anything that has
troubled you over a period of time or that is likely to affect
you over a period of time.' Respondents self-reporting
'Yes', were asked 'What is the matter with you' and
each health problem mentioned was recorded. The total
number of long-standing health problems self-reported
were summed, and recoded as None, 1, 2, 3 or more.
Statistical analysis
The proportions of men and women reporting sleep
problems on 4 or more nights a week are analysed for
each of the above variables, with chi-squared probability
values reported (Tables 1-2). Hierarchical logistic regression
models are presented to examine the two research
questions addressed in the paper.
Statistical analyses were performed using SPSS
(version 13, SPSS, Inc., Chicago, IL). Statistical tests
used .05 (2 tailed) significance levels. To allow the
strength and precision of the relationships to be assessed,
results are presented as odds ratios (OR) with
their associated 95% confidence intervals (CI).
Results
Gender differences in sleep problems
- bivariate analyses
Tables I and II show the proportion of
men and women reporting sleep problems
on 4 or more nights per week in relation to
socio-demographic, socio-economic, health
and other variables. More women (20%)
than men (14%) report sleep problems on 4
or more nights per week. A significant relationship
with age is found for women but not
for men. The highest proportion of women
reporting sleep problems are age 45-54
(24%) with a modest decline above this age.
The divorced/separated report the worst
sleep among both women (27%) and men
(26%), followed by the widowed (26%
women, 21% men). The gender difference
in sleep problems is greater among married
people (18% of married women and 12.5%
married men report sleep problems), than
among the single (16% vs 12%), or divorced.
This suggests that the sleep advantage of
marriage may be primarily experienced by
married men, rather than married women,
which would accord with results from a
range of qualitative studies [13,31].
Significant associations with sleep problems
are found for each socio-economic status
(SES) measure, which are broadly comparable
for men and women. Regarding
household income, only 10% in the highest
income group report sleep problems compared
with 25% in the lowest income group.
Adults with more education report fewer
sleep problems; only 12% with a degree report
sleep problems compared with 22%
who have no qualifications. The unemployed
(22%) and economically inactive (25%) are
more likely than the full-time employed
Przegl¹d Lekarski 2012 / 69 / 2
55

2
1,8
Figure 1
Odds Ratios of Women having Poor Sleep Quality
on 4 or more nights per week, compared with the
Men (reference category, men = 1.0) a .
1,6
1,4
1,2
1
1.49**
1.42**
1.23**
1.17*
1.27**
1. Age
2. + Marital Status, children
3. + Socioeconomic
4. + Worries/Smoking
5. + Health
a. Hierarchical logistic regression models adjusting for
sets of variables, as in Table III. Model 1, Age; Model 2,
+ Marital Status, Number of Children; Model 3, +
Socioeconomic (Educational Qualifications, Employment
Status, Household Income, Housing Tenure); Model 4,
+ Worries, Smoking; Model 5, + Subjective health, Health
worries, Number of chronic illnesses, Depressive
symptoms.
Significance of gender difference, ** p

£
Table I
Proportion reporting poor sleep on 4 or more nights a week by demographic and socio-economic variables
by gender, age 16-74
Men
Women
Total
% n = % N = % n=
All
14.
2 3851
19.
7 4727
17.
2 8578
Age
16-24
10.
9 385
15.
9 409
13.
5 794
25-34
12.
9 711
15.
7 972
14.
6 1683
35-44
13.
2 823
17.
5 1024
15.
6 1847
45-54
15.
1 747
24.
2 798
19.
8 1545
55-64
16.
7 646
22.
0 796
19.
6 1442
65-74
15.
4 539
22.
9 728
19.
7 1267
p=
0.091
0.000
0.000
Marital Status
Married
12.
5 2356
18.
0 2739
15.
5 5095
Single
12.
4 937
15.
8 849
14.
0 1786
Widowed
21.
4 131
25.
9 432
24.
9 563
Divorced/Separated
25.
5 427
27.
2 707
26.
5 1134
p=
0.000
0.000
0.000
Number of children
No
children
15.
2 2816
20.
4 3096
17.
9 5912
1 child
13.
2 440
19.
1 650
16.
7 1090
2 children
9.
3 454
18.
5 701
14.
9 1155
3+
children
13.
5 141
16.
4 280
15.
4 421
p=
0.008
0.303
0.055
Highest Educational Qualifications
Degree
9.
4 662
14.
7 577
11.
9 1239
Professional,
A Level
10.
7 857
17.
2 886
14.
0 1743
GCSE
or equivalent
15.2 1 2 1248
17.
9 1719
16.
8 2967
No
qualifications
18.
7 1047
25.
1 1517
22.
5 2564
p=
0.000
0.000
0.000
Employment Status
Working
full time
10.
2 2353
14.
5 1463
11.
8 3816
Working
part time
11.
6 285
16.
0 1168
15.
1 1453
Unemployed
20.
5 146
24.
6 114
22.
3 260
Economically
inactive
23.
1 1031
25.
5 1954
24.
7 2985
p=
0.000
0.000
0.000
Household Equivalised Income per week L£
< L150 £
24.
9 574
25.
7 1082
25.
4 1656
£ L150

Table II
Proportion reporting poor sleep on 4 or more nights a week by smoking, worries and health variables by
gender, age 16-74
Men Women Total
% n= % n= % n=
Cigarette Smoking
Current Smoker 19.0 1162 24.4 1402 22.0 2564
Ex Smoker 12.0 1808 18.3 1938 15.3 3746
Never Smoked 12.3 877 16.8 1383 15.0 2260
p= 0.000 0.000 0.000
Self-reported Worries
No 9.3 2612 12.5 2888 11.0 5500
Medium 14.6 591 22.2 825 19.0 1416
High 33.6 648 38.4 1014 36.5 1662
p= 0.000 0.000 0.000
Health worries
No 10.0 3228 15.3 3909 12.9 7137
Medium 28.4 342 33.7 466 31.4 808
High 45.6 281 49.7 352 47.9 633
p= 0.000 0.000 0.000
Self-reported Health
Very Good or Excellent 7.0 2111 12.3 2473 9.8 4584
Good 13.4 972 20.4 1253 17.3 2225
Fair or poor 35.1 767 37.4 999 36.4 1766
p= 0.000 0.000 0.000
Number of chronic illnesses
0 8.3 2108 12.7 2442 10.7 4550
1 16.7 1099 21.4 1296 19.2 2395
2 22.0 381 28.7 575 26.0 956
3 or more 40.0 260 44.0 411 42.5 671
p= 0.000 0.000 0.000
Source: Psychiatric Morbidity Survey, 2000 (authors' analysis)
Worries are implicated in the gender difference
in sleep problems. Women's sleep
is more likely to be disturbed by worries,
particularly associated with their gender role
as mothers or wives, and their concern for
the well-being of family members [13, 2, 31].
Previous sleep research has tended to view
'worries' as a mark of anxiety or psychological
problems, rather than embedded within
social roles and responsibilities. Worries
and concerns represent an important predictor
of sleep problems, but retain an independent
effect after controlling for health and
depression Table III, Model 5). In addition,
this analysis suggests that differences in
health status between men and women do
not explain the gender difference in sleep
problems. Indeed, the gender difference
becomes greater after adjusting for health
variables and depression (Table III, Model 5).
This research supports other studies
that have found poorer sleep quality among
those with low educational qualifications [15,
21, 24, 29], who are not working [22, 24],
and have low income [17, 9]. However, this
study goes beyond previous research in two
ways. First, by using representative national
data to simultaneously consider the independent
effects of four SES variables (education,
employment status, household income,
housing tenure). Second, it addresses previous
observations [24] that the higher prevalence
of insomnia among individuals with low
education, not working and with low income
may be confounded by poor physical and
mental health, through examining models
containing health measures and depression.
The relative importance of different sets
of factors in leading to sleep problems are
considered. 'Worries' are likely to be confounded
with socio-economic characteristics;
the relationships between sleep problems
and living on a low income or living in rented
housing were partially mediated through
worries and concerns. After adjusting for
smoking, worries, health and depression, a
significant independent association still remained
between sleep problems and both
low education and not being in paid work.
Lack of employment is linked to sleep problems
in two ways; for the unemployed, primarily
through its intrinsic relationship with
worries, while for the economically inactive,
primarily because of their poorer health status.
There is a much larger effect of marital
status on sleep problems among men than
among women, suggesting advantages
conferred by marriage for men in terms of
sleep quality. Divorced/separated men and
widowed men have particularly poor quality
sleep compared with married men. Although
losing a partner (whether through
divorce or death) has psychological consequences
which may adversely effect sleep
quality, this analysis shows that a major factor
explaining the higher rate of sleep problems
among the previously married is their
more disadvantaged SES.
Conclusions
A large part of the well-known gender
difference in reported sleep problems is
mediated by the more disadvantaged socioeconomic
status (SES) of women, casting
doubt on the primacy of physiological explanations
of this gender difference. In turn,
SES is shown to impact on psychological
distress and worries, which form part of the
mechanism through which disadvantaged
SES impacts on sleep problems.
There are gender differences in the association
of marital status with poor sleep
with greater effects of marital status for men
than women. Men who were previously married
(whether divorced or widowers) have
particularly poor sleep compared with their
married counterparts. Whereas for women,
these associations with marital status are
less strong. However, among both men
and women, a substantial proportion of the
higher reported sleep problems of the divorced
and widowed can be accounted for
by their more disadvantaged SES. These
findings that people with more disadvantaged
SES report greater sleep problems
need further consideration by health researchers.
Despite sleep being important for health
and well-being [3], previous research on
gender inequalities in health has hitherto
paid scant attention to gender differences
in sleep problems or to how women's disadvantaged
social position or gender roles
mediate these differences. This research
has shown strong linkages both between
socio-economic variables and reported
sleep problems, and between health variables
and sleep problems. This suggests
that low socio-economic status, family worries
and women's family roles, may potentially
be mechanisms that account for
women's greater experience of sleep disruption.
Acknowledgements
The author acknowledges funding from
the New Dynamics of Ageing initiative, a
multidisciplinary research programme sup-
58 Przegl¹d Lekarski 2012 / 69 / 2 S. Arber

Table III
Odds Ratios of poor sleep quality on 4 or more nights per week (n= 8240).
Model 1
Age +
Sex
Model 2
+
Demographic
-2 Log likelihood 7481.59 7415.25 7186.16 6711.49 6268.64
Ddf 6 6 12 4 9
Nagelkerke R Square .017 0.030 0.075 0.164 0.242
* p

25. Seeman T., Merkin S.S., Crimmins, E. et al.: Education,
income and ethnic differences in cumulative
biological risk profiles in a national sample of US
adults: NHANES III (1988-1994). Soc. Sci. Med.
2008, 66, 72
26. Sekine M., Chandola T., Martikainen P. et al.: Work
and family characteristics as determinants of socioeconomic
and gender inequalities in sleep: The Japanese
Civil Servants Study. SLEEP 2006, 29, 206.
27. Shishehbor M.H., Litaker D., Pothier C.E., Lauer
M.S.: Association of socio-economic status with functional
capacity, heart rate recovery, and all-cause
mortality. JAMA 2006, 295, 784.
28. Singleton N., Bumpstead R., O'Brien M., Lee A.,
Meltzer H.: Psychiatric Morbidity among Adults living
in Private Households, 2000, Office for National
Statistics, London: The Stationery Office, 2001.
29. Stewart R., Besset A., Bebbington R. et al.: Insomnia
comorbidity and impact and hypnotic use by
age group in a national survey population aged 16
to 74. SLEEP 2006, 29, 1391
30. Ustun T.B.: Cross-national epidemiology of depression
and gender. J. Gend. Specif. Med. 2000, 3, 54.
31. Venn S., Arber S., Meadows R. Hislop J.: The
Fourth Shift: Exploring the gendered nature of sleep
disruption among couples with children. Br. J. Sociol.
2008, 59, 79.
32. Vitiello M.V., Moe K.E., Prinz P.N.: Sleep complaints
cosegregate with illness in older adults. Clinical research
informed by and informing epidemiological
studies of sleep. J. Psychosom. Res. 2002, 53, 555.
33. Zhang B., Wing Y-K.: Sex differences in insomnia:
A meta-analysis. SLEEP 2006, 29, 85.
60 Przegl¹d Lekarski 2012 / 69 / 2 S. Arber