The midpoint of sleep on working days: A measure for chronodisruption and its association to individuals’ well-being

Abstract

There is consistent evidence suggesting a relationship between individuals' sleep-wake rhythms and well-being. The indiscriminate demands from daily working routines, which do not respect this individual physiological rhythm, might be mediating this phenomenon. The aim of the present study was to evaluate the relationship between the characteristics of sleep routines during working days and psychological well-being. This was a cross-sectional study on 825 individuals from rural communities from southern Brazil. The study protocol included a questionnaire on demographic characteristics, working routines, health complaints, and habits; the Munich Chronotype Questionnaire for sleep-wake rhythm and; the WHO-Five well-being index. Since sex has been shown to affect sleep circadian rhythm and well-being, analysis was performed on men and women separately. In the proposed hierarchical regression models, different factors contributed to well-being according to sex. Among men, sleep-wake and work-related variables did not predict well-being scores. Among women, later midpoints of sleep on working days (B = -1.243, SE B = 0.315, β = -0.220), working more days per week (B = -1.507, SE B = 0.494, β = 0.150), having longer working journeys (B = -0.293, SE B = 0.105, β = -0.166), earlier working journey midpoints (B = 0.465, SE B = 0.222, β = 0.115), and being exposed to less sunlight (B = 0.140, SE B = 0.064, β = 0.103) predicted worse well-being. For the subgroup of women with days free from work, we have found a correlation between later midpoints of sleep during the week with worse well-being (Pearson's r = -0.159, p = 0.045) while the same relationship was not significantly observed with the midpoint of sleep on non-working days (Pearson's r = -0.153, p = 0.054). Considering WHO-Five as categorical, based on proposed clinical cut-offs, among women working 7-d/week, those with worst well-being (WHO-Five < 13) had the latest midpoint of sleep (F = 4.514, p = 0.012). Thus, the midpoint of sleep on working days represents the interaction between individuals' sleep-wake behavior and working routines. It plays an important role as a stress factor and may be a useful alternative variable related to chronodisruption.

Full text

Chronobiology International, 2015; 32(3): 341–348
!Informa Healthcare USA, Inc.
ISSN: 0742-0528 print / 1525-6073 online
DOI: 10.3109/07420528.2014.979941
ORIGINAL ARTICLE
The midpoint of sleep on working days: A measure for
chronodisruption and its association to individuals’ well-being
Camila Morelatto de Souza
1,2
and Maria Paz Loayza Hidalgo
1,3
1
Laborato
´rio de Cronobiologia, Hospital de Clı
´nicas de Porto Alegre, Porto Alegre, Brazil,
2
Post-graduation program in
Psychiatry, and
3
Department of Psychiatric and Legal Medicine, UFRGS, Medical School, Porto Alegre, Brazi
There is consistent evidence suggesting a relationship between individuals’ sleep–wake rhythms and well-being. The
indiscriminate demands from daily working routines, which do not respect this individual physiological rhythm, might
be mediating this phenomenon. The aim of the present study was to evaluate the relationship between the
characteristics of sleep routines during working days and psychological well-being. This was a cross-sectional study on
825 individuals from rural communities from southern Brazil. The study protocol included a questionnaire on
demographic characteristics, working routines, health complaints, and habits; the Munich Chronotype Questionnaire
for sleep–wake rhythm and; the WHO-Five well-being index. Since sex has been shown to affect sleep circadian
rhythm and well-being, analysis was performed on men and women separately. In the proposed hierarchical
regression models, different factors contributed to well-being according to sex. Among men, sleep–wake and work-
related variables did not predict well-being scores. Among women, later midpoints of sleep on working days
(B¼1.243, SE B¼0.315, ¼0.220), working more days per week (B¼1.507, SE B¼0.494, ¼0.150), having
longer working journeys (B¼0.293, SE B¼0.105, ¼0.166), earlier working journey midpoints (B¼0.465,
SE B¼0.222, ¼0.115), and being exposed to less sunlight (B¼0.140, SE B¼0.064, ¼0.103) predicted worse well-
being. For the subgroup of women with days free from work, we have found a correlation between later midpoints of
sleep during the week with worse well-being (Pearson’s r¼0.159, p¼0.045) while the same relationship was not
significantly observed with the midpoint of sleep on non-working days (Pearson’s r¼0.153, p¼0.054). Considering
WHO-Five as categorical, based on proposed clinical cut-offs, among women working 7-d/week, those with worst
well-being (WHO-Five513) had the latest midpoint of sleep (F¼4.514, p¼0.012). Thus, the midpoint of sleep on
working days represents the interaction between individuals’ sleep–wake behavior and working routines. It plays an
important role as a stress factor and may be a useful alternative variable related to chronodisruption.
Keywords: Chronotype, circadian typology, entrainment, midpoint of sleep on workdays, sunlight, well-being, working
schedule
INTRODUCTION
Circadian sleep–wake rhythms vary greatly among
people. Early types are those individuals who are more
prone to sleep and wake up earlier, while late ones
are those who sleep and wake later, in relation to the
environmental night. This expressed behavior – termed
chronotype – is correlated with other physiological
functions, such as temperature and hormonal rhythms
(Adan et al., 2012; Roenneberg et al., 2003) and cellular
molecular mechanisms (Schulz & Steiner, 2009).
Later sleep–wake schedules have been associated
with poorer health-related outcomes, such as cardio-
vascular (Merikanto et al., 2013), metabolic diseases
(Antunes et al., 2010; Reutrakul et al., 2013), and obesity
(Roenneberg et al., 2012). In addition, reduced
well-being (Wittmann et al., 2010) and more depression
symptoms (Hidalgo et al., 2009; Levandovski et al., 2011;
Salgado-Delgado et al., 2011; de Souza & Hidalgo, 2014)
have been associated with these late types.
It has been hypothesized that the relationship con-
sistently found between chronotype and health com-
plaints, including depression, could be mediated by the
discrepancy between an individual’s inner rhythm and
societal demands, the necessary daily effort to adapt and
the chance to disrupt circadian physiological rhythmi-
city (Erren & Reiter, 2013; Foster et al., 2013; Klerman,
2005; Levandovski et al., 2011; Roenneberg et al., 2012;
Wittman et al., 2006; Wulff et al., 2010). Despite this, it
remains scarce the information on the relationship
between diurnal working routines, such as the starting
Correspondence: Prof. Maria Paz Loayza Hidalgo, Department of Psychiatric and Legal Medicine, UFRGS, Medical School, Ramiro
Barcelos St., 2400 2nd Floor, Porto Alegre 90035-00, Brazil. E-mail: mpaz@cpovo.net
Submitted May 17, 2014, Returned for revision October 20, 2014, Accepted October 20, 2014
341
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

and finishing times, the midpoint and the duration of
the working journey, with sleep rhythms and health
outcomes.
We hypothesized that the relationship between late
types and negative psychological outcomes, such as
well-being, could be mediated by individual’s diurnal
work routine. Thus, the aim of the present study was to
evaluate the relationship between the characteristics of
sleep routines during working days and psychological
well-being.
MATERIALS AND METHODS
Participants
The data presented here are part of a larger epidemio-
logical study that took place in Vale do Taquari,
Southern Brazil (for detailed information, see Carvalho
et al., 2014). Around 40% of the population in the
communities studied live in rural areas, and household
agriculture still represents an important share of eco-
nomic productivity. The population is predominantly
composed of Caucasians, particularly German descend-
ants, followed by Italian and a smaller group of Azorean
descendants. Participants were randomly enrolled fol-
lowing the draw of location and streets to be investi-
gated. Figure 1 shows study’s algorithm.
Study design and measurements
This was a cross-sectional study, in which subjects were
assessed in their homes by trained interviewers. In the
first evaluation, a questionnaire on demographics,
working characteristics, health habits, and the Munich
Chronotype Questionnaire (MCTQ) were applied. In
the second evaluation, participants answered the
World Health Organization five-item well-being index
(WHO-Five). The average interval between the two
assessments was 12.87 months.
Considering available evidence on factors either
related to chronotype or well-being, a questionnaire
on demographic characteristics included items related
to age, sex, level of formal education, working routine,
and health-related questions. The level of education was
reported as completed years of schooling, which was
further categorized into elementary school (8 years),
high school (9–11 years), and college, university, or
postgraduate study (12 years). We collected informa-
tion on the participants’ main working activities, and
they were categorized into the following groups: agri-
culture, industry, business or commerce, non-specia-
lized services, and specialized services. The working
schedule was investigated: starting and finishing
working times, working journey duration (finishing
time–starting time), and midpoint (starting time +
(duration/2)) of the journey.
The participants were asked to report any health
complaints, in addition to smoking habits and alcohol
consumption. All health complaints were reviewed
and were considered for the analysis those defined as
chronic/recurrent diseases.
The MCTQ (Brazilian–Portuguese version from
http://www.bioinfo.mpg.de/wepcronotipo) is a self-
reported instrument which was used to assess the
number of days in a week with regular working
routines, the midpoint of sleep (starting sleep time +
(sleep duration/2)), sleeping and wakening latencies
(time to fall asleep and to get out of bed, respectively),
sleep duration (finishing sleep time–starting sleep
time), and sunlight exposure for working and non-
working days (Di Milia et al., 2013; Levandovski et al.,
2013; Roenneberg et al., 2003). It also asks if partici-
pants are dependent on alarm clocks and if are there
other reasons besides work requiring them to wake at
the reported schedules on non-working days (Di Milia
et al., 2013; Levandovski et al., 2013; Roenneberg et al.,
2003). Because most (70.4%) of our sample works
during the 7 d of the week, not allowing the appropri-
ate calculation of parameters for non-working days,
the midpoint of sleep on working days was taken as
an alternative marker for an individual’s circadian
sleeping–waking rhythm. Unless otherwise specified,
all the sleeping–waking parameters described refer
to working days.
The midpoints on working and non-working days
were moderately correlated for both women (Pearson’s
r¼0.554, p0.001) and men (Pearson’s r¼0.468,
1st Phase of Assessment
6,506 participants
12 - 65 years old
2nd Phase of Assessment
1,127 participants
18 - 65 years old
Inclusion: more frequently those with more extreme than
intermediate chronotypes based on the midpoint of sleep on free
days corrected for the sleep debt accumulated on work days and
normalized by age and gender (Levandovski et al., 2011)
Exclusion: night shift workers (starting working activities earlier
than 5:00 a.m. and finishing later than 11:00 p.m.)
Selection studied
n=825
Exclusion: working less than 5 days/week (n=99); retired,
unemployed, or students (n=88); any incongruent or missing data
(n=24); any outlier data (n=91; defined as values ≥ or ≤3 standard
deviations from the mean for all studied variables).
Male
n=277
Female
n=548
5/6 working days
n=85
7 working days
n=192
5/6 working days
n=159
7 working days
n=389
FIGURE 1. Study algorithm.
342 C. M. de Souza & M. P. L. Hidalgo
Chronobiology International
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

p0.001). Since the midpoint of sleep on non-working
days was not correlated to the sleep duration on working
days (Spearman’s r¼0.097, p¼0.132) – not even to a
significant increase in the sleep duration on free days
(Spearman’s r¼0.073, p¼0.253) the advised correction
due to sleep debt accumulated during the workweek,
described in previous studies (Roenneberg et al., 2007),
was not used in the present.
The WHO-Five well-being index was used to evalu-
ate participants’ psychological well-being, in the pre-
vious 2 weeks. Questions were answered on a six-point
Likert scale from 0 (at no time) to 5 (all of the time),
generating a raw score ranging from 0 to 25. A cut-off
of 513 (or 550%) has been originally proposed
indicating poor well-being and the need for further
clinical investigation for depression (Bech, 2004). In a
Brazilian Portuguese validation study, the instrument
presented an alpha reliability of 0.83 and a new cut-off
of 519 with higher sensitivity has been proposed
(de Souza & Hidalgo, 2012). For most of the analysis
performed, we have used the final score as a continu-
ous variable. Categories for comparison were
generated according to the above-mentioned cut-offs
(13, 14–18, 19).
Statistical analysis
All data were included in the IBM Statistical Package for
the Social Sciences (SPSS, Chicao, IL) 22. Sample
demographics were expressed as means ± standard devi-
ation (SD) or the number of cases (–) and percentages.
Normality was defined as values of skewness and
kurtosis in between ±1, and this was taken into consid-
eration when choosing the most appropriate test for
each variable distribution. For comparison of two
continuous variables, Pearson’s/Spearman’s test was
used; for comparison of two categories, Student’s
t/Mann–Whitney U tests; and for categorical variables
with more than three groups, analysis of variance
(ANOVA)/Kruskal–Wallis test. To assess possible con-
founding effects and colinearity of variables, a hierarch-
ical multivariate linear regression analysis, which
included studied variables that we hypothesized might
be playing a role for the outcome well-being, was
performed for each sex. As factors, in the first step,
demographic characteristics and sleeping–waking circa-
dian behaviors were included, and in the second, the
working routines and daily sunlight exposure were
included. All the analyses were performed separately
for each sex. For all the analyses, two-sided tests for
significance at 5% level were performed.
Ethical aspects
The research ethics committee of Hospital de Clı
´nicas
de Porto Alegre approved the study protocol
(Project 08-087 GPPG/HCPA, CONEP 15155), and
written informed consent was obtained from all
participants.
RESULTS
Table 1 shows the distribution of the studied variables
for each sex separately. Most (57.5%) of the participants
in our sample did not use an alarm clock during the
workweek, and those who did use one often (46%) rose
before the alarm rang. Those who have days free from
work did not use alarm clocks, and most (76.6%) did
not report a reason to wake up at a specific time on
such days.
Relationship between sleep–wake rhythms and
working routines with well-being scores for each
sex separately
Figure 2 shows the distribution of WHO-Five scores for
sleeping–waking circadian parameters separately for
each sex. Among men, only the waking latency corre-
lated with well-being (Spearman’s r¼0.137, p¼0.022).
For women, a later midpoint of sleep (Pearson’s
r¼0.164, p0.001) and shorter sleep duration
(Pearson’s r¼0.114, p¼0.008) correlated with worse
well-being.
Figure 3 shows the distribution of WHO-Five scores
for the studied work-related variables separately for
each sex. Among males, none of the variables for
working routine were related to psychological well-
being. Among women, those who worked the entire
week showed significantly worse well-being than
those who had the chance to rest (WHO-Five scores
of 17.63, SD ¼4.803 versus 18.47, SD ¼3.828; t¼2.134,
p¼0.034).
Table 2 is a summary of the results of multiple
hierarchical regression analysis on the relationship
between sleep–wake rhythms during the workweek
and working routines with well-being scores for men
and women.
Relationship between sleep–wake rhythms and
working routines with well-being scores separately
for sex and for 7-d and 5- to 6-d workers
Table 3 shows values for sleep and work routines for
categories of sex and of the numbers of working days
per week. Among the men who worked continuously,
later wakening (Spearman’s r¼0.145, p¼0.045), and
taking longer to get out of bed (Spearman’s r¼0.188,
p¼0.009) were correlated to worse well-being. For
the group with days free from working obligations, the
least they have slept (Pearson’s r¼0.224, p¼0.040), the
worse their well-being.
Among the women who worked 7-d weeks, later
midpoints of sleep (Pearson’s r¼0.186, p0.001),
later wakening (Spearman’s r¼0.107, p¼0.035), and
shorter sleep duration (Pearson’s r¼0.124, p¼0.014)
were related to worse outcome. Among the women who
worked 5–6 d, later midpoints of sleep (Pearson’s
r¼0.159, p¼0.045) and lower sunlight exposure
(Spearman’s r¼0.273, p0.001) were related to
Midpoint of sleep on working days and well-being 343
!Informa Healthcare USA, Inc.
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

worse well-being. The midpoint of sleep on non-
working days was not correlated with well-being scores
for either sex (Pearson’s r¼0.018, p¼0.869; r¼0.153,
p¼0.054).
Comparison among different groups of well-being
accordingly to clinical cutoffs
Figure 4 shows the relationship between sleep–waking
and working routines for the clinically defined cutoffs
for the well-being categories, separately for each sex and
for the number of working days per week. Among men,
none of the studied variables could distinguish different
levels of well-being groups. For the women who
worked 7 d, the group with the worst well-being slept
later (F¼6.219, p¼0.002) had a later midpoint of sleep
(F¼4.514, p¼0.012), and shorter sleep duration
(F¼3.858, p¼0.022) compared with the group with
the best well-being. The group with the best well-being
had more sunlight exposure than the others (Kruskal–
Wallis Chi-square ¼17.121, p0.001).
DISCUSSION
Different factors are contributing to men and women
well-being. Among women we could observe an inter-
action between the sleep–wake rhythm and diurnal
working routines – represented by the midpoint of sleep
on working days – associating with their level of
psychological well-being. The later the midpoint of
their working and sleeping routines during the work-
week, even when including in the analysis other aspects
of the working routine, the worse their well-being.
Among men, none of the studied variables correlated to
their well-being.
Men and women behave differently and have differ-
ent physiological needs (Adan Sa
´nchez-Turet, 2001;
Adan & Natale, 2002; Lehnkering & Siegmund, 2007;
Roenneberg et al., 2007; Waterhouse et al., 2012). As has
been documented, the frequency and the intensity of
psychological complaints are higher among women
(Piccinelli & Wilkinson, 2000; Tesch-Ro
¨mer et al.,
2008). In our sample, women were younger, on an
TABLE 1. Characteristics of the study population, separately for each sex.
Male Female
Demography and health
Age* 44.65 (± 13.13) 42.28 (±11.38)
Education
Fundamental 74.01% 71.53%
High school 21.30% 20.98%
Superior 4.69% 7.48%
Self-reported disease
Yes 24.55% 29.20%
No 75.45% 70.80%
Smoking
Yes 16.24% 6.57%
No 83.75% 93.43%
Alcohol consumption
Yes 77.26% 40.15%
No 22.74% 59.85%
Working routines
Main activity***
Agriculture 67.15% 46.17%
Industry 8.30% 5.47%
Bussinesses 5.05% 4.38%
Non-specialized services
a
18.41% 40.51%
Specialized services
b
1.80% 3.47%
Working days/week
5/6 d 30.69% 29.01%
7 d 69.31% 70.99%
Working journey duration*** 11 h 15 min (±1 h 48 min) 10 h 43 min (±2 h 35 min)
Midpoint of working journey 12:35 (±58 min) 12:32 (±1 h 07 min)
Daily sunlight exposure* 8 h 57 min (±3 h 08 min) 8 h 20 min (±3 h 21 min)
Sleep–wake characteristics on workweek
Sleeping latency** 9.21 min (±15.07 min) 13.26 min (±18.81 min)
Midpoint of sleep*** 2:00 (±50 min) 2:15 (±48 min)
Awakening latency 7.09 min (±11.77 min) 6.78 min (±12.74 min)
Sleep duration 7h 55 min (±1 h 10 min) 7 h 51 min (±1 h 02 min)
WHO-Five
WHO-Five total score 19.36 (±4.04) 17.88 (±4.55)
*p0.05, **p0.01, ***p0.001
a
Without specific training.
b
With specific training such as teachers, physicians, lawyers.
344 C. M. de Souza & M. P. L. Hidalgo
Chronobiology International
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

average from a later chronotype and with worse well-
being scores than men (Table 1). Men were more
frequently working longer hours each day, as farmers
and for longer hours exposed to sunlight (Table 1).
Results should then be considered in the light of these
observations.
The participants examined in this study formed a
very homogenous sample with regard to a number of
working aspects: most worked in agriculture (53.2%) or
provided non-specialized services (33.1%); most worked
7 d a week (70.4%), and there was no great diversity in
starting and finishing hours of work (standard deviation
between 1 and 2 h). Also, the fact that many of them do
not use alarm clock (57.5%) or rose even before it ranges
(46%) suggests a regulatory effect on sleep rhythm from
working obligations. All these features yielded valuable
data to investigate the hypothesis that indiscriminate
working routines might be contributing factors in the
relationship between circadian sleep–wake rhythm and
well-being.
Among men, the factors examined in the present
study did not play a major role in the subjects’ well-
being. One possible explanation is that men, signifi-
cantly more frequently than women, worked outdoors
15
20
25
4 5 6 7 8 9 10 11 12
WHO-Five score
Sleep duration on work days (number of hours)
15
20
25
(A) (B)
(C) (D)
12:30 AM 1:30 AM 2:30 AM 3:30 AM 4:30 AM
WHO-Five score
Midpoint of sleep on work days (local time)
0
15
20
25
−150 153045607590105
WHO-Five score
Sleeping latency on work days (min)
15
20
25
−150 153045607590105
WHO-Five score
Awakening latency on work days (min)
Worse Better
Worse Better
Worse Better
Worse Better
0
00
FIGURE 2. Distribution of average WHO-Five scores for intervals of sleep–wake circadian rhythmic variables, separately for each sex. Male
sex is represented by dark gray rhombus and female by light gray dots. Vertical bars represent standard errors of the mean.
10
15
20
25
(A) (B)
WHO-Five score
5/6 working days/week 7 working days/week
*
15
20
25
2 4 6 8 1012141618
WHO-Five score
Working journey duration (number of hours)
Worse Better
Better
Worse
0
FIGURE 3. Distribution of average WHO-Five scores for groups working 5/6 d/week and 7 d/week and for hourly intervals of the working
journey, separately for each sex. Male sex is represented by dark gray columns/rhombus and female is represented by light gray columns/
dots. Vertical bars represent standard error of the mean. *p0.05.
Midpoint of sleep on working days and well-being 345
!Informa Healthcare USA, Inc.
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

as farmers with the opportunity to be exposed to the
natural light–dark cycle. Sunlight has been considered
the main zeitgeber and it is possibly facilitating, in this
case, individuals’ entrainment to their working obliga-
tions (Duffy & Wright, 2005; Golombek & Rosenstein,
2010; Roenneberg & Merrow, 2007; Martinez-Nicolas
et al., 2011).
Among women, we could replicate the previous
observation that later sleep routines are linked to
detrimental effects on health, here the worsening of
psychological well-being. In the univariate analysis,
less hours of sleep on the workweek and working 7 but
not 5/6 d per week were associated with a decrease in
well-being levels. The hierarchical regression analysis
allowed us to control for the interaction among this
significantly related variables to the outcome. It resulted
that the midpoint of sleep on working days remained
the biggest contributor in the proposed model. The
sleep duration on working days was no longer a
significant factor. It is possible that the relationship
observed through the univariate analysis was a conse-
quence of the co-linearity between these two variables.
All the working aspects related to scheduling
contributed to women’s well-being and, when included
in the model, even strengthened the relationship
between the midpoint of sleep and well-being, meaning
that they are mediating this association.
In a secondary analysis, considering apart workers
with and without days free from working demands,
we have found more varying associations, but,
among women, either working 5/6 or 7 d in a week,
the midpoint of sleep on working days was related to
well-being. Still, when considering clinical categories of
well-being, we have once more found that the category
of women with worst well-being presented the latest
midpoint of sleep on working days.
The WHO-Five well-being score has been proven
to be a useful instrument for depression screening
(Henkel et al., 2003; de Souza & Hidalgo, 2012), thus we
should be cautious when interpreting these findings,
because it is possible that the sleep behaviors we
associated with well-being are merely symptoms of
depression. Conversely, sleep complaints are commonly
reported in depression (Abe et al., 2011; Reid et al., 2012;
Selvi et al., 2007), and, in the present study, we have not
found that sleep rhythm on non-working days was
TABLE 2. Summary of the regression analysis for the outcome psychological well-being, sepretely for men and women.
Male (n¼277) Female (n¼548)
Model 1 Model 2 Model 1 Model 2
Variable BSE Bp BSE Bp BSE Bp BSE Bp
Age 0.003 0.020 0.011 0.001 0.021 0.004 0.021 0.018 0.052 0.031 0.019 0.076
Sleeping latency 0.005 0.017 0.018 0.005 0.017 0.018 0.000 0.11 0.004 0.001 0.011 0.005
Midpoint of sleep 0.549 0.360 0.113 0.668 0.419 0.138 0.721 0.264 0.128** 1.243 0.315 0.220***
Awakening latency 0.059 0.021 0.172** 0.063 0.021 0.182** 0.001 0.15 0.004 0.002 0.015 0.006
Sleep duration 0.159 0.242 0.046 0.118 0.258 0.034 0.348 0.197 0.079 0.241 0.205 0.055
Main activity – – – 0.023 0.195 0.009 – – – 0.085 0.120 0.035
Working days/week – – – 0.452 0.679 0.052 – – – 1.507 0.494 0.150**
Working journey duration – – – 0.034 0.175 0.015 – – – 0.293 0.105 0.166**
Midpoint of working journey – – – 0.253 0.325 0.060 – – – 0.465 0.222 0.115*
Daily sunlight exposure – – – 0.062 0.088 0.048 – – – 0.140 0.064 0.103*
R
2
0.042 0.048 0.035 0.075
Ffor change in R
2
2.391* 1.350 3.909** 4.363***
*p0.05, **p0.01, ***p0.001.
TABLE 3. Description of studied variables for the categories of sex and number of working days per week.
Male sex (n¼277) Female sex (n¼548)
5–6 d/week (n¼85) 7 d/week (n¼192) 5–6 d/week (n¼159) 7 d/week (n¼389)
Working routines
Work starting time 7 h 13 min ± 1 h 01 min 6 h 50 min ± 56 min*** 7 h 31 min ± 1 h 20 min 7 h 01 min ± 1 h 05 min***
Work finishing time 17 h 20 min ± 2 h 01 min 18 h 35 min ± 1 h 10 min*** 17 h 05 min ± 1 h 58 min 18 h 13 min ± 2 h 05 min***
Working journey duration 10 h 07 min ± 1 h 49 min 11 h 45 min ± 1 h 34 min*** 9 h 34 min ± 2 h 17 min 11 h 10 min ± 2 h 34 min***
Midpoint of working journey 12:17 ± 1 h 20 min 12:43 ± 43 min** 12:17 ± 1 h 14 min 12:37 ± 1 h 04 min***
Daily sunlight exposure 7 h 16 min ± 4 h 06 min 9 h 41 min ± 2 h 13 min*** 6 h 12 min ± 4 h 06 min 9 h 13 min ± 2 h 31 min***
Sleep–wake characteristics on workweek
Sleep starting time 22h 28 min ± 1 h 16 min 21 h 52 min ± 1 h 08 min*** 22 h 38 min ± 52 min 22h 12 min ± 1 h 07 min***
Sleep finishing time 6 h 13 min ± 53 min 5 h 51 min ± 42 min*** 6 h 17 min ± 46 min 6 h 08 min ± 52 min*
Sleep duration 7 h 44min ± 1 h 17 min 7 h 59min ± 1 h 06 min 7 h 40 min ± 58 min 7 h 56 min ± 1 h 03 min**
Midpoint of sleep 2:20 ± 53 min 1:51 ± 46 min*** 2:28 ± 40 min 2:10 ± 51 min***
*p0.05, **p0.01, ***p0.001.
346 C. M. de Souza & M. P. L. Hidalgo
Chronobiology International
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

correlated with well-being. Owing to the cross-sectional
design of this study, we are unable to derive any cause-
effect relationship.
Our study was limited by the fact that we have
studied a very homogeneous sample, thus we could only
generalize these findings to similar populations.
Moreover, we did not examine the flexibility of working
schedules, the possibility of taking naps during the
working day, earnings, and additional informal work
information.
However, to the best of our knowledge, no previous
study has evaluated different diurnal working schedules
and the rhythm of sleep during a working week.
A real-life study like the present one offers the chance
FIGURE 4. Relationship between sleep and working schedules for categories of well-being. Here is represented a 24 h day, dark gray bars
represent sleeping periods during the work week; dashed bars represent sleeping periods during non-working days; and white bars represent
working periods. Light shadowed area represents sunlight period for the longest day of the year for these coordinates and dark shadowed
represent the shortest. Horizontal lines are standard errors of the mean (SEM). *p0.05. See text for description of analysis results.
Midpoint of sleep on working days and well-being 347
!Informa Healthcare USA, Inc.
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.

to examine the full complexity of human behavior and
all the attendant factors. This study specifically under-
lines the importance of considering gender and social
routines when assessing sleep rhythms. The relevance of
these discoveries also rely on the fact that we are
identifying new ways of evaluating these variables,
which are potentially adjustable, in contrast to sex or
even the midpoint of sleep.
DECLARATION OF INTEREST
The authors declare that they have no conflicts of
interest. The Programa Brasil-Alemanha (PROBRAL)
between Coordenac¸a
˜o de Aperfeic¸oamento de Pessoal
de Nı
´vel Superior (CAPES) – Brazil and Deutscher
Akademischer Austausch Dienst (DAAD) – Germany
supported this work. M. P. L. H. received financial
support from Conselho Nacional de Desenvolvimento
Cientı
´fico e Tecnolo
´gico – CNPq.
REFERENCES
Abe T, Inoue Y, Komada Y, et al. (2011). Relation between
morningness–eveningness score and depressive symptoms
among patients with delayed sleep phase syndrome. Sleep
Med. 12:680–4.
Adan A, Archer SN, Hidalgo MP, et al. (2012). Circadian typology:
A comprehensive review. Chronobiol Int. 29:1–23.
Adan A, Natale V. (2002). Gender differences in morningness-
eveningness preference. Chronobiol Int. 19:709–20.
Adan A, Sa
´nchez-Turet M. (2001). Gender differences in diurnal
variations of subjective activation and mood. Chronobiol Int.
18:491–502.
Antunes LC, Jornada MN, Ramalho L, Hidalgo MPL. (2010).
Correlation of shift work and waist circumference, body mass
index, chronotype and depressive symptoms. Arq Bras
Endocrinol Metabol. 54:7.
Bech P. (2004). Measuring the dimension of psychological general
well-being by the WHO-Five. Qual Life Newslett. 32:15–6.
Carvalho FG, Hidalgo MP, Levandovski R. (2014). Differences in
circadian patterns between rural and urban populations:
An epidemiological study in countryside. Chronobiol Int. 313:
442–9.
Di Milia L, Adan A, Natale V, Randler C. (2013). Reviewing the
psychometric properties of contemporary circadian typology
measures. Chronobiol Int. 30:1261–71.
Duffy JF, Wright Jr KP. (2005). Entrainment of the human circadian
system by light. J Biol Rhythms. 20:326–38.
Erren TC, Reiter RJ. (2013). Revisiting chronodisruption: When the
physiological nexus between internal and external times splits
in humans. Naturwissenschaften. 100:291–8.
Foster RG, Peirson SN, Wulff K, et al. (2013). Sleep and circadian
rhythm disruption in social jetlag and mental illness. Prog Mol
Biol Transl Sci. 119:325–46.
Golombek DA, Rosenstein RE. (2010). Physiology of circadian
entrainment. Physiol Rev. 90:1063–102.
Henkel V, Mergl R, Kohnen R, et al. (2003). Identifying depression
in primary care: A comparison of different methods in a
prospective cohort study. Br Med J. 326:200–1.
Hidalgo MP, Caumo W, Posser M, et al. (2009). Relationship
between depressive mood and chronotype in healthy subjects.
Psychiatry Clin Neurosci. 63:283–90.
Klerman EB. (2005). Clinical aspects of human circadian rhythms.
J Biol Rhythms. 20:375–86.
Levandovski RM, Dantas G, Fernandes LC, et al. (2011). Depression
scores associate with chronotype and social jetlag in a rural
population. Chronobiol Int. 28:771–8.
Levandovski R, Sasso E, Hidalgo MP. (2013). Chronotype: A review
of the advances, limits and applicability of the main instru-
ments used in the literature to assess human phenotype.
Trends Psychiatry Psychother. 35:3–11.
Lehnkering H, Siegmund R. (2007). Chronotype, season, and sex of
subject on sleep behavior of young adults. Chronobiol Int.
24:875–88.
Martinez-Nicolas A, Ortiz-Tudela E, Madrid JA, Rol MA. (2011).
Crosstalk between environmental light and internal time in
humans. Chronobiol Int. 28:617–29.
Merikanto I, Lahti T, Puolijoki H, et al. (2013). Associations of
chronotype and sleep with cardiovascular diseases and type 2
diabetes. Chronobiol Int. 30:470–7.
Piccinelli M, Wilkinson G. (2000). Gender differences in depression:
Critical review. Br J Psychiatry. 177:486–92.
Reid KJ, Jaksa AA, Eisengart JB, et al. (2012). Systematic evaluation
of axis-I DSM diagnoses in delayed sleep phase disorder and
evening-type circadian preference. Sleep Med. 13:1171–7.
Reutrakul S, Hood MM, Crowley SJ, et al. (2013). Chronotype
is independently associated with glycemic control in type 2
diabetes. Diabetes Care. 36:2523–9.
Roenneberg T, Allebrandt KV, Merrow M, Vetter C. (2012).
Social jetlag and obesity. Curr Biol. 22:939–43.
Roenneberg T, Kuehnlea T, Judaa M, et al. (2007). Epidemiology of
the human circadian clock. Sleep Med Rev. 11:429–38.
Roenneberg T, Merrow M. (2007). Entrainment of the human
circadian clock. Cold Spring Harb Symp Quant Biol. 72:293–9.
Roenneberg T, Wirz-Justice A, Merrow M. (2003). Life between
clocks: Daily temporal patterns of human chronotypes. J Biol
Rhythms. 18:80–90.
Salgado-Delgado R, Osorio AT, Saderi N, Escobar C. (2011).
Disruption of circadian rhythms: A crucial factor in the etiology
of depression. Depress Res Treat. 2011:839743.
Schulz P, Steimer T. (2009). Neurobiology of circadian systems.
CNS Drugs. 23:3–13.
Selvi Y, Gulec M, Agargun YM, Besiroglu L. (2007). Mood changes
after sleep deprivation in morningness–eveningness chrono-
types in healthy individuals. J Sleep Res. 16:241–4.
de Souza CM, Hidalgo MP. (2012). World Health Organization
5-item well-being index: Validation of the Brazilian Portuguese
version. Eur Arch Psychiatry Clin Neurosci. 262:239–44.
de Souza CM, Hidalgo MP. (2014). Midpoint of sleep on school
days is associated with depression among adolescents.
Chronobiol Int. 31:199–205.
Tesch-Ro
¨mer C, Motel-Klingebiel A, Tomasik MJ. (2008). Gender
differences in subjective well-being: Comparing societies with
respect to gender equality. Soc Indic Res. 85:329–49.
Waterhouse J, Fukuda Y, Morita T. (2012). Daily rhythms of the
sleep–wake cycle. J Physiol Anthropol. 31:5.
Wittmann M, Paulus M, Roenneberg T. (2010). Decreased psycho-
logical well-being in late ‘chronotypes’ is mediated by smoking
and alcohol consumption. Subst Use Misuse. 45:15–30.
Wulff K, Gatti S, Wettstein JG, Foster RG. (2010). Sleep and
circadian rhythm disruption in psychiatric and neurodegen-
erative disease. Nat Rev Neurosci. 11:589–9.
348 C. M. de Souza & M. P. L. Hidalgo
Chronobiology International
Chronobiol Int Downloaded from informahealthcare.com by 201.21.113.118 on 03/25/15
For personal use only.