The Long-term Effect of Insomnia on Work Disability The HUNT-2 Historical Cohort Study

Article (PDF Available)inAmerican Journal of Epidemiology 163(11):1018-24 · July 2006with178 Reads
DOI: 10.1093/aje/kwj145 · Source: PubMed
Abstract
Chronic insomnia is common in the general population. Its effect on functioning and disability is usually attributed to an underlying condition, so the diagnosis of insomnia does not qualify for award of a disability pension in the United States or Europe. The aim of this study was to investigate whether insomnia, defined according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, contributed to long-term work disability. Using a historical cohort design, the authors gathered baseline data from a population-based Norwegian health study of 37,308 working-age people not claiming a disability pension through 1995–1997. The outcome was subsequent award of a disability pension (18–48 months after the health screening) as registered by the National Insurance Administration. Insomnia was a strong predictor of subsequent permanent work disability (adjusted odds ratio = 3.90, 95% confidence interval: 3.20, 4.76). Sociodemographic and shift-work characteristics had little confounding effect (adjusted odds ratio = 3.69, 95% confidence interval: 3.00, 4.53), and this association remained significant after adjustment for psychiatric and physical morbidity and for health-related behaviors (adjusted odds ratio = 1.75, 95% confidence interval: 1.40, 2.20). This study suggests that insomnia should receive increased attention as a robust predictor of subsequent work disability.
Original Contribution
The Long-term Effect of Insomnia on Work Disability
The HUNT-2 Historical Cohort Study
Børge Sivertsen
1
, Simon Overland
2
, Dag Neckelmann
3
, Nicholas Glozier
4
, Steinar Krokstad
5
,
Sta
˚
le Pallesen
6,7
, Inger Hilde Nordhus
1,7
, Bjørn Bjorvatn
7,8
, and Arnstein Mykletun
2,4,9
1
Department of Clinical Psychology, University of Bergen, Bergen, Norway.
2
Research Centre for Health Promotion, University of Bergen, Bergen, Norway.
3
Affective Disorder Section, Department of Psychiatry, Haukeland University Hospital, Bergen, Norway.
4
Division of Psychological Medicine, Institute of Psychiatry, King’s College London, London, United Kingdom.
5
HUNT Research Centre, Norwegian University of Technology and Science, Trondheim, Norway.
6
Department of Psychosocial Science, University of Bergen, Bergen, Norway.
7
Norwegian Competence Center for Sleep Disorders, University of Bergen, Bergen, Norway.
8
Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway.
9
Department of Mental Health, Division of Epidemiology, Norwegian Institute of Public Health, Oslo, Norway.
Received for publication September 19, 2005; accepted for publication January 5, 2006.
Chronic insomnia is common in the general population. Its effect on functioning and disability is usually attributed
to an underlying condition, so the diagnosis of insomnia does not qualify for award of a disability pension in the
United States or Europe. The aim of this study was to investigate whether insomnia, defined according to the
Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, contributed to long-term work disability.
Using a historical cohort design, the authors gathered baseline data from a population-based Norwegian health
study of 37,308 working-age people not claiming a disability pension through 1995–1997. The outcome was
subsequent award of a disability pension (18–48 months after the health screening) as registered by the National
Insurance Administration. Insomnia was a strong predictor of subsequent permanent work disability (adjusted odds
ratio ¼ 3.90, 95% confidence interval: 3.20, 4.76). Sociodemographic and shift-work characteristics had little
confounding effect (adjusted odds ratio ¼ 3.69, 95% confidence interval: 3.00, 4.53), and this association remained
significant after adjustment for psychiatric and physical morbidity and for health-related behaviors (adjusted odds
ratio ¼ 1.75, 95% confidence interval: 1.40, 2.20). This study suggests that insomnia should receive increased
attention as a robust predictor of subsequent work disability.
cohort studies; disability evaluation; logistic models; risk factors; sleep initiation and maintenance disorders; work
Abbreviations: CI, confidence interval; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition;
HUNT-2, Nord-Trøndelag Health Study; OR, odds ratio.
Difficulties in initiating or maintaining sleep are common
symptoms in the population. In the adult population of the
United Kingdom, 29 percent report sleep problems over the
past week (1). A diagnosis of insomnia is less common,
but several studies show that approximately 10 percent of
the adult population in Western societies suffers from
chronic insomnia (2, 3), as specified in the Diagnostic and
Statistical Manual of Mental Disorders, Fourth Edition
(DSM-IV) (4). Impaired sleep may be secondary to both
physical and mental disorders (5), but about 25 percent of
Reprint requests to Arnstein Mykletun, Research Centre for Health Promotion, University of Bergen, Christiesgt 13, N-5015 Bergen, Norway
(e-mail: A.Mykletun@iop.kcl.ac.uk; Arnstein.Mykletun@psyhp.uib.no).
1018 Am J Epidemiol 2006;163:1018–1024
American Journal of Epidemiology
Copyright
ª 2006 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved; printed in U.S.A.
Vol. 163, No. 11
DOI: 10.1093/aje/kwj145
Advance Access publication April 12, 2006
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
chronic insomniacs are assumed to suffer from primary in-
somnia (6).
The adverse individual consequences of insomnia are well
documented. Poor sleep is associated with cognitive and in-
tellectual impairment (7–9), current and subsequent affective
disorders (10, 11), and reduced immune function (12). More-
over, patients suffering from insomnia commonly report
a significant reduction in quality of life (13) and impaired
coping abilities (14, 15).
Despite substantial evidence that insomnia has significant
negative effects on various aspects of human functioning, sur-
prisingly few studies have focused on the societal impact of
insomnia. The economic costs resulting from sleep-related
accidents and lost productivity are estimated to be $92–
$107 billion annually (16) in the United States alone, includ-
ing such government-financed expenses as medical treatment
and drugs. However, the relation between insomnia and work
disability has received little attention, and insomnia is at pres-
ent not accepted as sufficient for issuing disability pensions.
In cross-sectional surveys, poorer sleep has been shown to be
associated with lower self-assessed working capacity and
employment status (17), dissatisfaction with work and high
levels of work stress (18), and self-reported sick leave (19, 20).
In contrast, a cohort study of Norwegian nurses found no
effect of sleep complaints on medium-term sickness absence
3 months later (21). To our knowledge, the only study so far
that has examined the effect of impaired sleep on subsequent
long-term work functioning found a single survey item mea-
suring subjective sleep quality to be a significant predictor for
self-reported sickness absence 4 years later, even after control-
ling for symptoms of both somatic and mental diseases (21).
The major problems of the published studies are 1) cross-
sectional designs, which limit understanding of causality;
2) disparate definitions of insomnia, reducing the interpret-
ability of previous results; and 3) that work disability is as-
sessed by self-report only. The aim of our study was to
estimate the direct effect of DSM-IV–defined insomnia on
permanent work disability using a historical cohort design
and, furthermore, to adjust the effect for physical and mental
symptoms and conditions as well as other potential socio-
demographic confounders and shift work.
MATERIALS AND METHODS
Design
A historical cohort study using baseline data on mental
and somatic health obtained in the Nord-Trøndelag Health
Study (HUNT-2) was carried out from August 1995 until
June 1997. The outcome variable was award of a disability
pension 18–48 months after participation in HUNT-2. By
excluding all disability pensions awarded 0–18 months after
participation in HUNT-2, we aimed to exclude subjects in the
process of applying for a disability pension while they at-
tended HUNT-2, thus reducing any possible protopathic bias.
Participants and procedures
All inhabitants of Nord-Trøndelag County, Norway, aged
20 years or older were invited to a clinical examination as
part of a general health screening. In all, 92,100 persons
aged 20–89 years were sent an initial questionnaire and an
invitation to participate in HUNT-2. Of these, 65,648 (71
percent) attended the physical examination, where they re-
ceived a second set of questionnaires, which 52,814 (80
percent) completed. Retired persons and individuals reach-
ing the retirement age of 67 years during follow-up were
excluded (n ¼ 11,123); retirement because of old age pre-
cludes the award of a disability pension. HUNT-2 respond-
ers who were receiving a disability pension at baseline (n ¼
3,964) or who were granted a disability pension within 18
months after baseline (n ¼ 419) were also excluded. Thus,
the final study population included 37,308 persons: 19,936
women and 17,372 men. In a follow-up study of random
dropouts (22), the two most common reasons for not attend-
ing the health screening were 1) not having time or experi-
encing a need for the health examinations and 2) grave
disease. No information on disability awards was available
for the nonresponders.
Measures
Disability pension award (dependent variable).
The
National Insurance Administration records all grants of
disability pensions, which, in Norway, is solely a public
responsibility. Correct registration is a prerequisite for
transfers of payment s; thus, the records are highly accurate.
The criterion for being awarded a disabili ty pension is an
application from the general practitioner stating cause-
specific and lasting reduced functional ability due to an ac-
knowledged medical condition. Further examinations from
organ-specific specialists are generally undertaken when ap-
propriate, although independent examination is not required.
In the last decade, approximately 20 percent of applications
were rejected by the insurance case managers, most often
because of incomplete evaluations or because the applicant
attempted to obtain another job and/or received medical re-
habilitation before the application was submitted.
Insomnia (exposure). The DSM-IV (4) criteria for in-
somnia include difficulty falling asleep, difficulty maintain-
ing sleep, or experiencing nonrestorative sleep for a period
of no less than 1 month. In addition, it is a prerequisite that
the sleep disturbance significantly impairs daily funct ioning.
In this study, a proxy for the DSM-IV insomnia diagnosis
was based on three items, encompassing persons reporting
sleep-onset or maintenance insomnia ‘often’ or ‘almost
every night’ in the last month, in addition to reporting im-
paired work performance caused by the sleep problems dur-
ing the last year.
To account for a possible circularity in classifying the out-
come variable (permanent work disability) by the item on
daytime work impairment, a second category, DSM-IV in-
somnia without daytime impairment, was constructed, other-
wise similar to the former category. The remai ning persons
were categorized as ‘unlike ly’ to suffer from insomnia. All
three categories were mutually exclusive.
Anxiety and depression. The Hospital Anxiety and De-
pression Scale (23) is a self-report questionnaire comprising
14 four-point Likert-scaled items, seven for anxiety (Hospital
Anxiety and Depression Scale-A) and seven for depression
Insomnia and Work Disability 1019
Am J Epidemiol 2006;163:1018–1024
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
(Hospital Anxiety and Depression Scale-D). No somatic
items or items regarding sleeping difficulties are included.
In a recent literature review covering 31 studies, the Hospital
Anxiety and Depression scale was found to have good case-
finding properties for anxiety and depression in patient pop-
ulations in primary care and hospital settings (24). A cutoff
score of 8 on either subscale provides an optimal balance
between sensitivity and specificity at about 0.80 for both
depression and anxiety according to the Diagnostic and
Statistical Manual of Mental Disorders, Third Edition, and
DSM-IV and to the International Classification of Diseases,
Eighth Revision and Ninth Revision. By using these cutoffs,
four groups were identified: baseline anxiety only, baseline
depression only, baseline comorbid anxiety and depres-
sion, and a reference group (no disorder). In addition, a
dummy variable, indicating anxiety and/or depression, was
constructed for tests of interactions with age and gender.
Physical health in HUNT-2. Physical health was assessed
in two separate ways, as described previously by Mykletun
et al. (25):
1. The index for somatic symptoms was computed as the
sum of organ systems from which symptoms were re-
ported. These organ systems were weighted as described
in the analyses. Included were gastrointestinal symptoms
(four questions on nausea, heartburn, diarrhea, and con-
stipation); musculoskeletal symptoms (pain in the neck,
shoulders, elbows, hands, breasts, back (three areas),
hips, knees, and ankles); the head (two questions on
headache and migraine), the senses (two questions on
hearing and sight); the heart (one question on palpita-
tions); and respiratory function (one question on re-
spiratory problems).
2. The index for self-reported somatic diagnoses comprised
asthma, angina pectoris, stroke, myocardial infarction,
diabetes, goiter, hypo- and hyperthyroid function, other
diseases of the thyroid gland, fibromyalgia, osteoporosis,
arthritis, rheumatism, ankylosing spondylitis, cancer, epi-
lepsy, high blood pressure (being treated or monitored),
and any other illness (one open item).
Sociodemographic variables and health-related behavior.
Information on gender and on age at the time of the HUNT-2
study was obtained from the national population registry.
Socioeconomic status and educational level have previously
been shown to be associated with receiving a disability pen-
sion (26). Socioeconomic status was measured by using a
validated approximation from the HUNT-2 study (27). Ed-
ucational level (three levels), night/shift work, daily ciga-
rette smoking, consuming too much alcohol during the last
14 days, and being physically active for 1 or more hours in
the last week were recorded by self-report.
Statistics
The physical health measures (indices for symptoms and
diagnoses) were weighted for their partial associations with
physician-certified sick leave; two separate regression mod-
els were used to estimate weights for the diagnoses and
symptoms included in the indices. Unstandardized regres-
sion coefficients were used as weights, and the indices were
computed as a sum of products between the standardized
symptoms and diagnoses and their weights.
Logistic regression analysis was used to examine the re-
lation between insomnia and award of a disabili ty pension.
In this paper, results are presented as odds ratios with 95
percent confidence intervals. Possible confounders were en-
tered into the model for adjustment in the following order:
gender and age, sociodemographic factors, night/shift work,
health behaviors, anxiety and depression (using ordinal
transformations), somatic diagnosis, and somatic symptoms.
Pearson chi-square tests were used to examine differences
between baseline demographic and clinical characteristics
in 1995–1997 and permanent work disability at follow-up.
The Pearson chi-square test was also used to examine the
causes of permanent work disability for those experiencing
or not experiencing insomnia.
Ethics
HUNT-2 was approved by the National Data Inspectorate
and the Board of Research Ethics in Health Region IV of
Norway. Informed consent in writing was obtained from all
subjects included in this study.
RESULTS
Sample characteristics
The baseline characteristics of the 37,308 persons who
completed the insomnia questionnaire are shown in table 1
stratified by diagnostic classification. The prevalence rate for
insomnia was 4.4 percent, while a further 5.8 percent of par-
ticipants experienced insomnia without impairment. Insom-
nia was more prevalent among women and persons with
lower educational levels, and sleep problems increased sig-
nificantly with older age. Greater cigarette use and alcohol
intake, and less exercise, were associated with increasing re-
ports of sleep problems at baseline. Both self-reported so-
matic symptoms and diagnosis were strongly associated with
increasing sleep disturbances, as were anxiety and depres-
sion (table 1).
Predictors of a disability pension award
In all, 915 persons were granted a permanent disability
pension 18–48 months after baseline assessment. Female
gender, older age, and the health-related behaviors of smok-
ing, excessive consumption of alcohol, and being physically
inactive were associated with a greater likelihood of receiv-
ing a subsequent disability pension, as were, unsurprisingly,
self-reported symptoms, somatic diagnoses, anxiety, and de-
pression (table 2).
The predictive effect of insomnia on disability
Persons with insomnia and impairment had a strongly ele-
vated risk of subsequently being awarded a disability pen-
sion during follow-up (odds ratio (OR) ¼ 3.90, 95 percent
1020 Sivertsen et al.
Am J Epidemiol 2006;163:1018–1024
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
confidence interval (CI): 3.20, 4.76; table 3). This effect was
only marginally attenuated after adjustment for age, gender,
and baseline socio demographic variables and health behav-
iors (OR ¼ 3.57, 95 percent CI: 2.91, 4.39). Adjusting for
anxiety and depression explained a substantial part of this
effect (OR ¼ 2.29, 95 percent CI: 1.83, 2.87), and further
adjustment for somatic diagnoses reduced the odds ratio to
2.02 (95 percent CI: 1.61, 2.53). The confo unding effects of
somatic symptoms further explained some of this associa-
tion, although the effect of insomnia remained highly sig-
nificant (OR ¼ 1.75, 95 percent CI: 1.40, 2.20; p < 0.001).
Persons with insomnia but no impairment showed a sig-
nificant, but reduced risk of subsequently being awarded a
disability pension (OR ¼ 2.45, 95 percent CI: 1.99, 3.02).
However, adjustment for the above confounders suggested
no independent effect on disability pension award (OR ¼
1.19, 95 percent CI: 0.95, 1.49).
Musculoskeletal disorders were the most prevalent causes,
accounting for almost 50 percent of all new disability awards
(table 4) in all three groups. Psychiatric disorders were a sig-
nificantly more prevalent cause among persons suffering
from insomnia with impairment (23.4 percent) compared
with insomnia without impairment (13.3 percent ) and no
insomnia (13.6 percent) (v
2
¼ 10.79, df ¼ 4, p ¼ 0.03).
The remaining causes, cancer and cardiovascular diseases
being the most prevalent, accounted for only 29 percent in
the insomnia group compared with 40.5 percent in the group
with no insomnia.
DISCUSSION
Using a historical cohort design, we found that insomnia
at baseline was a significant risk factor for subsequent
award of a disability pension. Almost half of this effect was
not explained by the sleep disturbance being associated
with anxiety and depressio n, somatic diagnoses, and somatic
symptoms.
TABLE 1. Baseline demographic and clinical characteristics of persons who completed the insomnia
questionnaire in the Nord-Trøndelag Health Study, Norway, 1995–1997*,y
Characteristic Unlikely insomnia
Insomnia without daytime
work impairment
Insomnia
No. of persons (%) 33,508 (89.8) 2,155 (5.8) 1,645 (4.4)
Age (years)z 41.9 (41.8, 42.1) 46.8 (46.3, 47.4) 43.4 (42.8, 44.0)
Gender: female (%) 52.9 58.1 58.4
Socioeconomic status (%)
I (highest) 9.4 8.8 8.4
II 17.4 13.2 17.7
III 19.4 19.0 19.4
IV 15.7 15.0 15.1
V þ VI 12.0 11.1 9.2
VII (lowest) 15.7 19.4 17.5
Not working or not applicable 10.4 13.5 12.6
Education (%)
Compulsory school only 23.9 23.9 29.4
College 50.0 50.0 47.4
University level 26.0 26.0 23.3
Night/shift work (%) 19.7 19.2 21.0
Daily cigarette smoking (%) 18.8 23.1 23.1
Consumption of too much alcohol in the
last 14 days (%) 11.4 15.0 16.0
Physically active 1 hour during the
last week (%) 76.3 72.9 70.0
Anxietyz 3.90 (3.86, 3.93) 5.61 (5.46, 5.76) 7.95 (7.75, 8.15)
Depressionz 2.87 (2.84, 2.89) 4.21 (4.07, 4.34) 5.70 (5.51, 5.88)
No. of somatic diagnosesz 0.25 (0.25, 0.26) 0.42 (0.39, 0.45) 0.47 (0.43, 0.50)
No. of self-reported symptomsz 1.46 (1.45, 1.47) 2.01 (1.96, 2.06) 2.53 (2.47, 2.58)
Disability pension at follow-up
(% (no. of persons)) 2.0 (686) 4.9 (105) 7.5 (124)
* All group differences for insomnia, except night/shift work, were statistically significant (p < 0.001).
y Some percentages do not total 100 because of rounding.
z Values are expressed as mean (95% confidence interval).
Insomnia and Work Disability 1021
Am J Epidemiol 2006;163:1018–1024
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
Methodological aspects
The present study has several strengths arising mainly
from the cohort design. The study sample was large, and
the participation rate at baseline was high. Both exposure
and outcome assessments should b e relatively unbiased. At
baseline measurement, neither participants nor administra-
tors were aware of the specific researc h hypotheses, which
reduced the possibility of information being biased by se-
lective symptom pres entation in order to gain access to, or
avoid, benefits. Disability pension status at baseline and at
follow-up was ascertained from the National Insurance Ad-
ministration. These data are complete (those persons mov-
ing to other parts of the country after participating in
HUNT-2 are still registered) and should not have been in-
fluenced by exposure status.
There are several limitations of our study. First, no data on
disability awards were available for the nonresponse group,
and selection bias cannot be ruled out. In the Hordaland
Health Study (HUSK; comparable to the HUNT study, but
the age range of participants was more limited), the rate of
disability pensioning was higher among the nonattendees
than among participants, which is likely also the case in the
HUNT study. Second, measurement of insomnia was estab-
lished by self-report rather than clinical diagnosis and did
not encompass an item on the duration of insomnia. Beyond
the duration criterion, the questionnaire used in this study
was tailored to meet the criteria for insomnia as specified
by DSM-IV. The prevalence rate found in the present study
is still som ewhat lower than previously reported, suggesting
a conservative estimate (2, 3). The absence of the duration
TABLE 2. Baseline demographic and clinical characteristics,
and cumulative incidence of disability pensions at follow-up,
for persons in the Nord-Trøndelag Health Study, Norway,
1995–1997
Characteristic
No. of
cases
Cumulative
incidence of
disability pensions
at follow-up
%
No. of
cases
Age (years)**
20–29 6,776 0.3 23
30–39 9,096 0.9 79
40–49 10,290 1.7 172
50–59 7,465 5.4 402
60–65 3,681 6.5 239
Gender*
Female 19,936 2.6 527
Male 17,372 2.2 388
Socioeconomic status**
I (highest) 3,487 1.5 54
II 6,405 2.0 129
III 7,229 2.6 187
IV 5,818 2.7 159
V þ VI 4,397 1.8 81
VII (lowest) 5,978 3.2 189
Not working or not
applicable 3,994 2.9 116
Education**
Compulsory school only 9,314 5.0 464
College 18,501 1.7 316
University level 9,493 1.4 135
Night/shift work*
No 29,963 2.6 765
Yes 7,345 2.0 150
Daily cigarette smoking**
No 30,136 2.1 634
Yes 7,172 3.9 281
Consumption of too much
alcohol in the last 14 days**
No 32,898 2.3 763
Yes 4,410 3.4 152
Physically active 1 hour
during the last week**
No 9,002 3.5 316
Yes 28,306 2.1 599
Anxiety**
HADSy-A score <8 32,011 2.1 676
HADS-A score 8 5,297 4.5 239
Depression**
HADS-D score <8 34,370 2.2 741
HADS-D score 8 2,938 5.9 174
No. of somatic diagnoses**
0 29,117 1.5 450
1 6,543 4.9 321
2 1,272 9.0 114
3 376 8.0 30
No. of self-reported symptoms*
0 8,365 0.9 73
1 11,154 1.9 212
2 9,582 2.7 256
3 5,720 3.7 211
4 2,487 6.6 163
* p < 0.05; **p < 0.001.
y HADS, Hospital Anxiety and Depression Scale.
TABLE 3. Effect of insomnia on risk of permanent work
disability for persons in the Nord-Trøndelag Health Study,
Norway, 1995–1997*
Adjustment variables
Insomnia without
daytime work
impairment
(n ¼ 2,155; 5.8%)
Insomnia
(n ¼ 1,645; 4.4%)
ORy 95% CIy OR 95% CI
Crude 2.45 1.99, 3.02 3.90 3.20, 4.76
þ Age and gender 1.77 1.42, 2.19 3.70 3.02, 4.54
þ Socioeconomic
status and
education 1.71 1.38, 2.12 3.69 3.00, 4.53
þ Night/shift work 1.71 1.38, 2.12 3.69 3.00, 4.53
þ Smoking, alcohol
consumption, and
physical activity 1.70 1.37, 2.11 3.57 2.91, 4.39
þ Anxiety and
depression 1.40 1.13, 1.75 2.29 1.83, 2.87
þ Somatic diagnoses 1.28 1.03, 1.60 2.02 1.61, 2.53
þ Somatic symptomsz 1.19 0.95, 1.49 1.75 1.40, 2.20
* All new disability pensions: n ¼ 915; population at risk: n ¼
37,308.
y OR, odds ratio; CI, confidence interval.
z Weighted for absence due to sickness 1 year before baseline
measurement.
1022 Sivertsen et al.
Am J Epidemiol 2006;163:1018–1024
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
criterion in our questions on insomnia may have reduced
the specificity of the measure and thus underestimated the
true association. However, when the curvilinear association
between age and insomnia is considered, the prevalence
rates most likely depict a reasonable estimate of insomnia
because our population consisted of all working persons
under the age of 67 years.
Third, our adjustments in the final analyses most likely
did not capture and fully attenuate all possible confounding
from chronic somatic or psychiatric conditions. Self-report
instruments are prone to error, and residual confounding
cannot be ruled out. Screening for psychiatric morbidity
was limited to symptoms of anxiety and depression only
and was conducted by using a relatively short inventory
not including vegetative symptoms of anxiety and depres-
sion. In some cases, depression may be largely presented
and experienced through these kinds of symptoms (28).
Thus, those with subthreshold depression may have been
overrepresented within the insomniacs, thus underestimat-
ing the true confounding effect of psychiatric morbidity.
Furthermore, the impacts of, for example, psychoses or
other psychiatric conditions are likely to have been only
partly captured. Although the subjects were given one open
answer category for ‘other diseases, this option is not
sufficient to capture all relevant diseases and symptoms
because it demands subjective insight into psychiatric mor-
bidity. However, it is commonly held that serious psycho -
pathology increases the risk of being a nonattendee in
epidemiologic studies (29), thus limiting the potential prob-
lem of psychopathology not being registered by the pro-
cedures we used. Information on somatic diagnoses and
symptoms was self-reported, and the categories used were
not exhaustive. If such diagnoses and symptoms were under-
reported, the effects of insomnia may in turn have been
overestimated. However, in our fully adjusted model, in-
somnia was considered to predict disability pension only
if reported somatic diagnoses , somatic symptoms, anxiety,
or depression did not serve as a possible explanation. Be-
cause this model assumed that the insomnia was secondary
in all instances of simultaneous occurrences, there is a risk
that we underestimated the strength of the association of
insomnia with later disability pension.
Finally, other potential confounders such as incipient cog-
nitive decline, poor coping, and work pressures were not
captured in our measures, but they may be relat ed to both
baseline insomnia (sleep disturbance symptoms are far more
prevalent in the population than the diagnoses used in these
regression models) and later disability pension. The lower
association of disability with insomnia without impairment
suggests that work problems and other functional limita-
tions may be acting as unmeasured confounders.
Insomnia and disability
To our knowledge, the present study is the first to docu-
ment an independent association between insomnia and sub-
sequent permanent work disability in a prospective study. In
contrast to all previous studies, ours used official records of
disability pensions as the measure of work disability. Self-
reports of work disability are likely to be less accurate than
official records, leaving open the possibility that subjective
experience of work disability is being measured, not de
facto work disability. Because previous research largely
has been cross-sectional, the direction of the relation and
causal pathways has been difficult to determine (17). The
association between poor sleep and work disability is com-
monly thought to be mediated through somatic or psycho-
logical factors, which may explain previous neglect of this
topic. This assumption is well founded because, in addition
to bein g linked to various medical conditions (30), impaired
sleep often is a core symptom of several psychiatric disor-
ders (4) and has been shown to precede both depression and
anxiety (31, 32). Psychological morbidity, in turn, has been
shown to play a considerable role in explaining permanent
work disability (25). However, although we cannot rule out
the possibility of residual confounding, our study suggests
that insomnia is inde pendently associated with subsequent
work disability and that this relation remains significant
even after adjustment for a range of clinical and demo-
graphic confounders. Furthermore, common behavioral fac-
tors often linked with sleep disturbance, such as shift work,
physical exercise, and alcohol intake, did not attenuate or
mediate this association. This finding suggests that the link
may be through unmeasured factors warrantin g further ex-
ploration, and policy attempts to reduce the rising disability
pension burden in Organisation for Economic Co-operation
and Development (OECD, Paris, France) countries need to
acknowledge this situation. Because insomnia is associated
with several social dysfunctions, establishing effective pre-
ventive strategies and low-threshold treatment options for
insomnia might be beneficial.
TABLE 4. Comparison of causes of disability in persons with and without insomnia in the Nord-Trøndelag
Health Study, Norway, 1995–1997
Primary cause
Reference group (n ¼ 686)
Insomnia without daytime
work impairment (n ¼ 105)
Insomnia with daytime
work impairment (n ¼ 124)
No. of
cases
%95%CI*
No. of
cases
% 95% CI
No. of
cases
% 95% CI
Psychiatric disorders 93 13.6 11.0, 16.1 14 13.3 16.8, 19.8 29 23.4 15.9, 30.8
Musculoskeletal disorders 315 45.9 42.2, 49.6 51 48.6 39.0, 58.1 59 47.6 38.8, 56.4
Other causes 278 40.5 36.9, 44.2 40 38.1 28.8, 47.4 36 29.0 21.0, 37.0
* CI, confidence interval.
Insomnia and Work Disability 1023
Am J Epidemiol 2006;163:1018–1024
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
ACKNOWLEDGMENTS
Conflict of interest: none declare d.
REFERENCES
1. Singleton N, Bumpstead R, O’Brien M, et al. Psychiatric
morbidity among adults living in private households, 2000.
London, United Kingdom: The Stationery Office, 2001.
2. Pallesen S, Nordhus IH, Nielsen GH, et al. Prevalence of in-
somnia in the adult Norwegian population. Sleep 2001;24:
771–9.
3. Ohayon MM. Prevalence of DSM-IV diagnostic criteria of
insomnia: distinguishing insomnia related to mental disorders
from sleep disorders. J Psychiatr Res 1997;31:333–46.
4. American Psychiatric Association. Diagnostic and statistical
manual of mental disorders: DSM-IV. 4th ed. Washington, DC:
Americal Psychiatric Association, 1994.
5. Roth T, Roehrs T. Insomnia: epidemiology, characteristics, and
consequences. Clin Cornerstone 2003;5:5–15.
6. Bixler EO, Vgontzas AN, Lin HM, et al. Insomnia in central
Pennsylvania. J Psychosom Res 2002;53:589–92.
7. Pilcher JJ, Huffcutt AI. Effects of sleep deprivation on per-
formance: a meta-analysis. Sleep 1996;19:318–26.
8. Simon GE, VonKorff M. Prevalence, burden, and treatment of
insomnia in primary care. Am J Psychiatry 1997;154:1417–23.
9. Szelenberger W, Niemcewicz S. Severity of insomnia corre-
lates with cognitive impairment. Acta Neurobiol Exp (Wars)
2000;60:373.
10. Mellinger GD, Balter MB, Uhlenhuth EH. Insomnia and its
treatment. Prevalence and correlates. Arch Gen Psychiatry
1985;42:225–32.
11. Vollrath M, Wicki W, Angst J. The Zurich study. VIII. Insom-
nia: association with depression, anxiety, somatic syndromes,
and course of insomnia. Eur Arch Psychiatry Neurol Sci
1989;239:113–24.
12. Savard J, Laroche L, Simard S, et al. Chronic insomnia and
immune functioning. Psychosom Med 2003;65:211–21.
13. Zammit GK, Weiner J, Damato N, et al. Quality of life in
people with insomnia. Sleep 1999;22(suppl 2):S379–85.
14. Morin CM, Rodrigue S, Ivers H. Role of stress, arousal, and
coping skills in primary insomnia. Psychosom Med 2003;65:
259–67.
15. Sadeh A, Keinan G, Daon K. Effects of stress on sleep: the
moderating role of coping style. Health Psychol 2004;23:
542–5.
16. Stoller MK. Economic effects of insomnia. Clin Ther 1994;
16:873–97; discussion 854.
17. Linton SJ, Bryngelsson I. Insomnia and its relationship to
work and health in a working-age population. J Occup Rehabil
2000;10:169–83.
18. Rosmond R, Lapidus L, Bjorntorp P. A cross-sectional study of
self-reported work conditions and psychiatric health in native
Swedes and immigrants. Occup Med (Lond) 1998;48:309–14.
19. Leigh JP. Employee and job attributes as predictors of absen-
teeism in a national sample of workers: the importance of
health and dangerous working conditions. Soc Sci Med
1991;33:127–37.
20. Ingemarsson AH, Nordholm L, Sivik T. Risk of long-term
disability among patients with back pain. Scand J Rehabil Med
1997;29:205–12.
21. Eriksen W, Natvig B, Bruusgaard D. Sleep problems: a pre-
dictor of long-term work disability? A four-year prospective
study. Scand J Public Health 2001;29:23–31.
22. HUNT Research Center. Non-responder study from HUNT 2,
2005. (http://www.hunt.ntnu.no/index.php?side¼forskning/
resultater/deskres/frammote/nonrespond).
23. Zigmond AS, Snaith RP. The hospital anxiety and depression
scale. Acta Psychiatr Scand 1983;67:361–70.
24. Bjelland I, Dahl AA, Haug TT, et al. The validity of the
Hospital Anxiety and Depression Scale. An updated literature
review. J Psychosom Res 2002;52:69–77.
25. Mykletun A, Øverland S, Dahl AA, et al. A population-based
cohort study of the effect of common mental disorder on
disability pension awards. Am J Psychiatry (in press).
26. Krokstad S, Westin S. Disability in society—medical and non-
medical determinants for disability pension in a Norwegian
total county population study. Soc Sci Med 2004;58:1837–48.
27. Krokstad S, Westin S. Health inequalities by socioeconomic
status among men in the Nord-Trøndelag Health Study,
Norway. Scand J Public Health 2002;30:113–24.
28. Davidson J, Turnbull CD. Diagnostic significance of vegeta-
tive symptoms in depression. Br J Psychiatry 1986;148:442–6.
29. Eaton WW, Holzer CE 3rd, Von Korff M, et al. The design of
the Epidemiologic Catchment Area surveys. The control and
measurement of error. Arch Gen Psychiatry 1984;41:942–8.
30. Shapiro CM, Devins GM, Hussain MR. ABC of sleep disor-
ders. Sleep problems in patients with medical illness. BMJ
1993;306:1532–5.
31. Ford DE, Kamerow DB. Epidemiologic study of sleep dis-
turbances and psychiatric disorders. An opportunity for pre-
vention? JAMA 1989;262:1479–84.
32. Vgontzas AN, Kales A. Sleep and its disorders. Annu Rev Med
1999;50:387–400.
1024 Sivertsen et al.
Am J Epidemiol 2006;163:1018–1024
by guest on June 12, 2013http://aje.oxfordjournals.org/Downloaded from
    • "Reduced efficiency at work and absenteeism due to insomnia increase over time. The relationship between efficiency at work and sleep quality is reciprocal; co-morbidities, self-reported poor efficiency at work and claims for earlier retirement because of acquired disability are more frequent in poor sleepers [105][106][107][108][109][110]. The prevalence of complaints about insomnia, poor sleep, and daytime sleepiness is higher in shift-workers than among those working on regular schedules [111][112][113]; in turn, sleep disorders reduce adaptation to shift schedules and increase the occurrence of accidents at work. "
    [Show abstract] [Hide abstract] ABSTRACT: Sleep disorders are frequent (18%-23%) and constitute a major risk factor for psychiatric, cardiovascular, metabolic or hormonal co-morbidity and mortality. Low social status or income, unemployment, life events such as divorce, negative lifestyle habits, and professional requirements (e.g., shift work) are often associated with sleep problems. Sleep disorders affect the quality of life and impair both professional and non-professional activities. Excessive daytime drowsiness resulting from sleep disorders impairs efficiency and safety at work or on the road, and increases the risk of accidents. Poor sleep (either professional or voluntary) has detrimental effects comparable to those of major sleep disorders, but is often neglected. The high incidence and direct/indirect healthcare and welfare costs of sleep disorders and poor sleep currently constitute a major medical problem. Investigation, monitoring and strategies are needed in order to prevent/reduce the effects of these disorders.
    Full-text · Article · Aug 2016
    • "The two most common medico-legal causes of disability pension (DP) are musculoskeletal and mental disorders (OECD, 2010). Non-medical risk factors of DP include female gender and higher age (OECD, 2010), low socio-economic status (Bruusgaard et al., 2010; Leinonen et al., 2011), smoking (Husemoen et al., 2004; Neovius et al., 2010), low leisure-time physical activity (Robroek et al., 2013), high body mass index (Neovius et al., 2010; Robroek et al., 2013), sleep problems (Sivertsen et al., 2006; Lallukka et al., 2011), high physical workload (Labriola et al., 2009; Lahelma et al., 2012) and adverse psychosocial factors at work (Sinokki et al., 2010; Lahelma et al., 2012; Mäntyniemi et al., 2012; Canivet et al., 2013). Pain is a leading symptom of most musculoskeletal disorders. "
    [Show abstract] [Hide abstract] ABSTRACT: Musculoskeletal pain at several sites (multisite pain) is more common than single-site pain. Little is known on its effects on disability pension (DP) retirement. A nationally representative sample comprised 4071 Finns in the workforce aged 30 to 63. Data (questionnaire, interview, clinical examination) were gathered in 2000-2001 and linked with national DP registers for 2000-2011. Pain during the preceding month in 18 locations was combined into four sites (neck, upper limbs, low back, lower limbs). Hazard ratios (HR) of DP were estimated by Cox regression. The HR of any DP (n = 477) was 1.6 (95% confidence interval 1.2-2.1) for one, 2.5 (1.9-3.3) for two, 3.1 (2.3-4.3) for three and 5.6 (4.0-7.8) for four pain sites, when adjusted for age and gender. When additionally adjusted for clinically assessed chronic diseases, the HRs varied from 1.4 (1.0-1.8) to 3.5 (2.5-4.9), respectively. When further adjusted for physical and psychosocial workload, education, body mass index, smoking, exercise and sleep disorders, the HRs were 1.3 (0.9-1.7), 1.6 (1.2-2.2), 1.8 (1.3-2.5) and 2.5 (1.8-3.6). The number of pain sites was especially strong in predicting DPs due to musculoskeletal diseases (HRs in the full model; 3.1 to 4.3), but it also predicted DPs due to other somatic diseases (respective HRs 1.3 to 2.3); pain in all four sites was also predictive of DPs due to mental disorders (full model HR 2.2). The number of pain sites independently predicted DP retirement. Employees with multisite pain may need specific support to maintain their work ability. © 2014 The Authors. European Journal of Pain published by John Wiley & Sons Ltd on behalf of European Pain Federation - EFIC®.
    Full-text · Article · Aug 2015
    • "Multi-site pain is associated with an increased risk of sickness absence (Haukka et al., 2013), disability pension (Kamaleri et al., 2009b) and poor self-rated work ability (Miranda et al., 2010). Sleep problems is a factor associated with both pain (Smith and Haythornthwaite, 2004; Fishbain et al., 2010; Alsaadi et al., 2011) and sickness absence (Sivertsen et al., 2006Sivertsen et al., , 2009 Akerstedt et al., 2007; Salo et al., 2010 ). In a review from 2010, investigating the relationship between chronic pain and sleep disorder, the presence of a bidirectional relationship between sleep and pain could not be excluded (Fishbain et al., 2010). "
    [Show abstract] [Hide abstract] ABSTRACT: Background Disturbed sleep and pain often co-exist and the relationship between the two conditions is complex and likely reciprocal. This 5-year prospective study examines whether disturbed sleep can predict the onset of multi-site pain, and whether non-disturbed sleep can predict the resolution of multi-site pain.Methods The cohort (n = 1599) was stratified by the number of self-reported pain sites: no pain, pain from 1-2 sites and multi-site pain (≥3 pain sites). Sleep was categorized by self-reported sleep disturbance: sleep A (best sleep), sleep B and sleep C (worst sleep). In the no-pain and pain-from-1-2 sites strata, the association between sleep (A, B and C) and multi-site pain 5 years later was analysed. Further, the prognostic value of sleep for the resolution of multi-site pain at follow-up was calculated for the stratum with multi-site pain at baseline. In the analyses, gender, age, body mass index, smoking, physical activity and work-related exposures were treated as potential confounders.ResultsFor individuals with no pain at baseline, a significantly higher odds ratio for multi-site pain 5 years later was seen for the tertile reporting worst sleep [odds ratio (OR) 4.55; 95% confidence interval (CI) 1.28-16.12]. Non-disturbed (or less disturbed) sleep had a significant effect when predicting the resolution of multi-site pain (to no pain) (OR 3.96; 95% CI 1.69-9.31).Conclusion In conclusion, sleep could be relevant for predicting both the onset and the resolution of multi-site pain. It seems to be a significant factor to include in research on multi-site pain and when conducting or evaluating intervention programmes for pain.
    Full-text · Article · Mar 2015
Show more