predictors of shift work disorder among nurses: a longitudinal study

25
Accepted Manuscript Title: Predictors of shift work disorder among nurses - a longitudinal study Author: Siri Waage, Ståle Pallesen, Bente Elisabeth Moen, Nils Magerøy, Elisabeth Flo, Lee Di Milia, Bjørn Bjorvatn PII: S1389-9457(14)00381-5 DOI: http://dx.doi.org/doi:10.1016/j.sleep.2014.07.014 Reference: SLEEP 2551 To appear in: Sleep Medicine Received date: 14-3-2014 Revised date: 7-6-2014 Accepted date: 3-7-2014 Please cite this article as: Siri Waage, Ståle Pallesen, Bente Elisabeth Moen, Nils Magerøy, Elisabeth Flo, Lee Di Milia, Bjørn Bjorvatn, Predictors of shift work disorder among nurses - a longitudinal study, Sleep Medicine (2014), http://dx.doi.org/doi:10.1016/j.sleep.2014.07.014. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Upload: bjorn

Post on 09-Feb-2017

214 views

Category:

Documents


1 download

TRANSCRIPT

Accepted Manuscript

Title: Predictors of shift work disorder among nurses - a longitudinal study

Author: Siri Waage, Ståle Pallesen, Bente Elisabeth Moen, Nils Magerøy,

Elisabeth Flo, Lee Di Milia, Bjørn Bjorvatn

PII: S1389-9457(14)00381-5

DOI: http://dx.doi.org/doi:10.1016/j.sleep.2014.07.014

Reference: SLEEP 2551

To appear in: Sleep Medicine

Received date: 14-3-2014

Revised date: 7-6-2014

Accepted date: 3-7-2014

Please cite this article as: Siri Waage, Ståle Pallesen, Bente Elisabeth Moen, Nils Magerøy,

Elisabeth Flo, Lee Di Milia, Bjørn Bjorvatn, Predictors of shift work disorder among nurses - a

longitudinal study, Sleep Medicine (2014), http://dx.doi.org/doi:10.1016/j.sleep.2014.07.014.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service

to our customers we are providing this early version of the manuscript. The manuscript will

undergo copyediting, typesetting, and review of the resulting proof before it is published in its

final form. Please note that during the production process errors may be discovered which could

affect the content, and all legal disclaimers that apply to the journal pertain.

Predictors of shift work disorder among nurses - a longitudinal study

Siri Waage, PhD1,2,*

, Ståle Pallesen, PhD2,3

, Bente Elisabeth Moen, MD, PhD1,4

,

Nils Magerøy, MD PhD5, Elisabeth Flo, PhD

1,2, Lee Di Milia, PhD

6, and

Bjørn Bjorvatn, MD, PhD1,2

1 Department of Global Public Health and Primary Care, University of Bergen, Norway

2 Norwegian Competence Center for Sleep Disorders, Haukeland University Hospital,

Norway

3 Department of Psychosocial Science, University of Bergen, Norway

4 Centre for International Health, University of Bergen, Norway

5 Department of occupational medicine, Haukeland University Hospital, Norway

6 Central Queensland University, School of Management and the Institute for Health and

Social Science Research, Rockhampton, QLD, Australia

*Corresponding author / Request for reprints:

Siri Waage

Department of Global Public Health and Primary Care

Kalfarveien 31

5018 Bergen, Norway

E-mail: [email protected]

keywords: Shift work, sleep problems, sleepiness, insomnia, shift work disorder

Highlights

Survey on shift work, sleep and health among 1533 nurses with a two year follow-up

There was a significant reduction in shift work disorder from baseline to follow-up

Several factors were identified to be predictors of shift work disorder

Page 1 of 24

Abstract

Background: Shift work is associated with sleep problems and impaired health. The main

aim of the present study was to explore predictors of developing shift work disorder (SWD)

among Norwegian nurses using a longitudinal design.

Methods: A total of 1533 nurses participating in a survey on shift work, sleep and health

responded to questionnaires at baseline and about two years later at follow-up. SWD was

defined as problems of excessive sleepiness and/or complaints of insomnia related to the work

schedule.

Results and Conclusions: There was a significant reduction (p<.001) in the prevalence of

SWD from baseline to follow-up, from 35.7% to 28.6%. Logistic regression analyses showed

significant risks of having SWD at follow-up and the following variables measured at

baseline; number of nights worked the last year (OR=1.01, 95% CI=1.01-1.02), having SWD

(OR=5.19, 95% CI=3.74-7.20), composite score on the Epworth Sleepiness Scale (OR=1.08,

95% CI=1.04-1.13), use of melatonin (OR=4.20, 95% CI=1.33-13.33), use of bright light

therapy (OR=3.10, 95% CI 1.14-8.39), and symptoms of depression measured by the Hospital

Anxiety and Depression Scale (OR=1.07, 95% CI=1.00-1.14). In addition, leaving night work

between baseline and follow-up was associated with a significant reduced risk of SWD at

follow-up (OR=0.12, 95% CI=0.07-0.22).

Page 2 of 24

Introduction

Estimates suggest that approximately 20% of workers in Europe are employed in some form

of shift work that involves night work [1]. Shift work is associated with a wide range of health

problems [2, 3]. In work schedules that include night work, sleep problems are among the

most frequently reported health problems [4, 5]. The most common sleep problems among

night shift workers are sleep onset and maintenance difficulties, reduced sleep duration and

excessive sleepiness during work [5]. For Norwegian nurses, it is common to work rotating

shifts, which is a work schedule that could involve morning, evening and night shifts within

the same work week. This is reported to lead to insufficient sleep duration [6], and night work

is reported to be the most important cause of long-term stress and fatigue [7]. Nurses in

rotating shifts have reported more mental health problems compared to nurses working non-

rotating shifts [8] and associations between shift work, anxiety and depression have also been

shown [9]. In a recent Norwegian study, nurses working in intensive care units were found to

report poorer sleep and higher levels of sleepiness, fatigue, anxiety and depression compared

to Norwegian norm groups [10]. However, results linking shift work and mental health are

inconsistent; anxiety and depression were for example not associated with night work in a

cross sectional study of Norwegian nurses [11].

Shift work disorder (SWD) is a circadian rhythm sleep disorder characterized by excessive

sleepiness and complaints of insomnia related to the work schedule [12]. The diagnostic

criteria for SWD are described in the second edition of the International Classification of

Sleep Disorders (ICSD-2) and include the following four criteria: 1) Complaint of insomnia or

Page 3 of 24

excessive sleepiness temporally associated with a recurring work schedule that overlaps the

usual time for sleep, 2) symptoms must be associated with the shift work schedule and present

over the course of at least one month, 3) circadian and sleep-time misalignment as

demonstrated by sleep log or actigraphical monitoring for 7 days or more and finally 4) sleep

disturbance is not explainable by another sleep disorder, a medical or neurological disorder,

mental disorder, medication use or substance use disorder [13].

Varying prevalences of SWD have been reported in cross-sectional studies, ranging from

14.5% among police officers working nights [14], 23.3% among oil rig workers [15], 24.4%

among shift working nurses [16], 32.1% among Australian night workers [17], and up to

44.3% among Norwegian nurses in rotating shift work [18]. SWD is a relatively new

diagnosis, and there is uncertainty and discussion about its operationalization, prevalence,

consequences and treatment [19]. At present, few studies have used standardized questions to

measure SWD and SWD seems to be underestimated in clinical settings. Epidemiological data

on SWD are scarce [20]. One limitation of previous research on SWD is that the studies are

based on cross-sectional design which prevent conclusions concerning directionality and

possible cause-effect relationships related to the development and consequences of SWD.

To address this limitation, the main aim of the present study was to explore predictors of

SWD among Norwegian shift working nurses using longitudinal data. We also aimed to

assess the prevalence of SWD among Norwegian nurses at baseline and follow-up.

Methods

Procedure and participants

Page 4 of 24

The data were collected from an on-going longitudinal cohort study “SUrvey of Shift work,

Sleep and Health (SUSSH)” that was initiated in 2008/2009. A sample of 5400 nurses was

randomly selected from the Norwegian Nurses Organization’s membership roll which

includes most of the nurses in Norway. Survey questionnaires have been sent to this sample

annually and this study presents findings from the first (2008/2009=baseline) and the third

wave (2011=follow-up) of the survey.

The sample comprised five equal strata based on the numbers of years since graduation from

nursing school (0-11 months, 1-3 years, 3.1-6 years, 6.1-9 years, and 9.1-12 years). A total of

2059 (response rate = 38.1%) nurses completed the questionnaire in the first wave (during the

period December 2008 to March 2009). During the spring of 2011, all the nurses that

completed the first wave received an invitation to participate in the third wave. A total of

1533 nurses (91% female) responded, yielding a response rate of 78.5%. All questionnaires

were administered by postal mail with a pre-paid envelope for returning the completed forms.

Up to two reminders were sent for each wave to those who did not respond.

Instruments

Demographics

The questionnaires comprised several sections. We collected data on socio-demographic

variables (gender, age, married/cohabiting, children in household), work-related variables

(work schedule, type and percentage of position worked in a full-time capacity, type of work

place, exposure to night work (years) and number of night shifts per year); as well as life-style

variables (e.g. present daily smoking and caffeine consumption/daily cups of coffee, tea or

cola, use of sleep medication, use of bright light treatment).

Page 5 of 24

Shift work disorder

In line with a number of previous studies three questions based on the minimal criteria from

the ICSD-2 were used to assess SWD [15, 17, 18]. The questions are found to adequately

assess SWD in epidemiological contexts [18]. The questions were: 1) Do you experience

either difficulties sleeping or experience excessive sleepiness? (yes/no), 2) Is the sleep or

sleepiness problem related to the work schedule that makes you work when you normally

would sleep? (yes/no), 3) Have you had this sleep or sleepiness problem related to the work

schedule for at least one month? (yes/no). Subjects were classified as suffering from SWD

when they responded “yes” to all three questions.

Insomnia

Insomnia symptoms were assessed with the Bergen Insomnia Scale (BIS) [21]. The

questionnaire consists of 6 items, where the response alternatives reflects the number of days

per week (0-7) the respondent had experienced a specific insomnia symptom, and is based on

the diagnostic criteria for insomnia found in the fourth edition of the Diagnostic and

Statistical Manual for Mental Disorders (DSM-IV, American Psychiatric Association, 2000).

The scores on each item are added to create a total composite score (0 to 42), and higher

values indicate a greater degree of insomnia symptoms [21]. The Norwegian version of the

BIS has been validated and had shown good psychometric properties [21]. In the present

study, the Cronbach’s alpha was .83 for the baseline data and .82 for the follow-up data.

Sleepiness

Sleepiness was measured with the Norwegian version of the Epworth Sleepiness Scale (ESS)

[22]. The ESS consists of 8 items that measure the subject’s general tendency to sleep or doze

off in 8 different situations. Each item is scored from 0 (no probability) to 3 (high probability),

Page 6 of 24

yielding a total score between 0 and 24. The Norwegian version has shown high validity and

reliability [23]. In the present study the Cronbach’s alpha coefficients were .74 for the

baseline data and .75 for the follow-up data, respectively.

Anxiety and depression

Symptoms of anxiety and depression were assessed with the Norwegian version of the

Hospital Anxiety and Depression Scale (HADS) [24]. This scale consists of 14 items

measuring symptoms of anxiety and depression experienced during the last week. Seven

questions specifically address symptoms of anxiety whereas the other seven questions pertain

specifically to symptoms of depression. The items are rated on a 4-point scale (0-3), yielding

two scores, one for anxiety and one for depression, each ranging from 0 and 21. The HADS

has shown good reliability [25]. In the present study the Cronbach’s alpha coefficients for

both the anxiety and depression subscales were .81 for the baseline data. For the follow-up

data the coefficients were .82 for both the scales.

Morningness

A Norwegian version of the Diurnal Type Scale [26] was used to assess the morningness-

eveningness dimension. The scale consists of 7 items related to this dimension, each rated on

a four-point scale. Higher scores indicate higher levels of morningness (e.g. preference for

rising relatively early in the day, performing activities relatively early in the day and getting

to bed relatively early in the evening). The Diurnal Type Scale has been shown to have high

internal reliability and validity. The Cronbach’s alpha coefficient was .64 at both baseline and

follow-up in the present study.

Ethics

Page 7 of 24

The Regional Committee for Medical and Health Research Ethics of Western Norway (REK-

West) approved the study.

Statistics

PASW Statistics 18 for Windows was used for the statistical analyses. For comparisons of

demographic variables between the nurses with and without SWD at baseline, independent t-

tests and Pearson chi-square tests were used. Significance level was set at .05.

The nurses were divided into four different groups based on having SWD or not at the two

time points of the study. Group one consisted of workers not having SWD in either of the two

waves (54.9%, n=817). The second group comprised nurses defined with SWD solely at

baseline, and not at follow-up (16.5%, n=246). Group three consisted of nurses defined with

SWD solely at follow-up and not at baseline (9.7%, n=144). Finally, the fourth group

comprised of nurses defined with SWD at both baseline and follow-up (18.9%, n=281).

Paired-sample t-tests were used to compare the change for each of the sleep and health

parameters from baseline to follow-up within each of the four groups.

Logistic regression analyses were performed to assess the impact of a number of factors on

the likelihood of nurses having SWD at follow-up. The latter comprised the dependent

variable (0=not having SWD, 1=having SWD). We included age, gender, marital status,

children in household, use of sleep medication (prescription or over the counter), use of

exogenous melatonin, bright light treatment, smoking, caffeine consumption, number of

nights worked the last 12 months, SWD, sleepiness score, insomnia score, anxiety score,

depression score and diurnal type all measured at baseline as independent variables.

Page 8 of 24

Furthermore, we included no longer working night shifts at follow-up (yes/no) also as an

independent variable. All variables were first entered separately (crude analyses) and

secondly they were all entered together at the same time in an adjusted analysis.

Furthermore, logistic regression analyses were performed to assess the impact of the same

factors on the likelihood of disappearance of SWD among nurses having SWS at baseline.

Results

Demographics

At baseline (n=1533), the mean age of the nurses was 33 years, range 21-63. Among the

whole sample, 76.3% worked in somatic hospital departments, 13.5% in psychiatric

departments, 3.5% in nursing homes, 3.7% in home care services, and 2.1% in other work

places, respectively. The mean hours worked per week were 33.9 with a distribution of 2.8%

working <50% position, 28.6% working between 50-75% of full time position, 13.4%

working between 76-90% of full time position, and 55.2% working more than 90% of full

time position. Previous or present night work was reported by 84.3% of the nurses. Working

only daytime was reported by 7.6% of the nurses, 0.1% reported only evening work, 25.0%

worked a two-shift schedule involving day work and evening work, 8.2% worked nights only,

55.0% worked a three-shift schedule involving day, evening and night work, and 3.1%

reported working other schedules involving night work. 14.0% of the nurses quitted night

work between baseline and follow-up. A total of 74.0% of the nurses were married or

cohabiting, and 48.7% reported having children at home. Mean body mass index (BMI)

among the nurses was 24.4 (range 17.3-47.9). The nurses reported a mean of 3 cups of

caffeinated beverages per day (range 0-30), and 10.0% were daily smokers (daily smoking,

yes/no).

Page 9 of 24

Shift work disorder

A total of 54.9% of the nurses did not have SWD at baseline or at follow-up, 16.5% were

defined with SWD at baseline only, 9.7% at follow-up only, and 18.9% of the nurses were

defined with SWD at both baseline and follow-up. Analysing data from all the nurses, there

was a significant reduction (p<.001) in the prevalence of SWD between baseline (35.7%,

n=538) and follow-up (28.6%, n=433). Among the nurses from wave 1 who did not respond

in wave 3, the prevalence of SWD was 33.2% at baseline. The differences in demographics

between nurses with and without SWD at baseline are described in table 1. Having SWD at

baseline was positively associated with male gender, age, currently or previously working

night shifts, numbers of nights the last year, and inversely related to the score on the

morningness dimension (table 1).

Insert table 1 about here

The differences between the two measurements (baseline and follow-up) across the four

groups are presented in table 2. Interestingly, the group that no longer met the criteria for

SWD at follow-up worked significantly fewer nights and also reported a reduction in scores

on the ESS, the BIS, and on symptoms of anxiety and depression. In that group 30.7% of the

nurses had quit night work between the two assessment points, and the mean number of nights

worked the last year was reduced from 27.9 to 18.5 nights. The reductions in the ESS, BIS,

anxiety and depression scores were similar among the nurses who no longer worked night

shifts, and those still working nights shifts at follow-up. For the group that developed SWD at

follow-up, exposure to night work and scores for BIS increased.

Insert table 2 about here

Page 10 of 24

The results from the logistic regression analyses showed a risk of having SWD at follow-up

and the following variables assessed at baseline; use of melatonin, use of bright light therapy,

number of nights worked the last year, having SWD, sleepiness scores, and depression score

(see table 3). In addition, leaving night work between baseline and follow-up significantly

reduced the risk for fulfilling the criteria for SWD at follow-up. In addition, logistic

regression analyses were performed among the nurses who had SWD at baseline to

investigate predictors of cessation of SWD. Not surprisingly, no longer working night shift at

follow-up was the strongest predictor of disappearance of SWD from baseline to follow-up.

Furthermore, numbers of nights worked per year and sleepiness measured by the ESS at

baseline were positively associated maintaining SWD (see table 4).

Insert table 3 and 4 about here

Discussion

Having SWD at baseline, use of exogenous melatonin, use of bright light therapy, number of

nights worked last year, sleepiness score, and depression score were all found to be predictors

of SWD among Norwegian nurses. In addition, quitting night shifts from baseline to follow-

up decreased the risk of SWD at follow-up. Interestingly, there was overall a significant

reduction in the prevalence of shift work disorder (SWD) from 35.7% at baseline to 28.6% at

follow-up.

Measured at baseline, the nurses with SWD were slightly older, comprised more males, were

presently or previously working nights, and worked more nights the last year than the nurses

without SWD. This is in line with studies showing that sleep problems related to shift work

Page 11 of 24

increase with age [27, 28]. However, the difference in age between the two groups was small,

only about one year, and the importance of this in terms of tolerance to shift work seems

negligible. Similarly, most of the other demographic differences between nurses with or

without SWD at baseline are small. One exception is number of nights worked the last year,

where nurses with SWD worked on average nearly 32 nights per year, while the nurses

without SWD worked about 22 nights per year. Also, morning types are reported to have

more difficulties adjusting their circadian rhythms to night work [29]. However, in this study

the nurses with SWD scored lower on the morningness dimension compared to the nurses

with SWD. In line with this, a study on oil rig workers found no significant difference in

diurnal type between workers with SWD and workers without SWD [15]. Not surprisingly,

subjects with SWD scored higher on measures of sleepiness, anxiety and depression than

subjects without SWD.

The strongest predictor for SWD at follow-up was having SWD (OR 5.2) at baseline. This

suggests that the sleep related problems that are experienced by shift workers are chronic for a

significant proportion of the nurses. The number of night shifts worked per year was also a

significant predictor, in line with results from a previous cross-sectional study from SUSSH

including nurses from the same cohort [18]. Changing the work schedule so that it no longer

included night work significantly predicted reduced risk of SWD at follow-up, supporting the

fact that night work seems to be a main determinant of SWD. One of the criteria for SWD is

to have complaints of insomnia or excessive sleepiness associated with a work schedule that

overlaps with the usual time for sleep. Previous studies have reported that SWD is particularly

prevalent among shift workers working night shift and early-morning shifts [30]. Excessive

sleepiness and/or insomnia are criteria for SWD and are frequently reported among shift

workers [4, 5]. Sleepiness score at baseline was one of the predictors of reporting SWD at

Page 12 of 24

follow-up. It therefore was somewhat surprising that this was not the case for insomnia.

However, in our study insomnia was measured by the Bergen Insomnia Scale, which is based

on the diagnostic criteria for insomnia stated in the DSM-IV (American Psychiatric

Association, 2000). The criteria for SWD do not require insomnia as defined by the DSM-IV

diagnostic criteria, which is likely to be more stringent than self-reported insomnia symptoms.

The DSM-IV criteria requests difficulty in falling or maintaining sleep, or the feeling of not

having had restitutional sleep for at least one month in addition to a daytime impairment,

while the questions used to define SWD in this study did for instance not include any

questions about daytime consequences of the sleep problem. However, at baseline the mean

total score of the Epworth Sleepiness Scale (ESS) was about 9.5 among the nurses with SWD

and even lower about (7.7) among the nurses without SWD. A value of 11 or higher on the

total score of the ESS is considered as pathological sleepiness, meaning that the nurses in this

study overall are feeling sleepy, albeit still not above the clinical cut-off value.

Shift workers struggling with adjustment to different work schedules are likely to search for

interventions in order to better adjust to shift work and the problems such work schedules

pose. Adapting circadian rhythms to the work schedule or to adjust it to a normal day

schedule following e.g. a period of night shifts is assumed to be curative of shift work related

problems and SWD [31]. Bright light therapy and exogenous melatonin are as such

recommended in the current treatment guidelines for SWD [30]. Both use of melatonin and

bright light were predictors of having SWD at follow-up. One possible explanation could be

that the nurses reporting use of melatonin or bright light treatment are the nurses with the

most severe sleep or circadian problems, and that the SWD remains as long as they continue

working shift work. A weakness of the study is that we neither collected information on the

timing and dosage of melatonin, nor information about the timing and dosage of bright light.

Page 13 of 24

Future research on the effect of treatment in large randomized clinical trials of shift workers

with SWD is needed to make recommendations regarding the effects of melatonin and bright

light treatment for this disorder.

One of the predictors of SWD was depression at baseline. This is in accordance with results

from a cross-sectional study showing that shift working nurses with SWD showed more

severe depressive symptoms than those without SWD [16]. The strength of our study is the

longitudinal design, being the first study to identify depression as a predictor of developing

SWD. In the cross-sectional study from SUSSH by Flo and co-workers (2012) an association

between SWD and anxiety was reported, but after adjusting for multiple variables, neither

symptoms of anxiety nor depression were associated with SWD [18].

The prevalence of SWD reported in our study was high. Cross-sectional data from the same

Norwegian nurse population including nearly 2000 nurses at baseline, have previously

reported an even higher SWD prevalence ranging from 32.4% to 37.6% depending on the

operationalization [18]. For comparison, a recent cross-sectional study from Japan reported a

prevalence of 24.4% among 997 female hospital nurses engaged in two-shift and three-shift

schedules [16]. The discrepancy between the prevalence rates across these studies could be

explained by differences in working hours and working conditions between Norway and

Japan. For instance the typical night shift start/end times differed between the samples (Japan;

24:00 - 09:00 compared to Norway; 21:30 - 07:00). Other possible explanations for variances

in reported prevalence rates may be differences in the estimation of SWD and methodology.

However, several of the reported studies have used the same questions to estimate SWD,

which highlights that in addition to work schedule differences, individual differences may

also play a role [17].

Page 14 of 24

In the present study, the prevalence of SWD decreased by about 7% from baseline to follow-

up. Different possible factors could be considered to explain this reduction. Like in other

studies on shift workers, it is reasonable to assume that a type of selection bias might be

present, where individuals not coping with shift work tend to terminate this type of work [32].

In the present study, 30.7% of the nurses who had SWD at baseline, but not at follow-up, had

quit night work between the two time points. Among the whole sample, 14% of the nurses

made the same change. This suggests that vulnerable nurses self-selected out of night shift

work. In addition, it is also possible that during the period from baseline to follow-up the

nurses developed better coping strategies to deal with irregular work hours. It is also possible

that this decline in prevalence can be explained by organizational changes in working

conditions or wages, which we do not have information about. Interestingly, the reduction in

SWD prevalence between baseline and follow-up was evident in all different shift work

schedules.

Limitations and strengths

Some limitations of the study should be mentioned. The response rate from the first wave was

only 38% and some may thus question the representativeness of the population. Still, the

response rate at follow-up was high (79%) ensuring that we compared a large sample across

the two time points. The sample comprised only nurses, mainly females, which also pose

some threat to the external validity. On the other hand the relative homogenous sample

reduces the risk of work related confounders to influence the results.

Research on shift work as a risk factor for negative health has some methodological

challenges regarding shift work exposure. A limitation of the study is that we do not have

information about shift work experience before being included in the present cohort study.

Page 15 of 24

Future research should therefore include workers early in their shift work career, as many

nurses may have been exposed to night work before and during their education. Still, a

strength of the study is that the questionnaire consists of detailed information about present

shift work exposure, in addition to the longitudinal design with information about changes in

work schedules among the nurses.

Another limitation of the study includes its reliance on self-reported data only, and no

objective assessments. However, as primarily baseline data were used as predictors for SWD

at follow-up, this significantly reduces the risk of the results being distorted by the common

method bias [33]. Lately, a new version of the ICSD has been published, with an important

change in the criteria for SWD. In ICSD-2 the symptoms associated with the SWD should be

present for at least one month, while in the ICSD-3 this is changed to three months. Future

research should be conducted in accordance with the changes in the criteria.

Many of the odds ratios of the predictors on SWD in the present study were quite small. Still,

it should be noted that many of the variables in the analyses are not dichotomous, and the

odds ratios are consequently not expected to be high. One important factor that was

significant, but had a low odd ratio, was number of night worked last year. This variable was

continuous ranging from 0 to 200 nights per year, meaning that with an odd ratio of 1.01, an

increase of one night shift per year would increase the risk of developing SWD with one

percent. Nevertheless, when interpreting the significant results, the size of the odds ratio

needs to be taken into careful consideration.

The present study also has some valuable assets worth mentioning. As far as we know, this

study is the first longitudinal study assessing SWD, providing an unique opportunity to

Page 16 of 24

investigate directionality between variables and to identify possible causes of SWD.

Longitudinal designs are superior to cross-sectional studies when it comes to discovering

directionalities between variables. Still, the present study would be better if we had included

more than two assessment times. The study furthermore included standardized and well-

validated instruments and the relatively large sample size provides adequate statistical power

to the analyses.

Conclusion

Several factors measured at baseline like reporting SWD, use of exogenous melatonin, use of

bright light therapy, number of nights worked last year, sleepiness score, and depression score

were found to be significant predictors of SWD at follow-up. In addition, quitting night work

between baseline and follow-up was significantly associated with a decreased risk of SWD at

follow-up, suggesting that night work may be a major cause of SWD. There was a significant

reduction in the SWD prevalence rate from 35.7% at baseline to 28.6% at follow up that

might reflect selection of SWD subjects to other types of work schedules, development of

better coping strategies over time or organizational changes.

Page 17 of 24

References

[1] Parent-Thirion A, Vermeylen G, van Houten G, Lyly-Yrjänäinen M, Biletta I, Cabrita J.

Fifth European Working Conditions Survey 2012.

[2] Costa G. The impact of shift and night work on health. Appl Ergon 1996;27:9-16.

[3] Harma M. Are long workhours a health risk? Scand J Work Environ Health 2003;29:167-

9.

[4] Akerstedt T. Shift work and disturbed sleep/wakefulness. Sleep Med Rev 1998;2:117-28.

[5] Akerstedt T. Shift work and disturbed sleep/wakefulness. Occup Med 2003;53:89-94.

[6] Chan MF. Factors associated with perceived sleep quality of nurses working on rotating

shifts. J Clin Nurs 2009;18:285-93.

[7] Winwood PC, Winefield AH, Lushington K. Work-related fatigue and recovery: the

contribution of age, domestic responsibilities and shiftwork. J Adv Nurs 2006;56:438-49.

[8] Jamal M. Shift Work Related to Job-Attitudes, Social-Participation and Withdrawal

Behavior - a Study of Nurses and Industrial-Workers. Pers Psychol 1981;34:535-47.

[9] Pati AK, Chandrawanshi A, Reinberg A. Shift work: consequences and management. Curr

Sci 2001;81:32-52.

[10] Bjorvatn B, Dale S, Hogstad-Erikstein R, Fiske E, Pallesen S, Waage S. Self-reported

sleep and health among Norwegian hospital nurses in intensive care units. Nurs Crit Care

2012;17:180-8.

[11] Oyane NM, Pallesen S, Moen BE, Akerstedt T, Bjorvatn B. Associations between night

work and anxiety, depression, insomnia, sleepiness and fatigue in a sample of norwegian

nurses. PLoS One 2013;8:e70228.

Page 18 of 24

[12] Schwartz JR, Roth T. Shift work sleep disorder: burden of illness and approaches to

management. Drugs 2006;66:2357-70.

[13] American Academy of Sleep M. International Classification of Sleep Disorders 2.nd ed,

Diagnostic and Coding Manual. Westchester, Illinois: American Academy of Sleep Medicine;

2005.

[14] Rajaratnam S, Barger L, Lockley S, Cade B, O'Brien C, White D, et al. Screening for

sleep disorders in north American police officers. Sleep 2007;30:A209-A.

[15] Waage S, Moen BE, Pallesen S, Eriksen HR, Ursin H, Akerstedt T, et al. Shift work

disorder among oil rig workers in the North Sea. Sleep 2009;32:558-65.

[16] Asaoka S, Aritake S, Komada Y, Ozaki A, Odagiri Y, Inoue S, et al. Factors Associated

With Shift Work Disorder in Nurses Working With Rapid-Rotation Schedules in Japan: The

Nurses' Sleep Health Project. Chronobiol Int 2013;30:628-36.

[17] Di Milia L, Waage S, Pallesen S, Bjorvatn B. Shift Work Disorder in a Random

Population Sample - Prevalence and Comorbidities. Plos One 2013;8.

[18] Flo E, Pallesen S, Mageroy N, Moen BE, Gronli J, Nordhus IH, et al. Shift work disorder

in nurses--assessment, prevalence and related health problems. PLoS One 2012;7:e33981.

[19] Sack RL, Auckley D, Auger RR, Carskadon MA, Wright KP, Jr., Vitiello MV, et al.

Circadian rhythm sleep disorders: part I, basic principles, shift work and jet lag disorders. An

American Academy of Sleep Medicine review. Sleep 2007;30:1460-83.

[20] Culpepper L. The social and economic burden of shift-work disorder. J Fam Pract

2010;59:S3-S11.

[21] Pallesen S, Bjorvatn B, Nordhus IH, Sivertsen B, Hjornevik M, Morin CM. A new scale

for measuring insomnia: The Bergen Insomnia Scale. Percept Mot Skills 2008;107:691-706.

[22] Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness

scale. Sleep 1991;14:540-5.

Page 19 of 24

[23] Pallesen S, Nordhus IH, Omvik S, Sivertsen B, Tell GS, Bjorvatn B. Prevalence and risk

factors of subjective sleepiness in the general adult population. Sleep 2007;30:619-24.

[24] Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiol Scand

1983;67:361-70.

[25] Bjelland I, Dahl AA, Haug TT, Neckelmann D. The validity of the Hospital Anxiety and

Depression Scale. An updated literature review. J Psychosom Res 2002;52:69-77.

[26] Torsvall L, Akerstedt T. A diurnal type scale. Construction, consistency and validation in

shift work. Scand J Work Environ Health 1980;6:283-90.

[27] Harma, M. Ageing, physical fitness and shiftwork tolerance. Appl Ergonom 1996;27, 25-

29.

[28] Marquié, J. C. Sleep, age, and shiftwork experience. J Sleep Res 1999;8, 297-304.[29]

[29] Arendt, J. Shift work: coping with the biological clock. Occup Med (Lond) 2010;60, 10-

20.

[30] Thorpy M. Understanding and Diagnosing Shift Work Disorder. Postgrad Med

2011;123:96-105.

[31] Wright KP, Bogan RK, Wyatt JK. Shift work and the assessment and management of

shift work disorder (SWD). Sleep Med Rev 2013;17:41-54.

[32] Knutsson A, Akerstedt T. The Healthy-Worker Effect - Self-Selection Among Swedish

Shift Workers. Work and Stress 1992;6:163-7.

[33] Podsakoff PM, MacKenzie SB, Lee JY, Podsakoff NP. Common method biases in

behavioral research: a critical review of the literature and recommended remedies. J Appl

Psychol 2003;88:879-903.

Page 20 of 24

Table 1. Comparison of demographic variables between nurses with shift work disorder

(SWD; n=525) and nurses without SWD (n=947) at baseline. Not SWD SWD p-value

Age mean years 33.0 34.2 .006 a

Sex female/male % 92.3/7.7 88.5/11.5 .012 b

Married/cohabiting in % 75.3 73.9 .545 b

Children living at home % yes 51.3 50.4 .740 b

Body Mass Index 24.5 24.5 .798 a

Daily caffeine use in cups 3.0 3.2 .053 a

Daily smoking % yes 9.5 12.5 .077 b

Work hours per week 33.9 33.9 .816 a

Nurse experience in years 5.1 5.4 .129 a

Night work (present or previous) % yes 83.6 88.6 .008 b

Number of nights worked last year 22.3 31.7 <.001 a

Mean hours of sleep per night 7.05 6.77 <.001 a

Sum Epworth Sleepiness Scale 7.67 9.52 <.001 a

Sum Bergen Insomnia Scale 10.47 18.16 <.001 a

Sum Anxiety (HADS-A) 3.81 5.96 <.001 a

Sum Depression (HADS-D) 2.01 3.95 <.001 a

Sum Diurnal Scale 18.1 16.9 <.001 a

a Independent samples t-test,

b Pearson chi-square test

Page 21 of 24

Table 2. Comparisons of four groups of nurses, with or without shift work disorder (SWD) at baseline and follow-up by paired-samples t-tests.

No SWD at baseline or

at follow-up

SWD at baseline

No SWD at follow-up

No SWD at baseline

SWD at follow-up

SWD at both baseline and

at follow-up

Baseline Follow-up p-value Baseline Follow-up p-value Baseline Follow-up p-value Baseline Follow-up p-value

Number of nights last year 21.7 21.3 .75 27.9 18.5 <.001 26.6 38.2 <.001 35.2 37.9 .17

Work hours per week 37.1 32.9 <.001 33.4 31.9 .01 33.5 32.3 .048 34.2 33.6 .10

Mean hours of sleep per night 7.1 7.1 .65 6.8 6.9 .10 7.0 6.9 .04 6.7 6.7 .79

Sum Epworth Sleepiness Scale 7.5 7.2 .01 9.4 7.7 <.001 8.5 9.1 .06 9.8 9.8 .65

Sum Bergen Insomnia Scale 10.1 9.6 .05 17.6 13.4 <.001 12.9 14.6 .005 18.3 18.4 .80

Sum Anxiety (HADS-A) 3.7 3.8 .32 5.8 5.0 <.001 4.2 4.3 .89 6.1 5.9 .48

Sum Depression (HADS-D) 1.9 1.9 .58 3.7 3.0 <.001 2.6 2.7 .78 4.2 4.0 .38

HADS = Hospital Anxiety and Depression Scale

Page 22 of 24

23

Table 3. Logistic regression analyses predicting shift work disorder (SWD) at follow-up

among Norwegian nurses. In the crude analyses each independent variable is analysed one by

one against SWD, in the adjusted analysis they are all included simultaneously. Crude

(N=1440 to 1516)

Adjusted

(N=1196)

Independent variable OR 95% CI OR 95% CI

Age 1.02 1.01-1.04 1.02 1.00-1.04

Gender (female = 1.00) Male 1.19 0.82-1.74 0.88 0.53-1.48

Married/Cohabiting (yes = 1.00) No 1.18 0.92-1.52 1.24 0.86-1.79

Children in household (yes = 1.00) No 0.99 0.79-1.24 .84 0.60-1.18

Sleep medication (no =1.00) Yes 2.40 1.64-3.51 1.03 0.59-1.83

Melatonin (no =1.00) Yes 3.36 1.62-6.99 4.20 1.33-13.33

Bright light (no =1.00) Yes 2.29 1.18-4.45 3.10 1.14-8.39

Over the counter sleep medication (no =1.00) Yes 2.64 1.64-4.25 1.51 0.78-2.93

Smoking (no = 1.00) Yes 1.37 0.96-1.95 0.98 0.60-1.62

Caffeine (daily consumption) 1.06 1.01-1.10 0.99 0.93-1.06

No longer working night shift at follow-up (no = 1.00) Yes 0.22 0.13-0.35 0.12 0.07-0.22

Number of nights per year at baseline 1.01 1.01-1.01 1.01 1.01-1.02

SWD at baseline (no = 1.00) Yes 6.48 5.07-8.29 5.19 3.74-7.20

ESS total score at baseline 1.11 1.07-1.14 1.08 1.04-1.13

BIS total score at baseline 1.07 1.06-1.08 1.02 1.00-1.04

Sum Anxiety (HADS-A) 1.11 1.07-1.14 0.99 0.94-1.06

Sum Depression (HADS-D) 1.16 1.12-1.21 1.07 1.00-1.14

Sum Diurnal Scale 0.94 0.91-0.97 1.01 0.96-1.06

ESS = Epworth Sleepiness Scale, BIS = Bergen Insomnia Scale, HADS = Hospital Anxiety

and Depression Scale, bold = p-values ≤ 0.05

Page 23 of 24

24

Table 4. Logistic regression analyses predicting disappearance of shift work disorder (SWD)

at follow-up among Norwegian nurses with SWD at baseline. In the crude analyses each

independent variable is analysed one by one against SWD, in the adjusted analysis they are all

included simultaneously. Crude

(N=503 to 527)

Adjusted

(N=429)

Independent variable OR 95% CI OR 95% CI

Age 0.98 0.96-1.00 0.97 0.94-1.00

Gender (female = 1.00) Male 0.93 0.55-1.59 1.02 0.51-2.04

Married/Cohabiting (yes = 1.00) No 0.77 0.52-1.14 0.77 0.46-1.30

Children in household (yes = 1.00) No 0.83 0.58-1.17 1.00 6.21-1.61

Sleep medication (no =1.00) Yes 0.53 0.31-0.89 0.71 0.35-1.46

Melatonin (no =1.00) Yes 0.56 0.22-1.41 0.38 0.84-1.69

Bright light (no =1.00) Yes 0.99 0.36-2.80 0.41 0.05-3.27

Over the counter sleep medication (no =1.00) Yes 0.65 0.35-1.21 0.54 0.23-1.25

Smoking (no = 1.00) Yes 1.09 0.65-1.85 1.14 0.58-2.23

Caffeine (daily consumption) 0.90 0.84-.0.97 0.97 0.88-1.07

No longer working night shift at follow-up (no = 1.00) Yes 8.23 4.50-15.05 8.46 4.20-17.00

Number of nights per year at baseline 0.99 0.99-1.00 0.99 0.98-1.00

ESS total score at baseline 0.98 0.93-1.02 0.94 0.89-1.00

BIS total score at baseline 0.99 0.97-1.01 1.00 0.98-1.03

Sum Anxiety (HADS-A) 0.98 0.93-1.02 1.01 0.94-1.09

Sum Depression (HADS-D) 0.96 0.90-1.01 0.93 0.85-1.02

Sum Diurnal Scale 1.02 0.97-1.07 0.99 0.93-1.06

ESS = Epworth Sleepiness Scale. BIS = Bergen Insomnia Scale. HADS = Hospital Anxiety

and Depression Scale. bold = p-values ≤ 0.05

Page 24 of 24