ORIGINAL RESEARCH: EMPIRICAL RESEARCH –

QUANTITATIVE

The associations of supervisor support and work overload with

burnout and depression: a cross-sectional study in two nursing
settings
Matthias Weigl, Nicole Stab, Isabel Herms, Peter Angerer, Winfried Hacker & Jürgen Glaser

Accepted for publication 26 January 2016

Correspondence to M. Weigl: WEIGL M., STAB N., HERMS I., ANGERER P., HACKER W. & GLASER J.
e-mail: [email protected] ( 2 0 1 6 ) The associations of supervisor support and work overload with burnout
and depression: a cross-sectional study in two nursing settings. Journal of
Matthias Weigl Dr Dipl-Psych
Advanced Nursing 72(8), 1774–1788. doi: 10.1111/jan.12948
Researcher in Organizational Behavior
and Health
Institute and Outpatient Clinic for Abstract
Occupational, Social, and Environmental Aims. To investigate the moderating effects of work overload and supervisor
Medicine, Ludwig-Maximilians-University support on the emotional exhaustion–depressive state relationship.
Munich, Germany Background. Burnout and depression are prevalent in human service
Nicole Stab MPH Dr professionals and have a detrimental impact on clients. Work overload and
Researcher in Work Organization and supervisor support are two key job demands and job resources, whose role and
Health interplay for the development and maintenance of burnout and depression are not
Department of Psychology, Technical fully understood yet.
University at Dresden, Germany
Design. Two consecutive cross-sectional surveys: survey 1 investigated 111
Isabel Herms Dip-Psych hospital nursing professionals and survey 2 examined 202 day care professionals.
Psychologist Data collection was completed in 2010.
BAD Occupational Health and Safety Results. After controlling for general well-being and sociodemographic
Group, Munich, Germany
characteristics, nurses’ emotional exhaustion was associated with increased
Peter Angerer Dr med depressive state in both samples. We found a meaningful three-way interaction:
Professor for Occupational and Social our results show consistently that the relationship between emotional exhaustion
Medicine
and depressive state was strongest for nurses with high work overload and low
Institute for Occupational and Social
supervisor support. Additionally, nurses with low work overload and low
Medicine, Medical Faculty, D€ usseldorf
University, Germany supervisor support were also found to have stronger associations between
emotional exhaustion and depressive state.
Winfried Hacker Dr
Conclusion. The findings indicate that nurses’ reported supervisor support exerts
Senior Professor for Work Psychology
Department of Psychology, Technical
its buffering effect on the burnout-depression link differentially and serves as an
University at Dresden, Germany important resource for nurses dealing with high self-reported work stress.

Jürgen Glaser Dr phil

Keywords: burnout, day care, depression, hospitals, nursing, supervisor support,
Professor for Applied Psychology
Institute of Psychology, University of work overload
Innsbruck, Austria

1774 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

Background
Why is this research needed?
Meta-analyses show that psychosocial stressors at work are
• The role of job demands and job resources in the develop-
associated with an increased risk of depression, while psy-
ment of depression has become an issue of growing interest
chological resources, such as job control and social support
in nursing practice and research.
• Particular concern has been expressed over the contribu- at work, are related to a lower risk (Tennant 2001, Bonde
tory role of nursing professionals’ burnout and contextual 2008, Netterstrom et al. 2008, Theorell et al. 2015). There
determinants in the workplace. is still inconclusive evidence for the complex mechanisms
and the interplay of work conditions and strain that
What are the key findings? contribute to depression (Bonde 2008).
Burnout and depression are conceptually and empirically
• In two cross-sectional surveys (hospital and day care nurs-
ing), nurses with higher emotional exhaustion reported distinct constructs (Schaufeli & Enzmann 1998, Maslach
more depressive symptoms. et al. 2001, Toker & Biron 2012). Emotional exhaustion is
• More supervisor support was associated with less depres- considered the core dimension among all burnout conceptu-
sive state. alizations (Maslach et al. 2001, Shirom 2003). Whereas
• We observed a three-way interaction between nurses’ emo- burnout is specific to the work environment, depression is a
tional exhaustion, work overload and supervisor support in feeling that is also pervasive outside of work, crossing into
determining depressive state, such that the association several domains of the private life (Schaufeli & Enzmann
between emotional exhaustion and depression disappeared 1998, Iacovides et al. 2003, Toker & Biron 2012). Depres-
if nurses were working in jobs with low work overload
sion is defined as a multi-faceted, psychiatric morbidity
and high supervisor support.
with affective, cognitive and physiological impairments with
a specific set of symptoms, such as diminished interest and
How should the findings be used to influence policy/
practice/research/education? pleasure, sad or empty mood, fatigue and loss of energy,
feelings of worthlessness and diminished ability to think
• Supervisors should be supportive to buffer depressive states
and concentrate (DSM V: APA 2013).
in highly strained nursing professionals.
Psychometric investigations confirm the discriminant valid-
• Health-oriented work design interventions that address
ity of burnout and depression measures (e.g. Schaufeli &
nurses’ work load and enhance social support by supervi-
sors should be implemented and evaluated. Enzmann 1998). However, several researchers pointed out
that fatigue and lack of energy are inherent to both con-
structs (e.g. Iacovides et al. 2003, Toker & Biron 2012).
Although both states are independent factors, they share a
Introduction
significant amount of common variance (Schaufeli & Enz-
Burnout and depression pose a major public health chal- mann 1998, Iacovides et al. 1999, Ahola et al. 2005, Peter-
lenge to the workforce. A significant proportion of the son et al. 2008, Toker & Biron 2012, Bianchi et al. 2015).
workforce experiences depressive symptoms, e.g. 12-month Occupational burnout has been considered as an early
major depression prevalence of 6
4% for the USA, 4
6% stage in the development of work-related depression (Ahola
for Canada and 6
5% for Germany (Kessler et al. 2006, et al. 2005, Ahola & Hakanen 2007). According to Hob-
Roesler et al. 2006, Blackmore et al. 2007). Burnout and foll’s Conservation of Resources Theory (Hobfoll & Wells
depression are a significant burden particularly among 1998, Hobfoll 2001), individuals seek to replenish missing
employees in human service professions (Iacovides et al. individual resources. But when such efforts fail, individuals
1999, Wieclaw et al. 2006, Peterson et al. 2008). Specifi- are at greater risk of further exacerbating or depleting indi-
cally, healthcare workers have an elevated risk for psychi- vidual resources. Failures in preserving valuable resources
atric morbidities, i.e. depression (Tennant 2001). In an US trigger depressive symptoms as employees feel incapable of
study, a depressive symptom rate was observed among responding efficiently to high job demands and of coping
18% of the surveyed nurses (Letvak et al. 2012). Depressed with high strain (Toker & Biron 2012). Research supports
employees struggle to maintain presence, functioning and this notion of a downward spiral that is initially triggered
productivity at the workplace (e.g. Lerner et al. 2010). by increased burnout (Ahola & Hakanen 2007, Hakanen
Specifically for healthcare professionals, depression is not et al. 2008, Toker & Biron 2012).
only a hazard for themselves but also poses significant Although the association of burnout and depression has
safety risks for patients (e.g. Gartner et al. 2010). received considerable attention (e.g. Iacovides et al. 1999,

© 2016 John Wiley & Sons Ltd 1775

M. Weigl et al.

Ahola et al. 2005, Peterson et al. 2008), work conditions

Aims and hypotheses
that attenuate this relationship have been largely unex-
plored. Studies show that high work demands (such as time Our aim was to disentangle the role of work conditions,
pressure or work overload) contribute to exhaustion specifically work overload and supervisor support, on the
(Demerouti et al. 2001, Bakker et al. 2004, Bakker & emotional exhaustion–depression association. In line with
Demerouti 2007). Similarly, intense work demands and previous research, we first sought to confirm that exhaus-
overload were found to contribute to depression (Rau et al. tion is associated with depressive state (e.g. Ahola et al.
2010, Shirom et al. 2010). 2005, Ahola & Hakanen 2007, Hakanen et al. 2008, Peter-
Work resources enable highly strained employees to son et al. 2008). Consistent with Conservation of Resources
maintain functioning (Demerouti et al. 2001, Bakker et al. Theory, employees with increased emotional exhaustion are
2004, Bakker & Demerouti 2007). Work resources com- at risk for depleted regulatory resources (Hobfoll 2011).
prise psychological, social, or organizational characteristics Chronic resource depletion like persistent fatigue leads to
that: (1) reduce job-related physiological and psychological reduced affective states, indicated by worsening mood
costs or stressors; (2) are instrumental in achieving work (Hobfoll 2001, Skapinakis et al. 2004). Affective deteriora-
goals; and (3) intrinsically motivate and stimulate personal tion at work renders nurses more vulnerable to subsequent
learning and development (Demerouti et al. 2001, Bakker affective deterioration in general personal life (Hobfoll
& Demerouti 2007). In regard to nursing, work resources 2001, Toker & Biron 2012). To this end, increased exhaus-
facilitate the renewal of personal vigour in the context of tion triggers a process leading to depression and health
persistent job stressors (Demerouti et al. 2001, Bakker complaints (Schaufeli & Bakker 2004, Ahola et al. 2005,
et al. 2004). Social support is an effective coping resource Ahola & Hakanen 2007, Hakanen et al. 2008):
that buffers the adverse effects of high job stressors, e.g.
Hypothesis 1: Nurses’ emotional exhaustion at work is positively
work overload (Johnson & Hall 1988, Karasek & Theorell
related to depressive state.
1990). Higher levels of support were related to lower emo-
tional exhaustion in nurses and had a buffering effect on Previous studies showed that support at work is associated
the relationship between work stressors and exhaustion with a lower risk for depression (for an overview, see Bonde
(Jenkins & Elliott 2004). Evidence on supervisor behaviour 2008, Netterstrom et al. 2008). Supervisor support promotes
and burnout in nursing is inconsistent. For example, Stor- employees’ positive sense of self and thus leads to self-efficacy
deur et al. (2001) reported negative associations of active and increased well-being (Bandura 1978, Cohen & Wills
management by exceptions and emotional exhaustion of 1985). Nurses develop a feeling of mastery and improved
nurses. However, no relationships between transformational self-efficacy if supervisors offer means to reduce work over-
and contingent reward leadership behaviours and nurses’ load and, thereby, to successfully achieve performance goals
exhaustion could be found in their study. under intense work demands (Bandura 1978).
To this end, our study sought to answer one of the core
Hypothesis 2: Supervisor support is negatively associated with
questions regarding the burnout-depression-link. Namely,
nurses’ depressive state.
how nurses can benefit from work resources in high strain
environments. Work overload increases the risk for exceeding individual
energetic and affective resources (Shirom et al. 2010, Alar-
con 2011). Nurses are often occupied with additional
The study
work-related duties, meaning limited opportunities exist to
Drawing on two consecutive surveys in nursing profession- recover from work. With greater fatigue, individual reserve
als we aimed to examine the role of work conditions for capacity at work declines, individual functioning is reduced
the burnout–depression relationship. Specifically, we and negative affective states and depressive thoughts occur
focused on work overload and supervisor support as two (Skapinakis et al. 2004). Eventually, nurses in overload
fundamental job conditions. Work overload is a prevalent environments who strive to maintain regulatory energetic
stressor for healthcare professionals, which is associated resources face an increased risk of depleting individual
with depression (e.g. Rau et al. 2010, Shirom et al. 2010). resources (Hobfoll & Wells 1998).
Supervisor support is crucial for social recognition, perfor-
mance feedback and development of competencies. And it Hypothesis 3: Work overload amplifies the emotional exhaustion–
exerts also an important influence on the development of depressive state association, such that the association is stronger
depressive states (Cohen & Wills 1985). for nurses with high work overload.

1776 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

We further assume that supervisor support acts as an Design

effective buffer for the emotional exhaustion–depression
We conducted two consecutive surveys where hospital
relationship. As mentioned above, several studies show that
nurses and day care nursing professionals were surveyed
lack of support at work increases the risk for depression
(both data collections were finished in 2010). The purpose
(for an overview, Netterstrom et al. 2008). Supervisor sup-
of a two sample design was to enhance the validity of our
port can be considered a key source for resource investment
findings by testing the generalizability of our results across
and protecting mechanism against resource loss. Supervisor
different nursing settings. This strengthens the comparative
support promotes a positive sense of self and appreciation
potential since both samples include nursing professionals
and yields additional social resources that counteract ener-
with similar professional backgrounds. To address the limi-
getic depletion and loss of regulatory resources (Hobfoll &
tations of survey 1 (i.e. one institution, small sample size,
Wells 1998).
inpatient care), to replicate findings and to test the external
Hypothesis 4: Supervisor support moderates the association validity in another nursing setting, we conducted a second
between emotional exhaustion and depressive state, such that the survey among day care professionals. However, both sam-
association is stronger for nurses with low supervisor support. ples were not analysed as one group because we assumed
that differences between both nursing work settings may
Finally, we assume that emotional exhaustion, work over-
confound the results (i.e. different organizational structures
load and supervisor support exhibit a three-way interaction
and patient characteristics). Furthermore, we applied differ-
on depressive state. We argue that supervisor support is espe-
ent measures for supervisor support. This increases the gen-
cially beneficial for exhausted employees with high work
eralizability of findings but restricts the opportunity to
overload, as supervisor support helps to counteract employ-
collate the two samples into one (i.e. varying psychometric
ees’ depletion of individual regulatory resources. We expect
properties may confound the reliability of the results).
supportive leadership to provide a resourceful environment
that fosters nurses’ well-being and relieves the adverse effect
of high work stressors. This rationale is in line with the Job– Sample and participants
Demand–Control–(Support) model, as social support is
Survey 1: Hospital nurses working in a mid-size community
assumed to buffer and to offset the detrimental effects of job
hospital in South Germany were examined. Of 269 question-
demands (buffer hypothesis) (Johnson & Hall 1988, Johnson
naires distributed, 115 were returned (42
75% response).
et al. 1989, Karasek & Theorell 1990).
Excluding cases with missing data in the study variables, we
Further justification is provided by the Conservation of
obtained N = 111 complete data sets (females: N = 96,
Resources approach, as highly exhausted nurses possess
86
5%). Mean age was 40
41 years (SD 9
65 years, range 23-
lower regulatory resources and abilities to cope with work
62 years) and mean tenure in current job was 14
06 years (SD
stressors. In chronic stressful conditions, nurses might be at
8
17, range 1-37 years). N = 57 nurses (51
4%) worked full-
increased risk of losing valuable resources (Hobfoll & Wells
time and 6
3% (N = 7) reported having a temporary con-
1998, Hobfoll 2011). As work stressors are highly corre-
tract. The majority were certified nurses (N = 104; 94
5%)
lated with emotional exhaustion (Alarcon 2011), supervisor
and N = 28 participants were working in a leadership posi-
support may prevent resource loss and promote energetic
tion (25
5%). Nurses worked in various areas: N = 67 in
resources and, thus, breaking or temporarily interrupting
inpatient wards (61
5%), N = 14 in the intensive care unit
the downward spiral (Hobfoll 2001). This should be partic-
(12
8%), N = 9 in the operating theatre (8
3%) and N = 19
ularly pronounced for nurses in high demanding work envi-
in other staff functions (radiology, emergency, and admission
ronments. Supervisor support promotes self-esteem and
and administration; 17
4%). Above reported sociodemo-
appreciation and results in further personal resources (i.e.
graphic characteristics did not differ between the different
self-efficacy and sense of belonging) that assist nurses in
areas of work (results not reported).
coping with work-related demands and further enhance
Survey 2: Nurses working in 145 day care homes were
well-being (Cohen & Wills 1985).
initially contacted. Out of 80 day care facilities (with over-
Hypothesis 5: Emotional exhaustion, work overload and supervisor all 307 nurses) which agreed to join, N = 202 nurses
support exhibit a three-way interaction on depressive state, such responded (response rate 65
8%); 93
1% (N = 188) were
that the negative relationship between emotional exhaustion and female and mean age was 46
1 years (SD 10
76, range 17-
depressive state is strongest for nurses with high work overload 65 years). Mean job tenure was 5
18 years (SD 4
89, range
and low supervisor support. 04-25
7 years); mean professional tenure was 14
5 years (SD

© 2016 John Wiley & Sons Ltd 1777

M. Weigl et al.

10
3, range 0
4-10
31); 40
1% of the surveyed nurses were Supervisor support
working part time (<20 h/w). In survey 1, we applied the German version of the widely
used LMX-7 scale (Scandura & Graen 1984, Schyns et al.
2005). This 7-item measure describes the perceived nature
Data collection
of the supervisor–subordinate exchange relationship. Sam-
Survey 1’s questionnaires and letters of informed consent ple items are: ‘How well does your leader understand your
were sent to nurses via the internal mail. Completed and job problems and needs?’ ‘What are the chances that your
anonymized questionnaires as well as signed consent forms leader would “bail you out” at his or her expense?’ A 5-
were returned to the study team. Survey 2’s questionnaires point scale was used (e.g. 1 = not a bit; 5 = to a great
with pre-stamped return envelopes were sent to day care extent). In survey 2, it was assessed by the scale Social Sup-
homes willing to participate in the study. Completed ques- port from the Supervisor of the Copenhagen Psychosocial
tionnaires were sent back to the study team and contained Questionnaire (Pejtersen et al. 2010). It consists of two
the following scales: items (‘How often do you get help and support from your
immediate superior? and How often is your superior willing
Depressive state symptoms to listen to your work-related problems?’). Answers ranged
A subscale of the German version of Spielberger’s State-Trait from 1 = never/hardly ever-5 = always.
Depression Scales (Form X-1) was used to measure depressive
symptoms (STDS; Spielberger 1995, Spaderna et al. 2002).
Control variables
This five-item scale assesses cognitive-affective symptoms of
Participants reported their age (in years), gender (1 = male,
depression as a negative emotional state (dysthymic affective
2 = female), type of employment (1 = full-time, 2 = part-
state). The scale has been well validated (Krohne et al. 2002,
time) and working in a supervisor position (1 = yes,
Spaderna et al. 2002) and successfully applied to measure the
2 = no). To control for potential effects of general well-
impact of job conditions on depression in healthcare workers
being, we applied two different measures of general health
(Weigl et al. 2012). This scale shows good convergent valid-
status: In survey 1, a scale subsuming two items with a 4-
ity with alternative self-report measures of depression (Beck
point response range was applied (‘I feel healthy’; ‘I feel
Depression Inventory: r = 0
63/0
58; Zung Self-Rating
well’; 1 = not at all, 4 = yes, very much). Both items origi-
Depression Scale: r = 0
60/0
66; Center for Epidemiologic
nate from the above-described STDS and cover euthymic
Studies Depression Scale: r = 0
71/0
62). The five items cap-
affect (STDS; Spielberger 1995, Spaderna et al. 2002). In
ture depressed affect or dysthymia through descriptions such
survey 2, the mental component subscale (MCS) of the 12-
as ‘blue’, ‘miserable’, ‘downhearted’, ‘sad’ and ‘gloomy’.
Item Short-Form Health Survey (Ware et al. 1996) was
Nurses indicated how they felt on a four-point intensity scale
used in a German-validated version (Bullinger & Kirch-
(1 = not at all-4 = very much so).
berger 1998). It subsumes six items asking for perceived
levels of energy, accomplishments or being peaceful during
Emotional exhaustion (Burnout)
the past 4 months. A higher overall MCS score indicates
The 9-item measure from the German version of the
better mental well-being and quality of life.
Maslach Burnout Inventory was used to measure emotional
exhaustion (B€ussing & Perrar 1992). Item examples are: ‘I
feel emotionally drained from my work’ or ‘I feel burned Ethical considerations
out from my work’. A five-point frequency scale was
applied (1 = never/occasionally-5 = very often). Survey 1 was approved by the Ethics committee of the Fac-
ulty of Medicine, Munich University (No. 016/04). For sur-
Work overload vey 2, Research Ethics Committee approval was not
Nurses reported work overload using a three-item measure required. All local workers councils and administrations
from a well-established questionnaire for work analysis in were sought for agreement and approval prior to start.
hospitals (B€
ussing & Glaser 2002, Weigl et al. 2010). Item
wordings were ‘Even in a hurry I often cannot finish all my
Data analyses for surveys 1 and 2
work’, ‘I often have too much work to do all at once’ and
‘I often have to work under pressure to meet short-term To establish the factorial validity of the scales, confirmatory
deadlines’. A 5-point scale was applied (1 = not at all- factor analyses (CFA) were conducted with AMOS 19
0
5 = to a very great extent). (maximum-likelihood estimation). We applied accepted

1778 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

goodness-of-fit indices and used conventional cut-offs (e.g.

comparison to (0)

4: General Well-Being; 3-factor model: Factor 1: Emotional Exhaustion, Depression; Factor 2: Work Overload, Supervisor Support; Factor 3: General Well-Being; 2-factor model: Fac-
N = 111 (survey 1), N = 202 (survey 2). 4-factor model: Factor 1: Emotional Exhaustion, Factor 2: Depressive State Symptoms; Factor 3: Work Overload, Supervisor Support; Factor
(4)***
(6)***
(8)***
(9)***
Brown 2006): v2/d.f. should not exceed values of 2
0; incre-
mental fit index (IFI), Tucker Lewis index (TLI) and com-

Dv2 (d.f.)

164
85
675
83

1002
11
764
9
parative fit index (CFI) should be >0
90; and a root mean

Model
square error of approximation (RMSEA) should be <0
05.
Model comparisons used v2 statistics.

RMSEA
The results confirm the factorial validity of the scales.

0
096
0
137
0
142
0
156
0
08
Table 1 reports fit indices and alternative models, respec-
tively, for both samples. First, we tested the hypothesized

0
90

0
84
0
67
0
64
0
56
CFI
5-factor model that was composed of emotional exhaustion,
depressive state, work overload, supervisor support and

0
88

0
82
0
63
0
61
0
52
TLI
general well-being. The hypothesized structure showed sat-
isfactory fit to the data; model (0). Next, we compared this

0
90

0
84
0
67
0
64
0
57
five-factor model with four alternative models: one four-fac-

IFI
tor model that combined the two work-related scales under

Survey 2: Day care nursing

v2/d.f.
one common factor; one three-factor model that addition-

2
31

2
94
4
97
5
29
6
21
ally collapsed the emotional exhaustion and depression
items into one factor; a two-factor model consisting of a

243

247
249
251
252
d.f.
‘strain factor’ (emotional exhaustion, depression, well-
being) and a ‘work-conditions factor’ (work overload,

562
32

727
17
1238
15
1327
22
1564
43
supervisory support); and one general one-factor model.

tor 1: Emotional Exhaustion, Depression, General Well-Being; Factor 2: Work Overload, Supervisor Support.
v2
None of the four alternative models attained satisfactory fit
and all showed significantly higher v2-discrepancy than our

comparison to (0)

279
12 (4)***
476
4 (7)***
496
19 (9)***
1020
72 (13)***
proposed five-factor model. Accordingly, the hypothesized
factor structure was fully supported.
Dv2 (d.f.)
Hypotheses were tested using hierarchical moderated
Model

regression analyses with mean-centred predictor variables

(Cohen et al. 2003). In a step-wise procedure, estimations
Table 1 Discriminant validity results and goodness-of-fit indices for study scales.

RMSEA

for relationships between the predictor and outcome vari-

0
059

0
104
0
124
0
126
0
169
ables were computed. Additionally, interaction terms of the
standardized predictor variables were included in the multi-
0
91

0
76
0
66
0
65
0
37
variate equation to estimate potential incremental explana-
CFI

tion according to the hypothesized interactions among the

predictor variables (Cohen et al. 2003). Additionally, we
0
90

0
74
0
63
0
62
0
33
TLI

computed variance inflation factors (VIF) to test for poten-

tial degree of multi-collinearity among the predictors.
0
91

0
77
0
66
0
65
0
38
IFI

Although the appropriate value is subject of discussion, we

set the acceptable cut-off at VIF >4 (O’Brien 2007).
v2/d.f.
Survey 1: Hospital care

1
46

2
19
2
70
2
74
4
09

Descriptive and multivariate analyses were conducted with

SPSS 23.0.
367

371
374
376
380
d.f.

To avoid potential confounding influences, we controlled

for age, gender, leadership position, part-time work and
534
90

814
02
1011
30
1031
09
1555
62

general well-being. Age and gender were included to remain

consistent with the majority of studies on job-related deter-
v2

minants of depression (Ahola & Hakanen 2007, Bonde

(1) 4-factor model
(2) 3 factor model
(3) 2 factor model
(4) 1-factor model
(0) Hypothesized

2008). Leadership position was included because supervi-

5-factor Model

***P < 0
001.

sors are likely to be confronted with increased work over-

load and enhanced risk of impaired well-being. Part-time
Model

work was included as a potential confounder because it

reflects opportunities for recreation and recovery. Most

© 2016 John Wiley & Sons Ltd 1779

M. Weigl et al.

importantly, we controlled for general well-being because it

0
90/0
91

20 s correlations below the diagonal; Cronbach’s alpha values for scaled study variables in diagonal (survey 1/survey 2); (1) Cronbach’s alpha cannot be obtained for the mental compo-
measures: survey 1, scale range 1 = not at all, 4 = yes, very much; survey 2, SF-12 scale’s range 0-100. Intercorrelations of survey 1 variables are depicted above the diagonal, survey
N = 111 (survey 1), N = 202 (survey 2). Leadership Position 1 = yes, 2 = no; Sex 1 = male, 2 = female; Part-time work: 1 = full-time work, 2 = part-time work; General Well-Being
0
52**
0
50**
0
29**
0
29**
influences self-reported health in the workplace (de Lange

0
02
0
01
0
04
0
17
et al. 2005).

0
93/0
86
Results

0
30**

0
27**

0
21**
0
22*
0
06
0
03
0
10
0
07
Correlational results and test of hypotheses

0
86/0
86
Descriptive statistics and correlations of study variables are

0
25**
0
43**

0
32**
0
27**
shown in Table 2. In both surveys, emotional exhaustion

0
06
0
04
0
01
0
16
was highly associated to depressive state symptoms (survey

7
1: r = 0
50, P < 0
01; survey 2: r = 0
49, P < 0
01). Work

0
89/0
91
overload was positively associated with emotional exhaus-

0
38**

0
56**
0
27**
0
49**
tion (r = 0
43, P < 0
01; r = 0
56, P < 0
01) and depressive

0
10
0
04
0
05
0
19
state (r = 0
29, P < 0
01; r = 0
27, P < 0
01). Supervisor

6
support was negatively related to emotional exhaustion

0
67/-(1)
(r = 0
22, P < 0
01; r = 0
27, P < 0
01) and to depres-

0
59**
0
45**
0
32**
0
62**
0
05
0
08
0
03
0
14
sive state (r = 0
29, P < 0
01; r = 0
21, P < 0
01). Only
in survey 1, nurses part-time work tended to be associated

5
with less exhaustion (r = 0
19, p = 0
052). Age in survey

0
35**

0
31**
0
23*

0
01
0
01
0
08
0
03
0
04
2 was associated with increased emotional exhaustion
(r = 0
17, P < 0
05), increased work overload (r = 0
18,

4
P < 0
05) and less supervisor support (r = 0
21,

0
21**
P < 0
01). Moreover, female nurses in survey 2 experienced

0
15*
0
08
0
13

0
10
0
06
0
04

0
00
more emotional exhaustion (r = 0
14, P < 0
05) and higher

nent subscale of SF-12 because items are aggregated using an item-specific pre-defined weight.
3
work overload than males (r = 0
18, P < 0
01). Finally,

0
18**
nurses without a supervisor position in survey 2 reported

0
15*

0
14*
0
10

0
03

0
09

0
09
0
10
Table 2 Means, standard deviations, correlations (Pearson) among study variables.

higher supervisor support (r = 0
15, P < 0
05).

2

Results of regression analyses are presented in Table 3.

In regard to control variables only general well-being
0
36**

0
22**

0
21**
0
17*
0
18*
0
04

0
06

0
04
showed a strong significant association with depressive state
symptoms (survey 1: b = 0
51, P < 0
01; survey 2:
1

b = 0
59, P < 0
01). In the second step, we included emo-

10
76
0
25
0
49
0
49
9
27
0
95
0
92
1
00
0
47
Survey 2: Day

tional exhaustion, work overload and supervisor support.

care nursing

SD

Consistent with hypothesis 1, emotional exhaustion was

46
11
1
93
1
59
1
40
49
35
2
44
2
40
4
05
1
26

strongly associated with depressive state (survey 1:

M

b = 0
35, P < 0
01; survey 2: b = 0
29, P < 0
01). Hypoth-

esis 2 was not supported, as supervisor support was not
9
65
0
34
0
44
0
50
0
58
0
83
0
78
0
74
0
44
Hospital care

SD

directly related to nurses’ depressive state.

Survey 1:

In the third step, we added the three two-way interaction

40
41
1
86
1
75
1
49
3
21
2
87
3
07
3
57
1
24

terms. The cross-product term of emotional exhaustion and

M

work overload (Hypothesis 3) predicted depressive state

Depressive state symptoms

only in survey 1 (b = 0
18, P < 0
05). Figure 1 depicts this

Emotional Exhaustion

moderation effect: we found in hospital nursing that the

*P < 0
05; **P < 0
01.
Leadership Position

General Well-Being

Supervisor Support

association between emotional exhaustion and depressive

Work Overload
Part-time work

state was stronger for the subgroup in high work overload

Study variables

jobs. In survey 2, this interaction effect was not significant.

Again, Hypothesis 4 was partly confirmed: in survey 1 we
Gender
Age

found a significant interaction effect for emotional exhaus-

tion and supervisor support on depressive state (b = 0
17,
1
2
3
4
5
6
7
8
9

1780 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

P < 0
05). Figure 2 displays the interaction, which confirms with areas of work as a control variable but results did not
that the burnout – depressive state relationship was weaker change (results not reported).
in nurses who reported high supervisor support.
In both nursing samples, the three-way interaction
Discussion
(Hypothesis 5) had a significant effect on depressive state
(survey 1: b = 0
28, P < 0
01; survey 2: b = 0
19, Understanding the role of job demands and job resources in
P < 0
05). To facilitate its interpretation, we plotted the the development of depression, with the goals of ensuring
interaction for survey 1 in Figure 3a and for survey 2 in individual functioning at work, is crucial for health promo-
Figure 3b. We found consistently, that there was no associa- tion in nursing. Our surveys investigated the moderating
tion between exhaustion and depressive state for nurses with role of work overload (as a predominant job demand) and
low work overload and high supervisor support, whereas supervisor support (as an important job resource) on the
strong associations between emotional exhaustion and emotional exhaustion–depressive state relationship. Two
depressive state were consistently found for nurses with high consecutive surveys were conducted among hospital inpa-
work overload and low supervisor support. tient and day care nursing professionals.
To elucidate the interaction patterns, we tested for differ- We found a strong association between emotional
ences between the regression (Dawson & Richter 2006). In exhaustion and depressive state in both surveys. This is in
survey 1, we found a significant slope difference between line with previous research: nurses with increased states of
nurses with low overload and low support and nurses with exhaustion are at greater risk of depleting regulatory
low overload and high supervisor support (t = 2
19, resources and to develop depression (e.g. Ahola et al. 2005,
P < 0
05). Further slope differences were not significant; Ahola & Hakanen 2007, Hakanen et al. 2008, Peterson
nurses with high work overload/high supervisor support vs. et al. 2008). Due to the cross-sectional design of both sur-
nurses with low work overload/high supervisor support: veys, we cannot infer the direction of the relationship,
t = 1
94, P = 0
056; nurses with high work overload/low though there is strong rationale that emotional exhaustion
supervisor support vs. nurses with low work overload/high triggers a health impairment process, which leads to
supervisor support: t = 1
91, P = 0
059. The slope differ- depression and deteriorated mental health (Schaufeli &
ences in survey 2 were not fully consistent with results of Bakker 2004, Ahola et al. 2005, Ahola & Hakanen 2007,
survey 1. The slopes were different between nurses with Hakanen et al. 2008).
high work overload and low supervisor support vs. nurses Concerning the relationship between supervisor support
with low work overload and high supervisor support: and depressive state (Hypothesis 2), we found substantial
t = 2
63, P < 0
01. Furthermore, nurses with higher over- associations in both surveys, such that supervisor support
load and low supervisor support showed a significantly was associated with less depressive state (cf. Table 2). This
stronger emotional exhaustion – depressive state association confirms previous research on social support at work and
compared with nurses in low work overload and high depression (for an overview, see Bonde 2008, Netterstrom
supervisor support settings (t = 2
05, P < 0
05). Further et al. 2008). If nurses experience supervisor support, their
results were not significant: nurses with high work over- positive sense of mastery is enhanced, which in turn pro-
load/high supervisor support vs. nurses with high work motes well-being (Kuoppala et al. 2008). Supervisors who
overload/low supervisor support (t = 1
85, P = 0
066); offer support help employees to complete tasks and achieve
nurses with high work overload/high supervisor support vs. work-related goals (Bakker & Demerouti 2007). However,
nurses with low work overload/high supervisor support in the regression analyses, we identified no unique contribu-
(t = 1
68, P = 0
095). tion of supervisor support to depressive state when taking
In survey 1, we observed differences between areas of nurses emotional exhaustion into account.
work in work overload [F(d.f.) = 3
30(2), P = 0
04] and We assumed that work overload moderates the associa-
supervisor support [F(d.f.) = 6
17(2), P < 0
01). Nurses in tion between emotional exhaustion and depressive state
inpatient wards reported significantly less work overload symptoms such that the association is stronger for nurses
(mean = 2
94 SD 0
73) than nurses in other units, e.g. ICUs, with high work overload (cf. Hypothesis 3). However, only
ORs (mean = 3
43 SD 0
91). Supervisor support was signifi- in survey 1, this assumption was confirmed. One potential
cantly higher in inpatient settings (mean = 3
76 SD 0
70) explanation refers to the different nursing domains: hospital
than in ORs/ICUs (mean = 3
31 SD 0
56) and other areas of nurses reported higher work overload than nurses in day
work in hospital inpatient nursing (mean = 3
27 SD 0
72). care (cf. Table 3). The hospital environment is often char-
For survey 1 data, we repeated our multivariate analyses acterized by high workload and limited opportunities to

© 2016 John Wiley & Sons Ltd 1781

M. Weigl et al.

Table 3 Hierarchical linear regression analyses predicting depressive state in hospital nursing (survey 1) and day care nursing (survey 2).
Dependent Variable Depressive State Symptoms

Survey 1: Hospital care

Predictor
variable (Step) B SE b B SE b B SE b B SE

Control
variables
(Step 1)
Age 0
00 0
00 0
03 0
00 0
00 0
03 0
00 0
00 0
05 0
00 0
00
Gender 0
04 0
11 0
03 0
06 0
10 0
04 0
05 0
10 0
04 0
04 0
09
Leadership 0
09 0
09 0
10 0
02 0
08 0
03 0
02 0
08 0
02 0
00 0
07
position
0
01 0
09 0
01 0
01 0
08 0
01 0
02 0
08 0
02 0
02 0
08
Parttime work
General 0
22 0
04 0
51*** 0
15 0
04 0
35*** 0
15 0
04 0
34** 0
15 0
04
well-being
Main Effects
(Step 2)
Emotional 0
15 0
04 0
35*** 0
11 0
04 0
26*** 0
10 0
04
Exhaustion (EE)
Work 0
01 0
04 0
01 0
03 0
04 0
06 0
01 0
04
Overload
Supervisor 0
05 0
04 0
11 0
06 0
04 0
13 0
06 0
03
Support
Two-way
Interactions
(Step 3)
EE * Work 0
07 0
03 0
18* 0
03 0
03
Overload
EE * Supervisor 0
08 0
04 0
17* 0
05 0
04
Support
Work Overload * 0
06 0
04 0
13 0
04 0
03
Supervisor
Support
Three-way
Interaction
(Step 4)
EE * Work 0
03 0
01
Overload *
Supervisor
Support
DR 0
28*** 0
12*** 0
07** 0
06***
R² 0
28 0
40 0
47 0
53

*P < 0
05; **P < 0
01; ***P < 0
001.

N = 111 (survey 1), N = 202 (survey 2). All predictor variables were mean-centred.

recover and withdraw from the job (Shirom et al. 2010, leads to a greater risk for depressive symptoms (Hobfoll &
Alarcon 2011). Eventually, hospital nurses may experience Wells 1998).
greater fatigue, an antecedent of exhaustion, which nega- Likewise, Hypothesis 4 was not confirmed consistently.
tively affects their functioning at work (Skapinakis et al. For survey 1, the negative association between emotional
2004). Therefore, overtaxing hospital work may decrease exhaustion and depressive state was found to be stronger
the opportunity to maintain energetic resources and for hospital nurses with low supervisor support (cf. Fig-
increases the risk of depleting individual resources, which ure 2). Here, we can infer that supervisor support acts as a
meaningful buffer for the emotional exhaustion–depression

1782 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

Survey 2: Day care nursing

b B SE b B SE b B SE b B SE b

0
08 0
00 0
00 0
06 0
00 0
00 0
03 0
00 0
00 0
03 0
00 0
00 0
04
0
03 0
08 0
11 0
05 0
05 0
11 0
03 0
05 0
11 0
03 0
05 0
11 0
03
0
00 0
07 0
06 0
08 0
07 0
06 0
08 0
07 0
06 0
07 0
05 0
06 0
06

0
02 0
07 0
06 0
07 0
04 0
06 0
05 0
05 0
06 0
05 0
05 0
06 0
05

0
33*** 0
29 0
03 0
59*** 0
21 0
04 0
43*** 0
21 0
04 0
42*** 0
21 0
04 0
43***

0
24*** 0
14 0
04 0
29*** 0
13 0
04 0
27*** 0
13 0
04 0
27**

0
03 0
02 0
04 0
03 0
01 0
04 0
03 0
00 0
04 0
01

0
13 0
01 0
03 0
03 0
01 0
03 0
03 0
02 0
03 0
05

0
08 0
02 0
03 0
06 0
06 0
03 0
14

0
11 0
05 0
03 0
11 0
05 0
03 0
12

0
09 0
04 0
03 0
09 0
02 0
03 0
04

0
28*** 0
01 0
01 0
19*

0
36*** 0
05*** 0
01 0
02*

0
36 0
41 0
42 0
44

relationship. Specifically, in nurses with less vigour and contact with their superiors, which often receive HR-based
energy, supportive leaders can promote a positive sense of training in leadership and supervision. In day care nursing,
self-esteem and appreciation, which yields further personal professionals usually work in small units with only occa-
resources and counteracts energy depletion (Hobfoll & sional contact with their superiors. On the contrary, day
Wells 1998). Hypothesis 4 was, however, not confirmed in care nursing organizations in Germany are usually small
survey 2 among day care nursing professionals. Post hoc, and rarely provide programmes for leadership development.
we offer two explanations for these inconsistent effects. Second, it may be attributed to the nature or characteristics
First, in the hospital setting, nurses usually work in close of supervisor support at work. In survey 1, we applied the

© 2016 John Wiley & Sons Ltd 1783

M. Weigl et al.

2·5 (a)
2·4 (1) High work overload, high supervisor support
Low work overload
(2) High work overload, low supervisor support
High work overload 2·2
(3) Low work overload, high supervisor support
Depressive state

(4) Low work overload, low supervisor support

Depressive state
2 2

1·8

1·5 1·6

1·4

1·2
1
Low emotional exhaustion High emotional exhaustion 1
Low emotional exhaustion High emotional exhaustion
Figure 1 Moderation of the Relationship between Emotional
(b)
Exhaustion and Depressive State by Work Overload (Survey 1: 2·4 (1) High work overload, high supervisor support
Hospital care nursing). (2) High work overload, low supervisor support
2·2
(3) Low work overload, high supervisor support
(4) Low work overload, low supervisor support

Depressive state
2
2·5
Low supervisor support (LMX-7) 1·8
High supervisor support (LMX-7) 1·6
Depressive state

2 1·4

1·2

1·5 1
Low emotional exhaustion High emotional exhaustion

Figure 3 (a) Three-way interaction of the Relationship between

1 Emotional Exhaustion and Depressive State by Supervisor Support
Low emotional exhaustion High emotional exhaustion
and Work Overload in hospital care (Survey 1). (b) Three-way
Figure 2 Moderation of the Relationship between Emotional interaction of the Relationship between Emotional Exhaustion and
Exhaustion and Depressive State by Supervisor Support (Survey 1: Depressive State by Supervisor Support and Work Overload in Day
Hospital care). Care Nursing (Survey 2).

LMX-7-Scale that indicates the quality of the supervisor– that high levels of supportive job conditions have the poten-
employee relationship, whereas in survey 2, we used a mea- tial to cause negative effects through curvilinear associa-
sure that directly refers to supervisors’ supportive beha- tions, i.e. high supervisor support might be misperceived as
viours. Although both scales are well validated and distrust or surveillance by nurses if they work in jobs with
appropriate to address supervisor support, their slightly dif- low demands (Warr 1990). Emotional exhaustion and
ferent meanings and characteristics may be responsible for depression, on the other hand, were strongly related in
the varying effects. Potentially, the perceived leader-member nursing professionals who reported high work overload and
exchange (LMX) may more efficiently capture the overall low supervisor support. Although we applied different mea-
nature and impact of supervisors’ supportive behaviours. sures in two nursing samples, this three-way interaction
Hence, supervisor support is a rather heterogeneous con- effect could be consistently observed.
struct with various facets, including for example emotional Our results reveal that nurses in highly demanding envi-
support and empathy with subordinates as well as ronments with insufficient social resources – namely high
instrumental support for subordinates’ work situation. work overload and low supervisor support – are at greater
The main finding of both surveys is a meaningful three- risk to further deplete valuable resources, indicating a
way interaction between nurses’ emotional exhaustion, downward spiral of resource loss (Hobfoll & Wells 1998,
work overload and supervisor support in determining Hobfoll 2001). Burned-out professionals, who persistently
depressive state (Hypothesis 5, Figure 3a,b). Specifically, perceive low social resources at work, might fail to recover
there is no association between emotional exhaustion and and regain vigour at work. This puts them at greater risks
depression if nurses were working in jobs with low work of developing depressive symptoms, as they are incapable
overload and high supervisor support. This could be inter- of performing well, responding efficiently to high job
preted in the light of Warr’s ‘Vitamin Model’ which states demands and successively coping with high strain (Toker &

1784 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

Biron 2012). Supervisor support acts as a meaningful ent meanings in the respective item contents might limit
contextual resource that promotes new energetic resources their comparability. Nevertheless, our findings offer the
(Hobfoll 2001). Moreover, supervisor support serves to halt opportunity to extend the validity of our results beyond
the loss spiral and to enable high-strained professionals to one specific measure. Third, due to the cross-sectional
replenish their individual resources (Hobfoll 2011, Toker & design of both surveys, the directions of associations should
Biron 2012). It is important to note that the very nature of be interpreted carefully. We cannot assert causality since
the three-way interactions was not identical across both our design limits validity of causal inferences and increases
nursing settings (cf. Figure 3a,b). Nurses with high support- the risk of circular reasoning (i.e. nurses might be more
ive supervisors and high workload were expected to report negative in their ratings of work overload and supervisor
weaker associations between exhaustion and depression. support when they are feeling depressed). Although a broad
However, this was not consistently the case for our two body of research already exists, the conceptual and empiri-
samples. This is similar to reviews of the Job–Demand– cal distinctiveness of burnout and depression needs further
Control–(Support) model which found inconclusive evi- clarification (e.g. Bianchi et al. 2015). Prospective follow-
dence on this buffer hypothesis (van der Doef & Maes up studies are better suited to address causal associations of
1999, de Lange et al. 2003). work conditions, burnout and depression (Ahola & Haka-
In regard to Hypothesis 5, another finding deserves fur- nen 2007, Bonde 2008, Weigl et al. 2012, Theorell et al.
ther consideration. We found high associations between 2015). Fourth, we acknowledge that work conditions influ-
exhaustion and depressive state in nurses with low overload ence the onset of burnout (cf., Schaufeli & Enzmann 1998,
and low supervisor support, especially in survey 1 (hospital Demerouti et al. 2001, Maslach et al. 2001). Fifth,
nurses). Post hoc we assume that this effect is particularly concerning our depressive state measure, we emphasize that
triggered by reduced feelings of self-efficacy and mastery it indicates negative affect associated with depressive feel-
(Bandura 1978). Potentially, low work overload triggers ings, but does not serve as a clinical diagnosis for depres-
feelings of low accomplishment inherent to burnout and sion (Krohne et al. 2002, Lehr et al. 2008). Although we
depression. This perception is exacerbated if supervisor sup- attempted to establish convergence across both samples and
port is missing. Alternatively, one might argue that nurses assumed comparability between the nursing settings, we
with high over-commitment or Type-A behaviour do not concede that work environments of acute care and day care
seek for supervisor support (Siegrist et al. 2004). facilities may be quite different. Notwithstanding, we col-
Beyond the above-mentioned contributions of our study, lated the standardized reports of each survey into one sam-
two other findings deserve consideration. Our consistent ple and repeated the multivariate analyses again (including
association between work overload and depressive state all control variables and nursing setting). The results were
confirms previous findings (Rau et al. 2010, Shirom et al. consistent with our findings reported here: we observed a
2010). A further contribution is that our associations significant three-way interaction (b = 0
34, P < 0
01). This
between job-related determinants and individual outcomes corroborates to assume external validity of our findings
remain robust after adjusting for general well-being. across both settings. Our findings should be also replicated
in samples beyond hospital and day care settings (e.g. in
nursing homes for the elder). Moreover, we cannot exclude
Limitations and implications for theory and nursing
selection bias, i.e. healthy worker effect or biases due to
practice
voluntary participation. Finally, replication studies should
Our findings should be interpreted in the light of several draw on externally assessed psychosocial work-characteris-
limitations. First, there are individual and work-related pre- tics instead of self-reported ones, since high states of burn-
dictors of depression which have been neglected, e.g. family out or depression might deteriorate the perception of work
history of depression, personality and bullying or previous conditions (Rau et al. 2010).
serious life events (Bonde 2008, Netterstrom et al. 2008). Concerning potential implications, our study contributes
Second, we used self-reports which raises concerns for com- to the growing evidence base indicating that psychosocial
mon method bias. Based on our data, we assume that com- work conditions are associated with depressive symptoms
mon method bias is limited since our measures were (Bonde 2008, Theorell et al. 2015). Further theoretical con-
empirically distinct in the antecedent CFA. As already men- sideration of the buffering influence of the social environ-
tioned before, we used different measures in both surveys ment at work – specifically supervisor support – for nurses
to assess supervisor support and general well-being. Both with elevated risks for burnout and depression is necessary
measures are conceptually similar. However, slightly differ- (Tennant 2001). Since supervisor support is a critical

© 2016 John Wiley & Sons Ltd 1785

M. Weigl et al.

occupational resource, in-depth investigations into the

Funding
amplifying effects of specific behaviours of supervisors are
needed. Our results corroborate that highly exhausted The study was supported by the German Federal Institute for
nurses may lack personal capabilities to actively shape their Occupational Safety and Health (Project Number F 2224)
work environment for personal benefit, for example, and the Munich Centre of Health Sciences (MC-Health).
through reducing workload and seeking supervisor support
when they successively perceive less resources (Bandura
Conflict of interest
1978, Hobfoll 2011, Toker & Biron 2012).
In terms of practical implications, we advise The authors declare that they have no conflict of interest.
administrators to take into account that nursing supervisors
are decisive in limiting work overload and that their sup-
Author contributions
port exerts an important buffering function for depressive
state in high-strained nurses. Interventions to help All authors have agreed on the final version and meet at
exhausted nurses should aim to maintain individual func- least one of the following criteria [recommended by the
tioning through providing supervisor support and to reduce ICMJE (http://www.icmje.org/recommendations/)]:
workload. Thus, work redesign measures that address the
capabilities of supervisors in establishing a health promot- • substantial contributions to conception and design,
acquisition of data, or analysis and interpretation of
ing work environment should also seek to establish feasible
data;
workloads (Kuoppala et al. 2008, Skakon et al. 2010). To
complement these leadership approaches, supplementary • drafting the article or revising it critically for important
intellectual content.
organizational or work redesign interventions – e.g. by
means of health circles – that reduce work load and
enhance contextual resources should be implemented and References
evaluated (Aust & Ducki 2004).
Ahola K. & Hakanen J. (2007) Job strain, burnout and depressive
symptoms: a prospective study among dentists. Journal of
Affective Disorders 104(1–3), 103–110.
Conclusions Ahola K., Honkonen T., Isometsa E., Kalimo R., Nykyri E.,
Aromaa A. & Lonnqvist J. (2005) The relationship between job-
Findings from both surveys provide consistent evidence
related burnout and depressive disorders: results from the Finnish
that the interplay of supervisor support and work overload Health 2000 Study. Journal of Affective Disorders 88(1), 55–62.
influence the association of burnout and depression. We Alarcon G.M. (2011) A meta-analysis of burnout with job
found that the relationship between emotional exhaustion demands, resources and attitudes. Journal of Vocational
and depression was lowest for nurses in work environ- Behavior 79(2), 549–562.
APA (2013) Diagnostic and Statistical Manual of Mental Disorders
ments with low work overload and high supervisor sup-
(5th edn, text rev.), American Psychiatric Association,
port. According to our results, the primary objective
Washington, DC.
should be to constrain and reduce work overload in nurs- Aust B. & Ducki A. (2004) Comprehensive health promotion
ing professionals. Our study documents challenges associ- interventions at the workplace: experiences with health circles in
ated with facilitating an organizational environment that Germany. Journal of Occupational Health Psychology 9(3),
promotes well-being. The findings can be used to develop 258–70.
Bakker A.B. & Demerouti E. (2007) The Job Demands-Resources
interventions that preserve nurses’ well-being in terms of
model: state of the art. Journal of Managerial Psychology 22(3),
maintaining functioning and vitality when facing states of 309–328.
increased exhaustion. To extend our cross-sectional find- Bakker A.B., Demerouti E. & Verbeke W. (2004) Using the job
ings, follow up studies are needed to detect cause and demands-resources model to predict burnout and performance.
effect, as well as evaluation studies to compare Human Resource Management 43(1), 83–104.
Bandura A. (1978) The self system in reciprocal determinism.
randomized interventions.
American Psychologist 33, 334–358.
Bianchi R., Schonfeld I.S. & Laurent E. (2015) Burnout-depression
overlap: a review. Clinical Psychology Review 36, 28–41.
Acknowledgements Blackmore E.R., Stansfeld S.A., Weller I., Munce S., Zagorski B.M.
& Stewart D.E. (2007) Major depressive episodes and work
We thank Jake Burns for proof reading of the manuscript stress: results from a national population survey. American
draft. Journal of Public Health 97(11), 2088–2093.

1786 © 2016 John Wiley & Sons Ltd

JAN: ORIGINAL RESEARCH: EMPIRICAL RESEARCH – QUANTITATIVE Nurses’ work conditions, exhaustion and depression

Bonde J.P. (2008) Psychosocial factors at work and risk of Jenkins R. & Elliott P. (2004) Stressors, burnout and social
depression: a systematic review of the epidemiological evidence. support: nurses in acute mental health settings. Journal of
Occupational Environmental Medicine 65(7), 438–445. Advanced Nursing 48(6), 622–631.
Brown T.A. (2006) Confirmatory Factor Analysis for Applied Johnson J.V. & Hall E.M. (1988) Job strain, work place social
Research. Guilford Press, New York. support and cardiovascular disease: a cross-sectional study of a
Bullinger M. & Kirchberger I. (1998) SF-36: Fragebogen zum random sample of the Swedish working population. American
Gesundheitszustand. Hogrefe, G€ ottingen. Journal of Public Health 78(10), 1336–1342.
B€ussing A. & Glaser J. (2002) Work Analysis Instrument for Johnson J.V., Hall E.M. & Theorell T. (1989) Combined effects of
Hospitals – Self Report Version (TAA-KH-S). Hogrefe, G€ ottingen. job strain and social isolation on cardiovascular disease
B€ussing A. & Perrar K.M. (1992) Measurement of Burnout. The morbidity and mortality in a random sample of the Swedish
German Version of the Maslach Burnout Inventory (MBI-D). male working population. Scandinavian Journal of Work
Diagnostica 38, 328–353. Environment & Health 15(4), 271–279.
Cohen S. & Wills T.A. (1985) Stress, social support and the Karasek R. & Theorell T. (1990) Healthy Work: Stress,
buffering hypothesis. Psychological Bulletin 98(2), 310–357. Productivity and the Reconstruction of Working Life. Basic
Cohen J., Cohen P., West S.G. & Aiken L.S. (2003) Applied Books Inc, New York, NY.
Multiple Regression/Correlation Analysis for the Behavioral Kessler R.C., Akiskal H.S., Ames M., Birnbaum H., Greenberg P.,
Sciences. Earlbaum, Mahwah, NJ. Hirschfeld R.M., Jin R., Merikangas K.R., Simon G.E. & Wang
Dawson J.F. & Richter A.W. (2006) Probing three-way P.S. (2006) Prevalence and effects of mood disorders on work
interactions in moderated multiple regression: development and performance in a nationally representative sample of U.S.
application of a slope difference test. Journal of Applied workers. American Journal of Psychiatry 163(9), 1561–1568.
Psychology 91(4), 917–926. Krohne H.W., Schmukle S.C., Spaderna H. & Spielberger C.D.
de Lange A.H., Taris T.W., Kompier M.A., Houtman I.L. & (2002) The state-trait depression scales: an international
Bongers P.M. (2003) ‘The very best of the millennium’: comparison. Anxiety Stress and Coping 15(2), 105–122.
longitudinal research and the demand-control-(support) model. Kuoppala J., Lamminpaa A., Liira J. & Vainio H. (2008)
Journal of Occupational Health Psychology 8(4), 282–305. Leadership, job well-being and health effects – a systematic
de Lange A.H., Taris T.W., Kompier M.A., Houtman I.L. & Bongers review and a meta-analysis. Journal of Occupational and
P.M. (2005) Different mechanisms to explain the reversed effects Environmental Medicine 50(8), 904–915.
of mental health on work characteristics. Scandinavian Journal of Lehr D., Hillert A., Schmitz E. & Sosnowsky N. (2008) Assessing
Work Environment & Health 31(1), 3–14. depressive disorders using the Center for Epidemiologic Studies-
Demerouti E., Bakker A.B., Nachreiner F. & Schaufeli W.B. (2001) Depression Scale (CES-D) and State-Trait Depression Scales
The job demands-resources model of burnout. Journal of Applied (STDS-T): a comparative analysis of cut-off scores. Diagnostica
Psychology 86(3), 499–512. 54(2), 61–70.
Gartner F.R., Nieuwenhuijsen K., van Dijk F.J. & Sluiter J.K. (2010) Lerner D., Adler D.A., Rogers W.H., Chang H., Lapitsky L.,
The impact of common mental disorders on the work functioning McLaughlin T. & Reed J. (2010) Work performance of
of nurses and allied health professionals: a systematic review. employees with depression: the impact of work stressors.
International Journal for Nursing Studies 47(8), 1047–1061. American Journal of Health Promotion 24(3), 205–213.
Hakanen J.J., Schaufeli W.B. & Ahola K. (2008) The Job Letvak S., Ruhm C.J. & McCoy T. (2012) Depression in hospital-
Demands-Resources model: a three-year cross-lagged study of employed nurses. Clinical Nurse Specialist 26(3), 177–182.
burnout, depression, commitment and work engagement. Work Maslach C., Schaufeli W.B. & Leiter M.P. (2001) Job burnout.
and Stress 22(3), 224–241. Annual Review of Psychology 52, 397–422.
Hobfoll S.E. (2001) The influence of culture, community and the Netterstrom B., Conrad N., Bech P., Fink P., Olsen O., Rugulies R.
nested-self in the stress process: advancing Conservation of & Stansfeld S. (2008) The relation between work-related
Resources theory. Applied Psychology – An International Review psychosocial factors and the development of depression.
50(3), 337–370. Epidemiological Review 30(1), 118–132.
Hobfoll S.E. (2011) Conservation of resource caravans and O’Brien R.M. (2007) A caution regarding rules of thumb
engaged settings. Journal of Occupational and Organizational for variance inflation factors. Quality & Quantity 41(5), 673–
Psychology 84(1), 116–122. 690.
Hobfoll S.E. & Wells J.D. (1998) Conservation of resources, stress Pejtersen J.H., Kristensen T.S., Borg V. & Bjorner J.B. (2010) The
and aging: why do some slide and some spring? In Handbook of second version of the Copenhagen Psychosocial Questionnaire.
Aging and Mental Health: An Integrative Approach (Lomranz J., Scandinavian Journal of Public Health 38(3 suppl), 8–24.
ed.), Plenum Press, New York, NY, pp. 121–134. Peterson U., Demerouti E., Bergstrom G., Samuelsson M., Ake
Iacovides A., Fountoulakis K.N., Moysidou C. & Ierodiakonou C. M.A. & Nygren A. (2008) Burnout and physical and mental
(1999) Burnout in nursing staff: is there a relationship between health among Swedish healthcare workers. Journal of Advanced
depression and burnout? The International Journal of Psychiatry Nursing 62(1), 84–95.
in Medicine 29(4), 421–33. Rau R., Morling K. & Rosler U. (2010) Is there a relationship
Iacovides A., Fountoulakis K.N., Kaprinis S. & Kaprinis G. (2003) between major depression and both objectively assessed and
The relationship between job stress, burnout and clinical perceived demands and control? Work and Stress 24(1),
depression. Journal of Affective Disorders 75(3), 209–221. 88–106.

© 2016 John Wiley & Sons Ltd 1787

M. Weigl et al.

Roesler U., Jacobi F. & Rau R. (2006) Work and mental disorders Spaderna H., Schmukle S.C. & Krohne H.W. (2002) Report about
in a German national representative sample. Work and Stress 20 the German adaptation of the State-Trait Depression Scales
(3), 234–244. (STDS). Diagnostica 48(2), 80–89.
Scandura T.A. & Graen G.B. (1984) Moderating effects of Spielberger C.D. (1995) State-Trait Depression Scales (Form X-1).
initial leader–member exchange status on the effects of a Mind Garden, Palo Alto, CA.
leadership intervention. Journal of Applied Psychology 69(3), Stordeur S., D’Hoore W. & Vandenberghe C. (2001) Leadership,
428–436. organizational stress and emotional exhaustion among hospital
Schaufeli W.B. & Bakker A.B. (2004) Job demands, job resources nursing staff. Journal of Advanced Nursing 35(4), 533–542.
and their relationship with burnout and engagement: a multi- Tennant C. (2001) Work-related stress and depressive disorders.
sample study. Journal of Organizational Behavior 25(3), 293– Journal of Psychosomatic Research 51(5), 697–704.
315. Theorell T., Hammarstrom A., Aronsson G., Bendz L.T., Grape T.,
Schaufeli W.B. & Enzmann D. (1998) The Burnout Companion to Hogstedt C., Marteinsdottir I., Skoog I. & Hall C. (2015) A
Study and Practice: A Critical Analysis. Taylor & Francis, systematic review including meta-analysis of work environment
Washington, DC. and depressive symptoms. BMC Public Health 15(1), 738.
Schyns B., Paul T., Mohr G. & Blank H. (2005) Comparing Toker S. & Biron M. (2012) Job burnout and depression:
antecedents and consequences of leader – member exchange in a unraveling their temporal relationship and considering the role of
German working context to findings in the US. European Journal physical activity. Journal of Applied Psychology 97(3), 699–710.
of Work and Organizational Psychology 14(1), 1–22. van der Doef M. & Maes S. (1999) The Job Demand-Control(-
Shirom A. (2003) Job-related burnout. In Handbook of Support) Model and psychological well-being: a review of
Occupational Health Psychology (Quick J.C. & Tetrick L.E., 20 years of empirical research. Work and Stress 13(2), 87–114.
eds), APA, Washington, DC, pp. 245–264. Ware J. Jr, Kosinski M. & Keller S.D. (1996) A 12-Item Short-
Shirom A., Nirel N. & Vinokur A.D. (2010) Work hours and Form Health Survey: construction of scales and preliminary tests
caseload as predictors of physician burnout: the mediating effects of reliability and validity. Medical Care 34(3), 220–233.
by perceived workload and by autonomy. Applied Psychology 59 Warr P.B. (1990) Decision latitude, job demands and employee
(4), 539–565. well-being. Work and Stress 4(4), 285–294.
Siegrist J., Starke D., Chandola T., Godin I., Marmot M., Weigl M., Hornung S., Parker S.K., Petru R., Glaser J. & Angerer
Niedhammer I. & Peter R. (2004) The measurement of effort- P. (2010) Work engagement and accumulation of task, social
reward imbalance at work: European comparisons. Social and personal resources: a three-wave structural equation model.
Science and Medicine 58(8), 1483–99. Journal of Vocational Behavior 77(2), 338.
Skakon J., Nielsen K., Borg V. & Guzman J. (2010) Are leaders’ Weigl M., Hornung S., Petru R., Glaser J. & Angerer P. (2012)
well-being, behaviours and style associated with the affective Depressive symptoms in junior doctors: a follow-up study on
well-being of their employees? A systematic review of three work-related determinants. International Archives of
decades of research. Work and Stress 24(2), 107–139. Occupational and Environmental Health 85(5), 559–570.
Skapinakis P., Lewis G. & Mavreas V. (2004) Temporal relations Wieclaw J., Agerbo E., Mortensen P.B. & Bonde J.P. (2006) Risk
between unexplained fatigue and depression: longitudinal data of affective and stress related disorders among employees in
from an international study in primary care. Psychosomatic human service professions. Occupational and Environmental
Medicine 66(3), 330–335. Medicine 63(5), 314–319.

The Journal of Advanced Nursing (JAN) is an international, peer-reviewed, scientific journal. JAN contributes to the advancement of
evidence-based nursing, midwifery and health care by disseminating high quality research and scholarship of contemporary relevance
and with potential to advance knowledge for practice, education, management or policy. JAN publishes research reviews, original
research reports and methodological and theoretical papers.

For further information, please visit JAN on the Wiley Online Library website: www.wileyonlinelibrary.com/journal/jan

Reasons to publish your work in JAN:

• High-impact forum: the world’s most cited nursing journal, with an Impact Factor of 1·741 – ranked 8/109 in the 2014 ISI Jour-
nal Citation Reports © (Nursing (Social Science)).
• Most read nursing journal in the world: over 3 million articles downloaded online per year and accessible in over 10,000 libraries
worldwide (including over 3,500 in developing countries with free or low cost access).
• Fast and easy online submission: online submission at http://mc.manuscriptcentral.com/jan.
• Positive publishing experience: rapid double-blind peer review with constructive feedback.
• Rapid online publication in five weeks: average time from final manuscript arriving in production to online publication.
• Online Open: the option to pay to make your article freely and openly accessible to non-subscribers upon publication on Wiley
Online Library, as well as the option to deposit the article in your own or your funding agency’s preferred archive (e.g. PubMed).

1788 © 2016 John Wiley & Sons Ltd