Occupational Medicine 2017;67:377–382

Advance Access publication 1 June 2017 doi:10.1093/occmed/kqx054

Impact of working hours on sleep and mental

health
P. Afonso1, M. Fonseca2 and J. F. Pires3
Department of Psychiatry, Faculty of Medicine, University of Lisbon, Lisbon, Portugal, 2Center of Mathematics and
1

Downloaded from https://academic.oup.com/occmed/article-abstract/67/5/377/3859790 by guest on 25 September 2019

Applications, Faculty of Sciences and Technology, Nova University of Lisbon, Lisbon, Portugal, 3Human Behaviour Department,
AESE-Business School, Lisbon, Portugal.
Correspondence to: P. Afonso, Department of Psychiatry, Faculty of Medicine, University of Lisbon, Av. Prof. Egas Moniz, 1640-
035 Lisbon, Portugal. Tel: +351 21 799 95 57; e-mail: pedromafonso@netcabo.pt

Background The number of hours people are required to work has a pervasive influence on both physical and
mental health. Excessive working hours can also negatively affect sleep quality. The impact at work of
mental health problems can have serious consequences for individuals’ as well as for organizations’
productivity.
Aims To evaluate differences in sleep quality and anxiety and depression symptoms between longer work-
ing hours group (LWHG) and regular working hours group (RWHG). To examine factors influenc-
ing weekly working hours, sleep quality and anxiety and depressive symptoms.
Methods Participants were divided into two groups, RWHG and LWHG, based on working hours, with a cut-
off of 48 h per week. We used the Hospital Anxiety and Depression Scale (HADS) to assess anxiety
and depression symptoms and the Pittsburgh Sleep Quality Index (PSQI) to measure the quality
and patterns of sleep.
Results The response rate was 23%. Among the 429 study participants, those in the LWHG group (n = 256,
53%) had significantly more depressive and anxiety symptoms and worse sleep quality than those in
RWHG (n = 223, 47%). Working time was significantly positively correlated with higher corporate
position and HADS scores. Moreover, HADS scores were positively correlated with PSQI scores
and negatively correlated with age.
Conclusions This study suggests that longer working hours are associated with poorer mental health status and
increasing levels of anxiety and depression symptoms. There was a positive correlation between these
symptoms and sleep disturbances.
Key words Anxiety; depression; long working hours; sleep quality.

Introduction made worldwide in establishing statutory limits on work-

Long working hours are commonplace worldwide and
have been one of the most important topics in occupa-
tional health since the late 19th century. The European
Working Time Directive (Directive 2003/88/EC) estab-
lished minimum general safety and health requirements
for the organization of certain aspects of working time
in the European Union’s 28 member states, including
for the health professions. The directive states in its pre­
amble that ‘the improvement of workers’ safety, hygiene
and health at work is an objective which should not
be subordinated to purely economic consideration’.
Furthermore, it stipulates that workers should have a
maximum weekly working time of 48 h, including over-
time. Over recent decades, significant progress has been
ing hours. As a result, the majority of countries now have
statutory limits <48 h, and around half have a 40-h limit
in place [1]. Despite this, ~22% of workers worldwide are
still working >48 h per week [2].
The number of hours people are required to work has
a pervasive influence not only on physical health but also
on mental health. Excessive weekly working time has
negative effects on workers’ health, including increasing
the risk of hypertension, cardiovascular disease, chronic
infection, diabetes, metabolic syndrome, sleep disturb­
ance, anxiety and depression [1,3–6]. Additionally,
excessive working time has been implicated in several
sudden deaths caused by cardiovascular diseases (e.g.
stroke, acute cardiac failure, myocardial infarction
and aortic aneurysm rupture) in middle-aged workers.

© The Author 2017. Published by Oxford University Press on behalf of the Society of Occupational Medicine.
All rights reserved. For Permissions, please email: journals.permissions@oup.com
378  OCCUPATIONAL MEDICINE
In Japan, such deaths are called karoshi, meaning ‘death
from overwork’ [7]. In Japan, the number of workers
suffering from cardiovascular disease, cerebrovascular
disease and mental disorders due to work has increased
3-fold in the last decade [3].
Mental health problems in the workplace have seri-
ous consequences not only for the individual, but also
for the productivity of the organization. In Organisation
for Economic Co-operation and Development countries,
mental ill-health is responsible for between one-third
and half of all long-term sickness and disability in the
working age population. Data show that many people
with common mental health problems struggle at work.
For example, 69% of people with moderate mental
health problems report having difficulty with job per-
formance, compared with 26% of those without men-
tal health problems [8]. Sleep disturbance is associated
with substantial impairment in an individual’s quality of
life. Compared with good sleepers, people with persist­
ent sleep disturbance are more prone to accidents, have
higher rates of work absenteeism, diminished job per-
formance, decreased quality of life and increased health
care use [9]. Several studies have revealed that working
overtime is related to short or disturbed sleep and can
correlate with reduced sleep quality in a dose-response
manner [10–12].
The aim of this study was to examine the effect of
longer working hours on sleep quality, anxiety and
depressive symptoms in white-collar workers.
Methods
We sent an e-mail invitation to participate in this survey,
with a unique hyperlink to a questionnaire, to all alumni
of the Portuguese AESE-Business School. Participation
in the study was voluntary and the questionnaire was
confidential and anonymous. The study received ethical
approval from the institutional review board. We defined
long working hours as >48 working hours per week.
According to this criterion, participants were divided into
a regular working hours group (RWHG) and a longer
working hours group (LWHG). We designed a question-
naire to collect demographic and work characteristics
data and used the Hospital Anxiety and Depression Scale
(HADS) to elicit anxiety and depressive symptoms [13].
HADS is a self-reported 14-item questionnaire com-
posed of two 7-item subscales, one measuring anxiety
and one depression, which are scored separately. Each
item is answered on a four-point (0–3) Likert scale, so
possible scores range from 0 to 21 for both anxiety and
depression, with higher scores indicating a higher level of
depression or anxiety. We assessed the quality and pat-
terns of sleep using the Pittsburgh Sleep Quality Index
(PSQI) [14]. This self-reported questionnaire rates sleep
quality and patterns during the previous month and
evaluates seven components of sleep: subjective quality,
latency, duration, usual efficiency, sleep disturbances,
medication use and daytime dysfunction. A score >5 is
suggestive of a sleep disorder.
Standard descriptive summary statistics were used
to characterize the full sample and the two groups.
Associations between variables were evaluated using
the χ2 test (categorical variables) and the independent
t-test with unequal variances and estimated degrees of
freedom (continuous variables). Regression analysis was
performed on average work hours per week, PQSI and
HADS scores, social and demographic variables, work-
Downloaded from https://academic.oup.com/occmed/article-abstract/67/5/377/3859790 by guest on 25 September 2019
ing habits and sleep patterns. The backward selection
method using the Akaike Information Criterion was
used for model selection. All analyses were performed
using commercially available and open source statistical
software (IBM SPSS 22.0, R 3.2.4). In all regressional
analysis, the symmetry of residuals and large sample size
guaranteed estimates’ quality. A P value <0.05 was con-
sidered significant.
Results
Surveys were sent to 2059 alumni and 479 completed sur-
veys (23%) were returned and validated. The responders
were divided into RWHG (n = 223, 47%) and LWHG
(256, 53%). The average weekly working hours were 38
(SD 11.8) in RWHG and 55 (SD 6.5) in LWHG. In addi-
tion, a larger proportion of the LWHG reported working
>5 days per week (32 versus 7%). Characteristics of the
study participants are shown in Table 1. In both groups,
participants were mostly married men with children. The
propensity to work long hours was associated with level
in the corporate hierarchy. Interestingly, LWHG mem-
bers had higher level corporate positions than RWHG
subjects. This means that those with higher level corpo-
rate positions were statistically significantly more likely
to work more weekly hours than participants with lower
positions. Moreover, LWHG subjects were more likely to
take work home (75%) than those in the RWHG (55%;
P < 0.001). We found no significant differences in shift
work between the two groups.
An independence test to rule out any confounding
between sex and job level, using the χ2 statistic, yielded
P >0.99. A similar test was performed for sex and hav-
ing children, with P = 0.83. Both of these factors can be
considered independent of sex.
Data on PSQI and HADS scores are reported in
Table  2. The RWHG had significantly more sleeping
time (mean 6.7  h, SD 0.8) than the LWHG (mean
6.4 h, SD 0.9; P < 0.001), so sleeping hours were signif-
icantly negatively associated with weekly working hours
(P < 0.05). As shown in Table 2, PSQI scores were not
significantly higher for those in LWHG. Nonetheless,
using a cut-off score of 5 on the PSQI scale, a signifi-
cantly higher number of subjects in LWHG (66%)
reported more sleep disorders than in the RWHG
 P. AFONSO ET AL.: IMPACT OF WORKING HOURS ON SLEEP AND MENTAL HEALTH  379

Table 1.  Characteristics of the study participants by groups with long and regular working hours

Regular working hours Long working hours P value

Age, mean (SD) 47.6 (8.3) 46.7 (8.4) NSa

Sex, n (%)
 Female 85 (38) 70 (27) <0.05b
 Male 138 (62) 186 (73)
Civil status, n (%)
 Married 151 (68) 193 (75) NSb
 Divorced 24 (11) 17 (7)
 Single 28 (13) 28 (11)

Downloaded from https://academic.oup.com/occmed/article-abstract/67/5/377/3859790 by guest on 25 September 2019

  Common law 19 (8) 16 (6)
 Widower 1 (0) 2 (1)
Children, n (%)
 No 47 (21) 39 (15) NSb
 Yes 176 (79) 217 (85)
Corporate position, n (%)
 President/CEO 33 (15) 52 (20) <0.001b
  Department head 86 (38) 143 (56)
  Section head 74 (33) 39 (15)
 Supervisor 22 (10) 14 (6)
 Staff 2 (1) 2 (1)
 Other 6 (3) 6 (2)
Education level, n (%)
  Doctoral degree 8 (4) 17 (7) NSb
  Bachelor or master degree 208 (93) 230 (90)
  High school 7 (3) 9 (3)
Annual days of vacation/holidays, mean (SD) 22.3 (4.1) 20.1 (4.8) <0.001a
Weekly working days, n (%)
 ≤4 1 (1) 1 (1) <0.001b
 5 206 (92) 172 (67)
 6 14 (6) 75 (29)
 7 2 (1) 8 (3)
Taking work home, n (%)
 No 100 (45) 63 (25) <0.001b
 Yes 123 (55) 193 (75)
Shift work or night work, n (%)
 No 216 (97) 244 (95) NSb
 Yes 7 (3) 12 (5)

NS, non-significant.
a
Student t-test.
χ test.
b 2

Table 2.  PSQI and HADS in groups with long and regular working hours#8232;

Scale Regular working hours Long working hours P valuea

PSQI, mean (SD) (total score) 6.40 (3.3) 7.00 (3.5) NS

Sleep disorder, n (%)
  Yes (PSQI ≥ 5) 116 (54) 163 (66) <0.01
  No (PSQI < 5) 98 (46) 83 (34)
HADS total score, mean (SD) 10.6 (6.3) 12.6 (6.2) <0.05
HADS-anxiety, mean (SD) 6.5 (3.6) 7.5 (3.6) <0.01
HADS-depression, mean (SD) 4.4 (3.3) 5.1 (3.4) <0.01

NS, non-significant.
a
Student t-test.

(54%). In addition, there were significantly higher total those in RWHG. Although HADS scores do not provide
HADS scores and HADS subscale scores for anxiety definitive diagnoses of anxiety and depressive disor-
and depression in LWHG members compared with ders, these results show that LWHG members reported
380  OCCUPATIONAL MEDICINE

significantly more anxiety and depression symptoms
than those in RWHG.
Results of the regression analysis are shown in Tables 3–5.
As shown in Table 3, a significant positive relationship
was found between working hours per week and work in
some professional areas, having a higher level of corpor­
ate responsibility, taking work home and the number of
working days per week. Additionally, a negative correla-
tion was found between working hours per week and the
number of vacation days annually.
The results from the regression analysis for PSQI
working hours compared with those reporting regular
hours. These findings are similar to those of several previ-
ous studies [7,15–17]. Moreover, there was a significant
negative correlation between total HADS score and age.
This relationship can probably be explained by burnout,
since age has been found to be related to burnout. For
instance, in younger employees, the level of burnout is
higher than in those over 30 or 40 years old [18], sug-
gesting that burnout appears early in a person’s career
and may be a risk factor for developing depression [19].
However, the apparent relationship between burnout

Downloaded from https://academic.oup.com/occmed/article-abstract/67/5/377/3859790 by guest on 25 September 2019

scores (Table 4) suggested a positive correlation between
PSQI scores and both HADS-depression and HADS-
anxiety subscale scores.
There were positive correlations between the total
HADS score and both PSQI score and taking work home.
In addition, there was a negative correlation between the
HADS total score and age.
Discussion
Our study found significantly higher anxiety and depres-
sion symptom scores in respondents reporting long
and age is inconsistent and further studies are needed to
clarify this [20].
There is general agreement that long hours of work
adversely affect sleep [21]. In our study, there were more
participants reporting sleep disturbance in the LWHG
than the RWHG, which is compatible with previous
observations [12,22]. Additionally, our results showed
a significant inverse relationship between working time
and sleep duration. Inadequate recovery because of sleep
deprivation is considered an important component of the
pathway from long work hours to increased fatigue and
risk of health problems [23,24]. Moreover, shift work is

Table 3.  Regression analysis of factors associated with working hours per week

Estimate (original Lower bound (95%) Upper Estimate P value (original

parametrization) bound (95%) parametrization)

(Constant) 35.09 25.30 44.92 39.92 <0.001

Age −0.098 −0.180 −0.016 −0.098 <0.05
Sex (male) 1.777 0.347 3.214 1.777 <0.05
Professional area
  Public service and 0.000 −3.631
ONGs
  Media and 1.541 −12.00 15.15 −2.090 NS
advertisement
  Logistics and 4.418 0.076 8.779 0.787 <0.05
commerce
  Education and culture 0.982 −4.750 6.740 −2.648 NS
 Finance 5.241 1.103 9.398 1.611 <0.05
 Management 3.600 −0.189 7.405 −0.031 NS
  Medicine and health 6.604 2.707 10.519 2.974 <0.05
industry
 Tourism 6.708 −0.799 14.248 3.077 NS
 Other 3.582 −0.419 7.600 −0.049 <0.05
Job level
 President/CEO 3.832 1.671 6.003 2.629 <0.05
  Department head 3.317 1.644 4.997 2.114 <0.001
  Section head 0.000 −1.203
 Supervisor −1.161 −3.826 1.515 −2.364 NS
 Staff 1.341 −5.497 8.208 0.138 NS
 Other −0.110 −4.486 4.286 −1.313 NS
Annual days of vacation/ −0.252 −0.407 −0.097 −0.252 <0.05
holidays
Weekly working days 4.095 2.828 5.366 4.095 <0.001
Taking work home 1.801 0.383 3.225 1.801 <0.05

The factors ‘Corporate position’ and ‘Education level’ were reparametrized so that their coefficients have a null mean. This allows the coefficients to be interpreted as
positive or negative deviations from a nominal level. Multiple R2 = 0.323. NS, non-significant.
 P. AFONSO ET AL.: IMPACT OF WORKING HOURS ON SLEEP AND MENTAL HEALTH  381

Table 4.  Regression analysis of factors associated with PSQI scores

Variable Estimate Lower bound (95%) Upper bound (95%) P value

(Constant) 4.254 1.649 6.860 <0.001

Age 0.027 −0.004 0.059 NS
Working hours/week 0.017 −0.004 0.039 NS
HADS-anxiety 0.225 −0.182 1.512 <0.001
HADS-depression 0.173 −0.533 −0.106 <0.001

Multiple R2 = 0.249. NS, non-significant.

Downloaded from https://academic.oup.com/occmed/article-abstract/67/5/377/3859790 by guest on 25 September 2019

Table 5.  Regression analysis of factors associated with the total HADS score

Variable Estimate Lower bound (95%) Upper bound (95%) P value

(Constant) 18.86 12.57 25.14 <0.001

Age −0.092 −0.147 −0.038 <0.001
Annual days of vacation/holidays −0.077 −0.184 0.029 NS
Weekly working days 0.691 −0.191 1.573 NS
Taking work home 1.304 0.329 2.279 <0.01
PSQI 0.613 0.098 3.871 <0.001

Multiple R2 = 0.407. NS, non-significant.

associated with sleep disturbance [1,21,25], which may
bias the results. However, there were only seven shift
workers in the RWHG (3%) and 12 in the LWHG (5%),
a non-significant difference unlikely to have biased the
results. Sleep problems may be associated with psychiat-
ric conditions. Thus, there is growing experimental evi-
dence that the relationship between psychiatric disorders
and sleep is complex and includes bi-directional causa-
tion [26]. The positive correlation between PSQI scores
and HADS-anxiety and HADS-depression scores seem
to support this relationship. In other words, our results
showed an association between poor sleep quality and
anxiety and depression symptoms.
A sex-specific difference was also found between
the two groups, with more men in the LWHG (73%)
than the RWHG (62%). Similar results were obtained
in a large survey regarding working conditions in the
European Union [27]. Women tend to work shorter
hours than men, presumably because they still retain pri-
mary responsibility for the majority of housework and
childcare. Not surprisingly, LWHG members had sig-
nificantly less vacation time, worked more days per week
and took more work home than RWHG members.
Our study has several limitations. Firstly, it was a cross-
sectional study, whereas longitudinal research would be
necessary to clarify the long-term effects of long working
hours on mental health and sleep quality. Secondly, long
working hours may in part affect mental health through
factors not measured in our study, such as work–fam-
ily conflicts [28], burnout [18], a prolonged increase in
cortisol levels [28,29] and alcohol abuse [30]. Finally,
the participants in the study were alumni from a busi-
ness school, so the results may not be applicable to other
workers. In particular, the participants were white-collar
workers and with a higher level of education than the
general population.
Almost a century has passed since the 1919 adop-
tion of the first international labour standard on work-
ing hours, which established the principle of the 8-h
working day and 48-h working week. Despite progress in
legislation, ~22% of workers worldwide still work >48 h
per week [2]. Surprisingly, 53% of workers in our study
reported working >48  h per week, suggesting that the
European Working Time Directive has been ineffective,
at least at higher corporate levels.
Work is important for economic self-sufficiency and
a meaningful way of living. However, as demonstrated
in this study, long working hours can also be associ-
ated with problems such as an increased risk of sleep
disturbance and symptoms of anxiety and depression.
Our findings suggest that we still need to advocate
shorter working hours in order to preserve health and
well-being.
Key points
•• In this study of Portuguese business school alumni
long weekly working hours were associated with
poorer mental health status, as evidenced by
higher levels of anxiety and depression symptoms.
•• Long weekly working hours were also associ-
ated with reduced sleep time and increased sleep
disturbance.
•• These results confirm the importance of main-
taining regular weekly working hours and avoid-
ing excessive overtime work in reducing the risk of
anxiety, depression and sleep disorders.
382  OCCUPATIONAL MEDICINE
Conflicts of interest
None declared.
References
1. Rodriguez-Jareño MC, Demou E, Vargas-Prada S et  al.
European Working Time Directive and doctors’ health:
a systematic review of the available epidemiological evi-
dence. BMJ Open 2014;4:e004916.
2. Lee S-H, McCann DM, Messenger JC. International
Labour Organization. Working Time Around the World :
Trends in Working Hours, Laws and Policies in a Global
Comparative Perspective. Routledge, 2007. http://www.ilo.
org/wcmsp5/groups/public/@dgreports/@dcomm/@publ/
documents/publication/wcms_104895.pdf (18 June 2016,
date last accessed).
3. Bannai A,Tamakoshi A.The association between long work-
ing hours and health: a systematic review of epidemiologi-
cal evidence. Scand J Work Environ Health 2014;40:5–18.
4. Dembe AE, Erickson JB, Delbos RG, Banks SM. The
impact of overtime and long work hours on occupational
injuries and illnesses: new evidence from the United States.
Occup Environ Med 2005;62:588–597.
5. Kivimäki M, Batty GD, Hamer M et al. Using additional
information on working hours to predict coronary heart
disease: a cohort study. Ann Intern Med 2011;154:457–463.
6. Virtanen M, Ferrie JE, Singh-Manoux A et  al. Long
working hours and symptoms of anxiety and depression:
a 5-year follow-up of the Whitehall II study. Psychol Med
2011;41:2485–2494.
7. Shields M. Long working hours and health. Perspect Labour
Income 2000;12:49–56.
8. OECD. Fit Mind, Fit Job. From Evidence to Practice in Mental
Health and Work. Paris, France: OECD Publishing, 2015.
9. Roth T. Insomnia: definition, prevalence, etiology, and con-
sequences. J Clin Sleep Med 2007;3:S7–10.
10. Basner M, Fomberstein KM, Razavi FM et  al. American
time use survey: sleep time and its relationship to waking
activities. Sleep 2007;30:1085–1095.
11. Dahlgren A, Kecklund G, Akerstedt T. Overtime work and
its effects on sleep, sleepiness, cortisol and blood pressure
in an experimental field study. Scand J Work Environ Health
2006;32:318–327.
12. Nakashima M, Morikawa Y, Sakurai M et  al. Association
between long working hours and sleep problems in white-
collar workers. J Sleep Res 2011;20:110–116.
13. Zigmond AS, Snaith RP. The Hospital Anxiety and
Depression Scale. Acta Psychiatr Scand 1983;67:361–370.
14. Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR,
Kupfer DJ. The Pittsburgh Sleep Quality Index: a new
instrument for psychiatric practice and research. Psychiatry
Res 1989;28:193–213.
15. Virtanen M, Stansfeld SA, Fuhrer R, Ferrie JE, Kivimäki
M. Overtime work as a predictor of major depressive epi-
sode: a 5-year follow-up of the Whitehall II study. PLoS
ONE 2012;7:e30719.
16. Chen KY, Yang CM, Lien CH et al. Burnout, job satisfac-
tion, and medical malpractice among physicians. Int J Med
Sci 2013;10:1471–1478.
17. Nagashima S, Suwazono Y, Okubo Y et al. Working hours
Downloaded from https://academic.oup.com/occmed/article-abstract/67/5/377/3859790 by guest on 25 September 2019
and mental and physical fatigue in Japanese workers. Occup
Med (Lond) 2007;57:449–452.
18. Maslach C, Schaufeli WB, Leiter MP. Job burnout. Annu
Rev Psychol 2001;52:397–422.
19. Govardhan LM, Pinelli V, Schnatz PF. Burnout, depression
and job satisfaction in obstetrics and gynecology residents.
Conn Med 2012;76:389–395.
20. Ahola K, Honkonen T, Virtanen M, Aromaa A, Lönnqvist
J. Burnout in relation to age in the adult working popula-
tion. J Occup Health 2008;50:362–365.
21. Harrington J. Health effects of shift work and extended
hours of work. Occup Environ Med 2001;58:68–72.
22. Virtanen M, Ferrie JE, Gimeno D et  al. Long working
hours and sleep disturbances: the Whitehall II prospective
cohort study. Sleep 2009;32:737–745.
23. Park J, Kim Y, Chung HK, Hisanaga N. Long work-
ing hours and subjective fatigue symptoms. Ind Health
2001;39:250–254.
24. Härmä M. Work hours in relation to work stress, recovery
and health. Scand J Work Environ Health 2006;32:502–514.
25. Wright KP Jr, Bogan RK, Wyatt JK. Shift work and the
assessment and management of shift work disorder (SWD).
Sleep Med Rev 2013;17:41–54.
26. Krystal AD. Psychiatric disorders and sleep. Neurol Clin
2012;30:1389–1413.
27. Paoli P, Merlin D. Third European Survey on Working
Conditions 2000 | Eurofound. 2012. http://www.eurofound.
europa.eu/sites/default/files/ef_files/pubdocs/2001/21/en/1/
ef0121en.pdf (29 July 2016, date last accessed).
28. Crouter AC, Bumpus MF, Head MR, McHale SM.
Implications of overwork and overload for the qual-
ity of men’s family relationships. J Marriage Fam
2001;63:404–416.
29. Lundberg U, Hellström B. Workload and morning salivary
cortisol in women. Work Stress 2002;16:356–363.
30. Nash LM, Daly MG, Kelly PJ et  al. Factors associated
with psychiatric morbidity and hazardous alcohol use in
Australian doctors. Med J Aust 2010;193:161–166.