Bauernhofer et al.

BMC Psychiatry (2018) 18:10

DOI 10.1186/s12888-018-1589-y

RESEARCH ARTICLE Open Access

Subtypes in clinical burnout patients

enrolled in an employee rehabilitation
program: differences in burnout profiles,
depression, and recovery/resources-stress
balance
Kathrin Bauernhofer1* , Daniela Bassa1, Markus Canazei2, Paulino Jiménez1, Manuela Paechter1, Ilona Papousek1,
Andreas Fink1 and Elisabeth M. Weiss1

Abstract
Background: Burnout is generally perceived a unified disorder with homogeneous symptomatology across people
(exhaustion, cynicism, and reduced professional efficacy). However, increasing evidence points to intra-individual
patterns of burnout symptoms in non-clinical samples such as students, athletes, healthy, and burned-out employees.
Different burnout subtypes might therefore exist. Yet, burnout subtypes based on burnout profiles have hardly been
explored in clinical patients, and the samples investigated in previous studies were rather heterogeneous including
patients with various physical, psychological, and social limitations, symptoms, and disabilities. Therefore, the aim of this
study is to explore burnout subtypes based on burnout profiles in clinically diagnosed burnout patients enrolled in an
employee rehabilitation program, and to investigate whether the subtypes differ in depression, recovery/resources-
stress balance, and sociodemographic characteristics.
Methods: One hundred three patients (66 women, 37 men) with a clinical burnout diagnosis, who were enrolled in
a 5 week employee rehabilitation program in two specialized psychosomatic clinics in Austria, completed a series of
questionnaires including the Maslach Burnout Inventory – General Survey (MBI-GS), the Beck Depression Inventory, and
the Recovery-Stress-Questionnaire for Work. Cluster analyses with the three MBI-GS subscales as clustering variables
were used to identify the burnout subtypes. Subsequent multivariate/univariate analysis of variance and Pearson
chi-square tests were performed to investigate differences in depression, recovery/resources-stress balance, and
sociodemographic characteristics.
Results: Three different burnout subtypes were discovered: the exhausted subtype, the exhausted/cynical subtype,
and the burned-out subtype. The burned-out subtype and the exhausted/cynical subtype showed both more severe
depression symptoms and a worse recovery/resources-stress balance than the exhausted subtype. Furthermore, the
burned-out subtype was more depressed than the exhausted/cynical subtype, but no difference was observed between
these two subtypes with regard to perceived stress, recovery, and resources. Sociodemographic characteristics were
not associated with the subtypes.
(Continued on next page)

* Correspondence: [email protected]
1
Department of Psychology, University of Graz, Universitätsplatz 2/DG, 8010
Graz, Austria
Full list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 2 of 13

(Continued from previous page)

Conclusions: The present study indicates that there are different subtypes in clinical burnout patients (exhausted,
exhausted/cynical, and burned-out), which might represent patients at different developmental stages in the burnout
cycle. Future studies need to replicate the current findings, investigate the stability of the symptom patterns, and
examine the efficacy of rehabilitation interventions in different subtypes.
Keywords: Burnout, Burnout subtypes, Burnout profiles, Person-oriented approach, Cluster analysis, Depression,
Stress-recovery

Background of the three dimensions of the Maslach Burnout

During the last decade, burnout has become a public Inventory (MBI) [28], were the burned-out profile and
health issue affecting between 4% and 7% of the working the healthy or engaged profile representing people with
population [1]. Nevertheless, the diagnosis is not yet in- severe or low symptomatology in all three burnout di-
cluded in clinical classification systems such as the mensions. Another common symptom profile was the
Diagnostic and Statistical Manual of Mental Disorders exhausted/cynical profile characterized by high exhaus-
(DSM-5) [2] or the International Statistical Classification tion and cynicism, but simultaneously high professional
of Diseases and Related Health Problems (ICD-10) [3]. efficacy. Typical incongruent scoring patterns included
To date, there is also no consensus on the definition the exhausted or overextended, the cynical or disengaged,
of burnout and its core symptoms [4–7], and the and the reduced professional efficacy profile representing
diagnosis overlaps tremendously in symptomatology people with rather mild symptomatology where only one
with other diagnoses, especially chronic fatigue [8] symptom is pronounced. However, one has to keep in
and depression [9–11]. mind that most previous studies have been conducted in
Despite some conceptual controversies [4–7, 12], the non-clinical samples, which consist of people who report
most widely accepted definition of burnout is the one by symptoms of a burnout but have not been diagnosed by
Maslach, Schaufeli, and Leiter [6], who defined it as a a psychiatrist or clinical psychologist and who are all still
work-related stress syndrome characterized by three working [13–15]. Therefore, the so-called healthy worker
main symptoms: exhaustion, cynicism, and reduced effect [29] occurred in these studies, and not all of the
professional efficacy. Exhaustion refers to feelings of de- prior discussed burnout profiles might genuinely repre-
pletion caused by various work demands; cynicism re- sent clinical burnout subtypes who need or seek
flects a distant attitude towards work and the people one treatment in specialized psychosomatic clinics or re-
is working with; and reduced professional efficacy habilitation programs.
represents the negative self-evaluation that one is in- To date, only one study [30] has investigated subtypes
competent and no longer able to perform work tasks based on burnout profiles in working-aged adults whose
adequately. However, past research has shown that burn- working ability was reduced or threatened by disease,
out manifests in different ways. A varying number of disability or some other disorders, and who were there-
burnout profiles [13–15] has been observed across stud- fore enrolled in an employee rehabilitation program.
ies, and multifaceted longitudinal trajectories of the This study has identified four of the above mentioned
burnout symptoms [16–21] indicate different pathways subtypes: two subtypes with severe symptoms, namely
into burnout. Different burnout subtypes might there- the burned-out profile and the exhausted/cynical profile,
fore exist. one subtype with milder symptoms, namely the reduced
The person-oriented approach to burnout [15, 22] is professional efficacy profile, and the healthy profile.
one possibility to investigate burnout subtypes. It identi- However, some of these profiles might have occurred be-
fies different groups in a population that are intraindivi- cause Hätinen et al. [30] explored a heterogeneous
dually homogenous, but interindividually heterogeneous patient sample suffering from various physical, psycho-
in various aspects such as symptomatology, psychopath- logical, and social limitations, symptoms, and disabilities.
ology, and job- and person-related factors. In past For that reason, the present study aims to examine
research, the person-oriented approach has been used to whether meaningful burnout subtypes may be found in a
identify typical burnout profiles [15], which might more homogeneous group of clinically diagnosed burn-
represent people at different developmental stages in the out patients treated in a specialized psychosomatic clinic
burnout cycle or truly different burnout subtypes [13, if the three main burnout symptoms suggested by
23–27]. Maslach, Schaufeli and Leiter [6] are considered and ex-
The most frequent burnout profiles that emerged in plored within a person-oriented approach [15].
these studies [13–15], predominantly from the analysis Additionally, we will investigate the relationship of these
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 3 of 13

subtypes to depression since an intense discussion has Inconsistencies in organizational risk factors have been
been going on during the last years [9–12, 31–54] found in burnout subtypes as well. According to the job
whether burnout can be seen as a distinct construct or demands-resources (JD-R) model [57], burnout develops
rather a new label of an already known state. when job demands (e.g., workload, time pressure, con-
A growing body of literature has examined the overlap flict) are high, while resources (e.g., autonomy, social
between burnout (subtypes) and depression-level, both support, positive relationship with supervisor) are lim-
in the working population [9, 12, 23, 31, 32, 39–48, 50– ited. Resources are therefore no longer able to buffer the
52] and in clinical patients [30, 54], but findings are het- negative impact of high demands on stress reactions
erogeneous and inconsistent. Several studies in the [58]. Job demands and resources have also been linked
working population reported strong correlations to specific burnout symptoms: exhaustion is caused by
between burnout and depression [9–11, 31, 32], espe- high workload and emotional demands, whereas
cially between exhaustion and depression [10, 31, 32], cynicism, reduced professional efficacy, and disengage-
while in clinically diagnosed burnout patients, the strong ment have been associated with a lack of resources [59–
exhaustion-depression overlap could not be replicated 61]. Previous studies on burnout subtypes in the work-
[55]. Low to moderate correlations between burnout and ing population [13, 23, 27] likewise found that workload
depression were also found by others in the working was high in subtypes experiencing high exhaustion
population [40, 41, 47], whereas Ahola et al. [45] (burned-out, exhausted/cynical, and exhausted), but re-
reported that burnout and depression overlapped par- sources were only low in subtypes with severe burnout
ticularly in severe burnout. These heterogeneous find- symptoms (burned-out and exhausted/cynical), particu-
ings may partly result from using different measures of larly in the burned-out subtype [13]. Yet, in clinical re-
burnout and depression [11, 37] with some measures habilitation patients, Hätinen et al. [30] did not find any
reflecting a larger concept redundancy between both dis- differences in job stressors and resources between three
orders (for a discussion on this topic, see Maslach & different burnout subtypes (burned-out, exhausted/cyn-
Leiter [37]). Furthermore, longitudinal studies have been ical, and low professional efficacy), although recovery
used to study the complex relationship between burnout was associated with a decrease in job demands and an
and depression, but yielded inconsistent results as well. increase in job resources [56].
Some authors reported reciprocal relations between Because most prior research was conducted within the
burnout and depression [42–44], while others found a scope of the JD-R model of burnout, recovery has re-
unidirectional relationship from burnout to depression ceived comparatively less attention in relevant research.
[46–50] or vice versa [51–53] or no predictive relation The recovery/resources-stress-balance model [62–64] is
[36]. Studies applying a person-oriented approach to similar to the JD-R model but focusses additionally on
explore the burnout-depression overlap [9, 31] detected recovery as crucial aspect to prevent burnout. According
that burnout and depression were not separable from to the model, burnout develops after prolonged periods
each other. Moreover, both disorders developed longitu- of stress without sufficient recovery and resources. More
dinally in tandem supporting the hypothesis that precisely, the homeostatic balance between stress and re-
burnout and depression may be the same disorder. In covery is impaired because resources that were depleted
clinical burnout patients, longitudinal studies are still during phases of stress are not adequately restored in
sparse [55]. Besides, only a few studies have explored the the recovery phase [63]. The role of recovery has, how-
burnout-depression overlap by applying a person- ever, not been systematically explored in burnout
oriented approach in clinical burnout patients [54] or subtypes, particularly not in clinical burnout patients
burnout rehabilitation clients [30, 56], respectively. Yet, enrolled in an employee rehabilitation program.
the overlap between burnout symptoms and depression Therefore, the current study aims to explore different
might differ between burnout subtypes. Boersma and burnout subtypes in clinical burnout patients, who are
Lindblom [23] found in the working population that sub- enrolled in an employee rehabilitation program in a
types experiencing high exhaustion were more depressed psychosomatic clinic, by performing cluster analysis with
than other burnout subtypes, and van Dam [54] could dis- the three subscales of the Maslach Burnout Inventory –
tinguish a group with mild symptoms from a group with General Survey (MBI-GS) [28] as clustering variables.
severe symptoms on several measures (burnout, depres- Furthermore, depression levels will be compared among
sion, anxiety, and fatigue) in clinically diagnosed burnout the subtypes, based on the expectation that the strength
patients. In contrast, Hätinen et al. [30] found in working- of the obviously existing overlap between burnout symp-
aged rehabilitation clients that all subtypes were equally toms and depression might differ between burnout
depressed, but in this study, a mixed patient sample suffer- subtypes. Finally, differences between the burnout
ing from various physiological, psychological, and social subtypes in the recovery/resources-stress balance and
limitations, symptoms, and disabilities was explored. sociodemographic characteristics will be explored.
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 4 of 13

Methods Depression
Study design and participants The German version of the Beck Depression Inventory
This study used a person-oriented approach [15] to ex- (BDI) was used to assess depression severity [67]. The
plore different subtypes based on burnout profiles in BDI is a 21-item self-rating questionnaire that covers a
clinical burnout patients. A total of 103 patients (64% variety of depressive symptoms along a continuum from
women) with a clinical burnout diagnosis were 0 (absent or mild) to 3 (severe) symptoms. The BDI is
recruited from two specialized psychosomatic clinics in widely used in treatment settings. A cut-off score ≥ 18
Austria. The patients were between 23 and 58 years old indicates a clinically relevant level of depression [68].
(M = 44.82 years, SD = 8.08) and had various educa- Additionally, all patients were screened for a past diag-
tional backgrounds: 31% had a university degree or the nosis of major depression using the Structured Clinical
A-level (high education), while 69% completed primary Interview for Axis I DSM-IV Disorders (SCID-I) [69].
education, an apprenticeship or some other type of
education or vocational training without the A-level
Recovery/resources-stress balance
(lower education).
The Recovery-Stress-Questionnaire for Work (RESTQ-
Work) is based on the recovery/resources-stress balance
Procedure model [63] and consists overall of 92 items [62]. On a
Burnout diagnosis was established by a team of psychia- 7-point frequency rating scale ranging from 0 (never) to
trists and clinical psychologists. Since separate diagnostic 6 (always), it measures the degree of stress and the
codes for burnout are not yet included in clinical classi- extent of recovery and resources in the past 7 days/
fication systems such as DSM-5 or ICD-10 [2, 3], the nights. The RESTQ-Work has seven subscales: social-
burnout diagnosis was based on the ICD-10 criteria of emotional stress (e.g., being mentally stressed, irritated,
work-related neurasthenia, which has been proposed as and frequently in arguments with others), performance
the psychiatric equivalent of clinical burnout [38]. Simi- (−related) stress (e.g., time pressure, interruptions at
lar to previous studies [54, 65], we included only patients work), loss of meaning/burnout (e.g., emotional exhaus-
that scored ≥2.20 on exhaustion and either ≥2.00 on tion, loss of control, meaninglessness), overall recovery
cynicism or ≤3.67 on professional efficacy in the (e.g., physical recovery, relaxation, satisfying sleep),
MBI-GS. Due to their burnout symptoms, all patients leisure/breaks (e.g., undisturbed leisure time without too
were enrolled in a 5 week employee rehabilitation many high-duty activities such as household chores, effi-
program at the clinics, and they were on sick leave for at cient breaks at work), psychosocial resources (e.g., social
least 1 week prior to study admission (M = 106.20 days, support from family, friends, and colleagues), and work-
SD = 113.46). Data were collected during the first week related resources (e.g., autonomy, participation, experi-
of the rehabilitation program and informed consent was ence of personal growth). The RESTQ-Work displays
obtained from all patients prior to participation. The good internal reliability and validity, and it has been
study was in accordance with the 1964 Declaration of used in burnout research before [58].
Helsinki and was approved by the ethics committee of
the University of Graz, Austria.
Statistical analyses
The burnout-depression overlap for the entire patient
Psychometric measures sample was analyzed using basic correlations and corre-
Burnout lations corrected for attenuation [70] between the three
Burnout symptom severity was assessed with the MBI-GS subscales and the BDI score. To explore differ-
German version [66] of the MBI-GS [28]. It has 16 items ent burnout subtypes, cluster analyses [71] were
and consists of three subscales: exhaustion, cynicism, performed. The three MBI-GS subscales (exhaustion,
and reduced professional efficacy. The exhaustion sub- cynicism, and professional efficacy) served as the cluster-
scale consists of five items (e.g., feeling emotionally ing variables and were standardized (z-score; M = 0, SD
drained from work), the cynicism subscale includes also = 1) prior to the analyses. First, a hierarchical agglomera-
five items (e.g., enthusiasm has decreased since work tive cluster analysis with Ward’s method as a linkage
was started), and the professional efficacy subscale con- method and squared Euclidean distance as a similarity
sists of six items (e.g., feeling that one gets done things measure was conducted. The optimal number of clusters
effectively). The items are answered on a frequency rat- was determined by the dendrogram, which is a visual
ing scale ranging from 0 (never) to 6 (daily). Since all illustration of how the individual cases are arranged into
participants were on sick leave, they were instructed to the clusters produced by hierarchical clustering. Next, a
respond to the items of the MBI-GS according to how K-means cluster analysis was conducted to improve the
they would feel if they were working at the moment. cluster fit [71]. To identify burnout subtypes based on
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 5 of 13

burnout profiles, the same approach has been applied by Burnout subtypes in clinically diagnosed burnout patients
others as well [23, 30]. Cluster analysis with the three MBI-GS subscales indi-
To describe the burnout subtypes further, one-way cated two to four burnout subtypes in the present
analyses of variance (ANOVAs) with the clusters as in- sample. In the two-cluster solution, cluster 1 and cluster
dependent variable and the three MBI-GS subscales as 2 grouped together although the patients of cluster 2
dependent variables were conducted. Differences in had more severe burnout symptoms than the patients of
depression between the subtypes were evaluated using cluster 1 (see Table 2). In the four-cluster solution, a
Pearson’s chi-square tests (for BDI ≥ 18 and depression cluster of patients with elevated burnout symptoms
diagnosis in the past) and a one-way ANOVA. Differ- emerged that lay in between cluster 1 and cluster 2.
ences between the subtypes in recovery/resources-stress However, since the four-cluster solution did not provide
balance were evaluated using a one-way MANOVA. In more conceptual clarity than the three-cluster solution,
the ANOVAs and MANOVA, the clusters served as the we considered the three-cluster solution representing
independent variable, while the BDI and the seven three different burnout subtypes the best. Subsequent
RESTQ-Work subscales were the dependent variables, ANOVAs showed how the three subtypes differed in
respectively. Sociodemographic differences between the burnout symptomatology (see Table 2).
subtypes were investigated with one-way ANOVAs (for Figure 1 displays the symptom profile of the burnout
age and days on sick leave) and Pearson’s chi-square tests subtypes graphically, both in z-scores and in mean
(for gender, education, and rehabilitation clinic). In case of scores. Each burnout subtype was described based on its
homogeneous variances, post-hoc comparisons were made mean score profile and based on which symptoms were
with the Bonferroni test; in case of inhomogeneous vari- most pronounced within the subtype. The three
ances, the Games-Howell test was used. Effect sizes are subtypes were labeled as follows: burned-out (subtype 3),
reported as partial eta-square (ηp 2) indicating small exhausted/cynical (subtype 2), and exhausted (subtype
(0.01 ≤ ηp 2 < 0.06), medium (0.06 ≤ ηp 2 < 0.14), and large 1). The burned-out subtype (n = 30) had the most severe
(ηp 2 ≥ 0.14) effects, respectively [72]. All statistical proce- burnout symptomatology. It displayed high exhaustion,
dures were calculated in SPSS 24 and were performed high cynicism, and low professional efficacy. The
with α = 0.05 (two-tailed). exhausted/cynical subtype (n = 39) had severe burnout
symptoms as well. It was characterized by high
Results exhaustion, elevated levels of cynicism, but simultan-
Burnout-depression overlap in the entire patient sample eously high professional efficacy. Finally, the exhausted
To explore the burnout-depression overlap in the entire subtype (n = 34) showed the least severe burnout symp-
patient sample, basic correlations and correlations cor- toms. It displayed elevated levels of exhaustion, the
rected for attenuation [70] between the MBI-GS lowest scores on cynicism, and high professional efficacy.
subscales and the BDI were inspected (see Table 1). The Based on the z-score profile, the three subtypes could
three burnout subscales correlated moderately with de- have been labeled as burned-out (subtype 3), exhausted
pression. The highest correlation emerged between cyni- (subtype 2), and healthy (subtype 1). However, we de-
cism and depression (r = 0.41; p < 0.01), the lowest cided to label the subtypes based on the means score
correlation occurred between exhaustion and depression profile because of the general higher symptom severity
(r = 0.30; p < 0.01). Regarding the correlations between in clinical burnout patients. Based on the z-score profile,
the three MBI-GS subscales, both exhaustion and cyni- particularly the label healthy for subtype 1 and the label
cism (r = 0.53; p < 0.01) and cynicism and professional exhausted for subtype 2 would have been misleading be-
efficacy (r = −0.51; p < 0.01) correlated strongly with each cause both subtypes had elevated levels of exhaustion
other, whereas a low correlation emerged between and subtype 2 had additionally elevated levels of cyni-
exhaustion and professional efficacy (r = −0.16; p = 0.10). cism. Hence, subtype 1 did not seem to be healthy and
subtype 2 did not seem to be solely exhausted.
Table 1 Cronbach’s alpha (α) and correlations (correlations
corrected for attenuation are in italics) between the MBI-GS Level of depression in the burnout subtypes
subscales and the BDI Depression score was highest in the burned-out subtype,
1. 2. 3. 4. α followed by the exhausted/cynical subtype, and the
1. MBI – exhaustion 1 0.53** −0.16 0.30** 0.86
exhausted subtype (see Table 2; Bonferroni-corrected
post-hoc tests: burned-out subtype vs. exhausted/cynical
2. MBI – cynicism 0.65** 1 −0.51** 0.41** 0.77
subtype, p = 0.02; burned-out subtype vs. exhausted
3. MBI – professional efficacy −0.20* −0.66** 1 −0.35** 0.78 subtype, p < 0.01; exhausted/cynical subtype vs. exhausted
4. BDI 0.35** 0.51** −0.42** 1 0.88 subtype, p = 0.05). Moreover, there were significant differ-
*p < 0.05; **p < 0.01 ences between the subtypes with regard to how many
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 6 of 13

Table 2 Differences between the burnout subtypes in burnout, depression, and recovery/resources-stress balance
subtype 1 subtype 2 subtype 3
overall (N = 103) exhausted (N = 34) exhausted/ cynical (N = 39) burned-out (N = 30)
M SD M SD M SD M SD F(2,100) ηp 2 contrasts
MBI
exhaustion 5.06 0.84 4.13 0.62 5.60 0.37 5.43 0.54 84.51** 0.63 3,2 > 1
cynicism 4.13 1.01 3.28 0.71 4.17 0.85 5.03 0.63 43.84** 0.47 3 > 2 > 1
professional efficacy 4.56 0.80 4.83 0.55 5.05 0.43 3.60 0.54 77.73** 0.61 1,2 > 3
BDI
depression 21.46 9.65 16.41 8.68 21.31 8.74 27.39 8.69 12.70** 0.20 3 > 2 > 1
RESTQ-Work
social-emotional stress 2.58 1.12 1.88 0.79 2.69 1.18 3.22 0.95 14.85** 0.23 3,2 > 1
performance (−related) stress 3.08 1.10 2.32 0.92 3.32 0.94 3.63 1.03 16.96** 0.25 3,2 > 1
loss of meaning/burnout 3.04 1.20 2.04 0.83 3.47 0.94 3.62 1.18 26.63** 0.35 3,2 > 1
overall recovery 2.12 0.81 2.65 0.77 1.98 0.74 1.70 0.61 15.27** 0.23 1 > 2,3
leisure/breaks 2.95 1.21 3.56 1.13 2.72 1.12 2.56 1.18 7.39** 0.13 1 > 2,3
psychosocial resources 2.88 1.28 3.34 0.98 2.75 1.37 2.51 1.34 3.84* 0.07 1 > 3
work-related resources 2.70 0.90 3.02 0.86 2.73 0.86 2.31 0.85 5.39** 0.10 1 > 3
*p < 0.05; **p < 0.01

patients scored above the BDI cut-off score ≥ 18 for (using Pillai’s trace, V = 0.48, F(14, 190) = 4.33, p < 0.01,
clinical depression (burned-out subtype: 27 patients η2p = 0.24). Subsequent ANOVAs indicated significant
(90%); exhausted/cynical subtype: 27 patients (69%); differences in all seven RESTQ-Work subscales (see
exhausted subtype: 14 patients (41%); χ2 (2, N = 103) Table 2). Post-hoc analyses further revealed that both
= 17.22, p < 0.01). Concerning previous depression diagno- the burned-out and the exhausted/cynical subtype
sis, no significant difference was observed between the experienced significantly more social-emotional stress
subtypes, χ2 (2, N = 103) = 0.77, p = 0.68, (see Table 3). (p < 0.01), performance(−related) stress (p < 0.01), and loss
of meaning/burnout (p < 0.01) as well as significantly less
Level of recovery/resources-stress balance in the burnout overall recovery (p < 0.01) and less leisure/breaks during
subtypes work (p < 0.01) than the exhausted subtype. Furthermore,
The MANOVA showed significant differences in recov- the burned-out subtype had significantly less psycho-
ery/resources-stress balance between the three subtypes social resources (p = 0.02) and less work-related

Fig. 1 Burnout profiles based on MBI z-scores and based on MBI mean scores. a subtype 1 = exhausted; subtype 2 = exhausted/cynical; subtype 3 = burned-
out. b post-hoc tests MBI-GS: exhaustion (burned-out subtype vs. exhausted subtype, p < 0.01; exhausted/cynical subtype vs. exhausted subtype, p < 0.01;
burned-out subtype vs. exhausted/cynical subtype, p = 0.30); cynicism (burned-out subtype vs. exhausted subtype, p < 0.01; exhausted/cynical subtype vs.
exhausted subtype, p < 0.01; burned-out subtype vs. exhausted/cynical, p < 0.01); professional efficacy (burned-out subtype vs. exhausted subtype, p < 0.01;
exhausted/cynical subtype vs. exhausted subtype, p = 0.20; burned-out subtype vs. exhausted/cynical, p < 0.01).
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 7 of 13

Table 3 Sociodemographic characteristics of the burnout subtypes

subtype 1 subtype 2 subtype 3
overall (N = 103) exhausted (N = 34) exhausted/ cynical (N = 39) burned-out (N = 30)
gender female 66 (64%) 21 (32%) 24 (36%) 21 (32%)
male 37 (36%) 13 (35%) 15 (41%) 09 (24%)
education high 32 (31%) 08 (25%) 12 (38%) 12 (38%)
lower 71 (69%) 26 (37%) 27 (38%) 18 (25%)
BDI ≥ 18 yes 68 (66%) 14 (21%) 27 (40%) 27 (40%)
no 35 (34%) 20 (57%) 12 (34%) 03 (9%)
depression diagnosis in the past yes 70 (68%) 25 (36%) 25 (36%) 20 (29%)
no 33 (32%) 09 (27%) 14 (42%) 10 (30%)
rehabilitation clinica A 51 (49%) 15 (29%) 21 (41%) 15 (29%)
B 52 (51%) 19 (37%) 18 (35%) 15 (29%)
age M (SD) 44.82 (8.08) 44.96 (9.72) 45.76 (6.26) 43.43 (8.22)
days sick leave M (SD) 106.20 (113.46) 73.65 (94.66) 115.56 (104.04) 130.93 (137.13)
a
data from two different rehabilitation clinics

resources (p < 0.01) than the exhausted subtype, while Dam [54] investigated subgroups in clinically diagnosed
the burned-out and the exhausted/cynical subtype did burnout patients by means of cluster analysis using
not differ from each other in any of the RESTQ- fatigue (CIS), depression (SCL-90-D), and anxiety
Work subscales. (SCL-90-A) as clustering variables and found two clus-
ters that differed from one another in terms of
Sociodemographic characteristics of the burnout symptom-severity on the three aforementioned mea-
subtypes sures. To date, only Hätinen et al. [30] have explored
Sociodemographic data for the burnout subtypes are pre- MBI burnout profiles in a mixed sample of working-
sented in Table 3. There were no significant differences aged rehabilitation clients, in which four different symp-
between the three subtypes with regard to sociodemo- tom patterns occurred: the burned-out profile, the
graphic characteristics. Men and women were equally exhausted/cynical profile, the reduced professional
distributed across the three clusters, χ2 (2, N = 103) = 0.65, efficacy profile, and the healthy profile. However, some
p = 0.72, and the subtypes did not differ in age, F(2, 100) of these profiles might have occurred because Hätinen et
= 0.71, p = 0.50, and education, χ2 (2, N = 103) = 2.02, al. [30] explored a heterogeneous patient sample suffer-
p = 0.36, respectively. The rehabilitation clinic was not ing from various physical, psychological, and social limi-
significantly associated with the burnout subtype, χ2 tations, symptoms, and disabilities. A study analyzing
(2, N = 103) = 0.69, p = 0.71, and there was no signifi- burnout subtypes in a group of clinically diagnosed
cant difference between the subtypes in how many burnout patients with the three MBI-GS dimensions
days the patients had already been on sick leave prior (exhaustion, cynicism, professional efficacy) as clustering
to study admission, F(2, 100) = 2.30, p = 0.10. variables, is to the very best of our knowledge not avail-
able yet. Cluster analysis with the three MBI-GS
Discussion subscales as clustering variables revealed three distinct
Burnout subtypes in clinical burnout patients burnout subtypes in the present study.
The present study explored different burnout subtypes In line with Hätinen et al. [30], we found two subtypes
based on MBI profiles in clinical burnout patients with severe burnout symptoms, namely the burned-out
enrolled in an employee rehabilitation program in a and the exhausted/cynical subtype, whereas the reduced
psychosomatic clinic. As increasing evidence points to professional efficacy subtype and the healthy burnout
intra-individual patterns of burnout symptoms in non- profile could not be replicated in the present study.
clinical samples such as students, athletes, healthy and Instead, we identified an exhausted subtype representing
burned-out employees [13, 15], this study addressed the patients with milder symptomatology. Sociodemographic
specific research question of whether different burnout characteristics were not systematically associated with
subtypes can also be identified in clinically diagnosed the subtypes in the present study, but all groups differed
burnout patients and whether these burnout subtypes in depression severity. The burned-out subtype showed
differ in depression, recovery/resources-stress balance, overall the highest level of depression, while the
and sociodemographic characteristics. Previously, van exhausted subtype was the least depressed. Moreover,
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 8 of 13

the burned-out and the exhausted/cynical subtype had a sample from an employee rehabilitation clinic suffering
worse recovery/resources-stress balance compared to from various physical, psychological, and social limita-
the exhausted subtype, but they did not differ from each tions, symptoms, and disabilities. Patients with physical
in perceived stress, recovery, and resources. Since impairments displayed particularly the healthy burnout
neither the depression scale, nor the RESTQ-Work sub- profile in their study, but physical impairments can lead to
scales were used for clustering, these differences further the experience of reduced professional efficacy as well,
support the validity of the cluster solution. Taken when one is no longer able to conduct one’s job properly.
together our findings suggest that different burnout sub- Regarding the possible antecedents of burnout, we
types can be identified in clinical patients who were revealed that the three burnout subtypes in our study
diagnosed with burnout by a professional team of psy- differed in their recovery/resources-stress balance. The
chiatrists and clinical psychologists working in a special- two subtypes with severe burnout symptoms, namely the
ized psychosomatic clinic. burned-out subtype and the exhausted/cynical subtype,
In line with past research [30, 54], our study indicates experienced more social-emotional stress, performan-
that the three subtypes represent patients at different ce(−related) stress, and loss of meaning in their job than
stages in the burnout cycle. The subtypes correspond the exhausted subtype. Both subtypes reported also the
well with the process model of burnout [18], according lowest overall recovery, leisure, and breaks during work.
to which burnout starts with exhaustion due to pro- Furthermore, the burned-out subtype tended to have less
longed periods of stress, followed by cynicism as an at- psychosocial and work-related resources than the
tempt to cope with the intense emotional strain, before exhausted subtype. According to the JD-R model [57]
finally reduced professional efficacy occurs because a and the recovery/resources-stress balance model [62–
negative attitude towards one’s job precludes achieving 64], burnout develops due to high job demands and sim-
work goals. Yet, it cannot be ruled out that the burned- ultaneously low resources and recovery. Both would be
out and the exhausted/cynical subtype might be truly necessary to buffer the negative effects of high job de-
different burnout subtypes, each characterizing the final mands on stress reactions [58]. Our findings are in line
stage in the burnout cycle. Several researchers have with these assumptions, which have also received sup-
shown that professional efficacy develops rather inde- port in burnout subtypes in the working population [13,
pendently of the other two burnout symptoms [21, 60], 23, 27] and in longitudinal studies in clinical rehabilita-
and both burnout profiles seem to be relatively stable tion clients [56]. Yet, the burned-out and the exhausted/
over time [23]. Moreover, an alternative longitudinal tra- cynical subtype did not differ in recovery/resources-
jectory of the burnout symptoms starting with reduced stress balance from each other in the present study. This
professional efficacy, followed by cynicism, and ending is in line with some previous studies [23, 30] but not
in emotional exhaustion, has been found as well [21]. with others [13] reporting that the burned-out subtype
The reduced professional efficacy subtype is a common had even less organizational resources than the
subtype in the general working population [15], and two exhausted/cynical subtype. To clarify, whether the
different pathways into burnout—one starting with exhausted/cynical subtype and the burned-out subtype
exhaustion, the other one starting with reduced profes- represent people at different developmental stages in the
sional efficacy—have been addressed [13, 15, 24]. burnout cycle or truly different burnout subtypes, future
Truly different burnout subtypes might also explain studies should explore longitudinally which characteris-
why Hätinen et al. [30] found the reduced professional tics differentiate between these two subtypes.
efficacy subtype, while the present study found the
exhausted subtype representing patients with milder Burnout-depression overlap
symptomatology, although the same methodology was During the last years, an intense discussion has been go-
applied in both studies, namely cluster analysis with the ing on whether burnout can be seen as a distinct con-
three MBI-GS subscales. Nevertheless, one has to keep struct or rather a new label of an already known state,
in mind that our study sample markedly differed from and much of this debate has focused around the
the rehabilitation clients analyzed in the study by burnout-depression overlap [9–12, 31–54]. Until now, a
Hätinen et al. [30]. The present study solely included vast majority of studies confirms that burnout and de-
clinical burnout patients diagnosed with the ICD-10 cri- pression are not independent and especially in severe
teria of work-related neurasthenia, which has been pro- burnout, both disorders strongly overlap [11, 32, 45, 54].
posed as the psychiatric equivalent of clinical burnout This was also demonstrated in our own data. In the
[38]. In addition, all patients scored high on exhaustion present study, 90% of the patients in the burned-out
(≥ 2.20) and either high on cynicism (≥ 2.00) or low on subtype scored above the BDI cut-off score ≥ 18 for clin-
professional efficacy (≤ 3.67) in the MBI-GS [54, 65]. In ical depression [68], whereas only 69% / 41% of the pa-
contrast, Hätinen et al. [30] explored a mixed patient tients reached this threshold in the exhausted/cynical
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 9 of 13

subtype and the exhausted subtype, respectively. reciprocal relation between burnout and depression, with
Similarly, Ahola et al. [45] showed that 90% of people each predicting subsequent developments in the other,
with severe burnout reported a physical or mental dis- while other authors [46–50] reported a predictive relation
ease; more specifically, pain and depression. In contrast, from burnout to depression. Furthermore, three longitu-
Hätinen et al. [30] could not replicate these findings; the dinal studies showed a unidirectional relationship from
authors found that several burnout subtypes in rehabili- depression to burnout [51–53] and one study failed to find
tation clients were equally depressed. Yet, in their study, any predictive relation [36]. Bianchi, Schonfeld and Laur-
a mixed sample of rehabilitation clients suffering from ent [31] focused on how burnout and depressive symp-
various physiological, psychological, and social limita- toms clustered at baseline and follow-up by applying a
tions, symptoms, and disorders was explored, while person-oriented approach. In this study, burnout was
other studies [23, 54] using a person-oriented approach measured with a global burnout index (the combination
to explore burnout subtypes found that subtypes with of the MBI subscales emotional exhaustion and cynicism);
severe burnout symptoms had higher depression scores depression was assessed with the Patient Health Question-
than subtypes with milder burnout symptoms, which is naire (PHQ-9). A similar person-oriented approach by
in line with our study. using cluster analysis to study the relationship between
Interestingly, our results differed also from previous burnout and depressive symptoms at baseline and over 7
findings with regard to the burnout-depression overlap years in dentists was applied by Ahola et al. [9]. Again,
when the correlations between the three MBI-GS burnout was measured with a global index (more
subscales and the BDI score were considered [9, 10, 31, specifically, a weighted sum score of the MBI dimen-
32]. For example, Bianchi et al. [10] found a high correl- sions was calculated, so that emotional exhaustion,
ation between the MBI sum score and the BDI (r = 0.68) depersonalization, and diminished personal accom-
in the working population. Importantly, in their study, plishment had different weights in the syndrome); de-
the correlation between emotional exhaustion and pressive symptoms were assessed using the short
depression was much stronger (r = 0.74) than the corre- form of the Beck Depression Inventory (BDI-SF). In
lations between the three dimensions of the MBI. both studies, similar findings were obtained as both
Similar to the study of van Dam [54], the present study studies could show that the participants formed three
could not replicate such a strong correlation between clusters with low, intermediate, or high levels of
exhaustion and depression (r = 0.30) in clinical burnout burnout and depressive symptoms at baseline. Add-
patients. For more than half a century [73], researchers itionally, in both studies long-term development oc-
have debated about the utility of using non-clinical sam- curred in tandem, either remaining stable or changing
ples such as students as analogues for patients with clin- in synchrony. Therefore, both studies support the hy-
ical diagnosis. Generalizing findings from non-clinical pothesis that burnout and depression overlap or may
burnout samples to clinical burnout patients might even be the same disorder.
therefore be problematic. However, another explanation Up to now, longitudinal studies in clinical burnout
for the lower correlation between exhaustion and de- patients are still sparse. Oosterholt et al. [55] examined
pression in the present study and in the study of van the course of cognitive performance and cortisol level in
Dam [54] might be that the range of exhaustion and de- clinical burnout patients and non-clinical burnout indi-
pression scores is limited in clinical burnout patients, viduals over a time period of 1.5 years. After 1.5 years,
which might have flattened the correlation between both clinical burnout patients showed improvement of burn-
constructs. Furthermore, some of the heterogeneity in out symptoms and general physical and psychological
study results concerning the burnout-depression overlap complaints, but the patients still reported subjective
may result from using different measures of burnout cognitive impairments that were also evident in a cogni-
and depression [11, 12, 31, 32, 37, 40, 41, 47]. Hence, tive test. In contrast, the non-clinical burnout group
the high correlations between burnout and depression, expressed the same elevated level of burnout symptoms,
especially with the burnout dimension emotional general physical and psychological complaints, and cog-
exhaustion in some studies, might partly reflect a large nitive problems, but did not show objective cognitive
level of concept redundancy between specific burnout deficits after 1.5 years.
and depression measures (for a discussion on this topic, Hätinen et al. [30] used a person-oriented approach to
please see Maslach & Leiter [37]). explore MBI burnout profiles and the changes in
Especially longitudinal studies provide the opportunity burnout 4 months after a rehabilitation intervention in a
to study the complex relationship between burnout and mixed sample of working-aged rehabilitation clients. In
depression. Longitudinal studies focusing on the question this study, four different burnout profiles (burned-out
whether burnout predicted depression or vice versa, re- profile, exhausted/cynical profile, reduced professional
vealed inconsistent results. Some authors [42–44] found a efficacy profile, and healthy profile) were found and the
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 10 of 13

burnout changes observed 4-month after the interven- was observed on cynicism or professional inefficacy.
tion depended on the burnout profile membership. Moreover, the reported effect sizes were overall rather
While exhaustion showed a decreasing trend in the small. Yet, most studies included in this meta-analysis
burned-out and exhausted/cynical profile, the decreasing investigated non-clinical burnout employees who do
trend in cynicism was only evident in the burned-out not necessarily possess a high level of burnout, since
profile compared to the other profiles. Furthermore, the interventions were aimed at decreasing the level
reduced professional efficacy showed an increasing trend of stress.
in the exhausted/cynical profile, whereas in the reduced Studies assessing the efficacy of interventions in burn-
professional efficacy profile the trend was decreasing. In out patients who are treated in clinical settings are rare
a subsequent study, Hätinen et al. [56] investigated [30, 55, 56]. Preliminary results suggest that in clinical
burnout trajectories in a one-year rehabilitation program burnout samples distinct symptom profiles exist and that
with a six-month follow-up and found inter-individual not all burnout profiles benefit equally from rehabilita-
changes in how clients reacted to the interventions and tion in terms of reduction in burnout symptoms, job-
how these reactions were manifested in burnout symp- related antecedents, and consequences [30, 56]. For the
toms during follow-up. The authors found three burnout burnout subtypes with severe symptomatology such as
patterns: two patterns with high levels of burnout and the exhausted/cynical and the burned-out subtype, the
one pattern with low levels of burnout. However, only in first step in rehabilitation should lie on the alleviation of
one group with high burnout levels, exhaustion and the burnout and depression symptoms (e.g., psychother-
cynicism were decreased at follow-up. This group was apy and if indicated antidepressants). Additionally,
labeled as “high burnout-benefited trajectory” by the rehabilitation should focus on organization-directed
authors. For the “low burnout trajectory” and the “high intervention activities in these subtypes, since burnout
burnout-benefited trajectory”, positive changes were patients with severe symptomatology reported worse
detected in job-related antecedents (e.g., time pressure interpersonal relations, poor workplace climate, and
at work, job control, workplace climate) and conse- dissatisfaction with supervision in previous studies [30].
quences (e.g. depressive symptoms, job satisfaction). Similarly, in our own study, the burned-out and the
Thus, preliminary results suggest that in clinical burnout exhausted/cynical subtype experienced high social-
samples distinct symptom profiles exist and that not all emotional stress, performance-related stress and limited
burnout profiles benefit equally from rehabilitation in (work-related) resources. There is still an ongoing debate
terms of reduction in burnout symptoms, job-related about the efficacy of antidepressants and psychother-
antecedents, and consequences [30, 56]. apies in the treatment of subclinical depression, which is
defined as a level of clinically relevant depressive symp-
Practical implications toms in the absence of a major depressive disorder [79,
Burnout prevention/intervention programs in non- 80]. Therefore, the exhausted subtype might profit from
clinical burnout employees are either person-directed a person-directed approach focusing on factors contrib-
(individual level interventions), organization-directed or uting to exhaustion and recovery (e.g., interventions tar-
a combination of both [74]. Person-directed interven- geting workload, relaxation, and work-life balance [27].
tions usually consist of different kinds of relaxation In sum, identifying different subtypes of clinical burnout
exercises and/or cognitive behavioral techniques to en- patients and analyzing the relationship of these subtypes
hance social support, job competence, and personal cop- to depression will enable psychiatrists and clinical
ing skills. Organization-directed interventions target on psychologists to focus their intervention activities more
poor workplace climate, dissatisfaction with supervision, effectively.
and changes in work procedures such as task restructur-
ing. Furthermore, organization-directed interventions Study limitations
aim at decreasing job demand, increasing job control Several limitations of the current study need to be
and expanding the level of participation in decision- addressed. Similar to previous studies [54, 65], in the
making. Several reviews in non-clinical populations [74– current paper the term clinical burnout is used to distin-
77] concluded that burnout intervention programs had a guish our study population of clinically diagnosed
beneficial effect on burnout outcome measures, espe- patients treated in a specialized psychosomatic clinic
cially when person-directed and organization-directed from non-clinical burnout groups who report symptoms
interventions were combined. Nonetheless, a recent of a burnout but are neither diagnosed by a psychiatrist
meta-analysis [78] on the effectiveness of controlled in- or clinical psychologist as such and who are all still
terventions to reduce burnout indicated statistically sig- working. To date, there are no consensual diagnostic
nificant effects of interventions only on general burnout criteria for the burnout syndrome and the disorder is
scales (d = 0.224) and exhaustion (d = 0.172); no effect not yet included in clinical classification systems such as
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 11 of 13

DSM-5 or ICD-10 [2, 3]. Therefore, in the current study, psychosomatic clinic. Our results provide evidence that
the burnout diagnosis was based on the ICD-10 criteria in clinical burnout patients three different burnout
of work-related neurasthenia, which has been proposed subtypes can be revealed: the exhausted subtype, the
as the psychiatric equivalent of clinical burnout [38]. exhausted/cynical subtype, and the burned-out subtype,
Additionally, we included only patients who scored which might represent patients at different developmen-
≥2.20 on exhaustion and either ≥2.00 on cynicism or tal stages in the burnout cycle. Furthermore, the
≤3.67 on professional efficacy in the MBI-GS. Hence, the exhausted subtype displayed the lowest depression level
findings of our study might not be generalizable to and the best recovery/resources-stress balance, while the
clinical burnout patients who are diagnosed differently. burned-out subtype exhibited the most severe depression
An extensive discussion about the problem of simplifica- symptoms and a worse recovery/resources-stress balance
tion burnout to ICD-10/DSM-5 codes such as neuras- than the exhausted subtype. Clearly, the current findings
thenia, somatic symptom disorder or exhaustion, can be need to be replicated in future studies involving larger,
found in a current review by Maslach and Leiter [37]. ideally longitudinal data sets to investigate the efficacy of
Furthermore, the patients in the present study were on rehabilitation interventions in different burnout subtypes
average three and a half months on sick leave, and and to track the stability and change patterns between
hence, a recall bias may have occurred; especially in the burnout subtypes over time.
patients who had been longer on sick leave. However,
Abbreviations
the burnout subtypes did not differ in how many days ANOVA: Analysis of variance; BDI: Beck depression inventory; BDI-SF: Beck
the patients had been on sick leave; thus, the distortion depression inventory – short form; CIS: Checklist individual strength; DSM-
bias was at least equally distributed between the sub- 5: Diagnostic and statistical manual of mental disorders; ICD-10: International
statistical classification of diseases and related health problems; JD-R: Job
types. Additionally, data were collected during the first demands-resources; MANOVA: Multivariate analysis of variance; MBI: Maslach
week of the rehabilitation program. Therefore, it cannot burnout inventory; MBI-GS: Maslach burnout inventory – general survey;
be ruled out that the burnout profiles we found occurred MINI: Mini-international neuropsychiatric interview; PHQ-9: Patient health
questionnaire; RESTQ-Work: Recovery-stress-questionnaire for work; SCID-
because some patients already started benefiting from I: Structured clinical interview for Axis I DSM-IV disorders; SCL-90-A: Symptom
the positive effects of the rehabilitation program before checklist – anxiety subscale; SCL-90-D: Symptom checklist – depression
they filled in the questionnaires. A further limitation of subscale

the present study is that the patients were not screened Acknowledgements
for comorbid psychiatric disorders using the Structured We would like to thank two rehabilitation clinics in Austria for support in this
Clinical Interview for Axis I DSM-IV Disorders (SCID-I) research by permitting access to their burnout patients.

[69] or the Mini-International Neuropsychiatric Funding

Interview (MINI) [81]. Therefore, we cannot ensure that This work was supported by the Austrian Research Promotion Agency (Grant
comorbid disorder are equally divided over the burnout No. 827511/BULT).

types. Finally, the analyses were based on cross-sectional Availability of data and materials
report questionnaires. Hence, we were not able to inves- The authors confirm that some access restrictions apply to the data underlying
tigate whether the exhausted/cynical and the burned-out the findings. The data set cannot be made publicly available because informed
consent from study participants did not cover public deposition of data.
subtype represent people at different developmental
stages in the burnout cycle or truly different burnout Authors’ contributions
subtypes. Due to the rather small sample size and differ- KB, EW, MC and DB contributed in the conceptualization of the study. DB
participated in collecting the data. KB conducted the statistical analyses,
ent interventions in the two rehabilitation clinics, we supported by MP and IP. KB drafted the manuscript, and EW, PJ, IP and AF
were also not able to investigate if the three burnout contributed to the interpretation of the data. All authors revised the
subtypes responded differently to therapy. manuscript and gave final approval of the version to be published.

Ethics approval and consent to participate

Conclusions The study was in accordance with the 1964 Declaration of Helsinki and was
In recent years, different burnout subtypes or profiles, approved by the ethics committee of the University of Graz, Austria. Informed
consent was obtained from all patients prior to participation. Data confidentiality
based on cluster/profile analysis with the three MBI was guaranteed and the patients were informed that they had the right to
subscales, have been proposed by several research withdraw from the study at any time.
groups [13, 15]. However, until now, intra-individual
Consent for publication
patterns of burnout symptoms have primarily been in- Not applicable
vestigated in non-clinical samples such as students,
athletes, healthy, and burned-out employees. Therefore, Competing interests
The authors declare that they have no competing interests.
the current study aimed to investigate whether different
burnout subtypes can also be identified in a more homo-
Publisher’s Note
geneous group of clinically diagnosed burnout patients Springer Nature remains neutral with regard to jurisdictional claims in
enrolled in an employee rehabilitation program in a published maps and institutional affiliations.
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 12 of 13

Author details 25. Innanen H, Tolvanen A, Salmela-Aro K. Burnout, work engagement and
1
Department of Psychology, University of Graz, Universitätsplatz 2/DG, 8010 workaholism among highly educated employees: profiles, antecedents and
Graz, Austria. 2Department of Psychology, University of Innsbruck, outcomes. Burnout Res. 2014;1:38–49.
Bruno-Sander-Haus Innrain 52f, 6020 Innsbruck, Austria. 26. Lee SM, Cho SH, Kissinger D, Ogle NT. A typology of burnout in professional
counselors. J Couns Dev. 2010;88:131–8.
Received: 2 June 2017 Accepted: 2 January 2018 27. Timms C, Brough P, Graham D. Burnt-out but engaged: the co-existence of
psychological burnout and engagement. J Educ Adm. 2012;50:327–45.
28. Maslach C, Jackson SE, Leiter M. Maslach burnout inventory. Manual. 3rd ed.
Palo Alto, CA: Consulting Psychologists Press; 1996.
References 29. Karasek R, Theorell T. Healthy work: stress, productivity and the
1. Schaufeli W. Past performance and future perspectives of burnout research. reconstruction of working life. New York: Basic Books; 1990.
SA J Ind Psychol. 2003;29:1–15. 30. Hätinen M, Kinnunen U, Pekkonen M, Aro A. Burnout patterns in
2. American Psychiatric Association. Diagnostic and statistical manual of mental rehabilitation: short-term changes in job conditions, personal resources, and
disorders. 5th ed. Arlington, VA: American Psychiatric Publishing; 2013. health. J Occup Health Psychol. 2004;9:220–37.
3. World Health Organization. The ICD-10 classification of mental and 31. Bianchi R, Schonfeld IS, Laurent E. Is burnout separable from depression in
Behavioural disorders: diagnostic criteria for research. Geneva: World Health cluster analysis? A longitudinal study. Soc Psychiatry Psychiatr Epidemiol.
Organization; 1993. 2015;50:1005–11.
4. Demerouti E, Bakker AB. The Oldenburg burnout inventory: a good 32. Schonfeld IS, Bianchi R. Burnout and depression: two entities or one? J Clin
alternative to measure burnout and engagement. In: Halbesleben JRB, Psychol. 2016;72:22–37.
editor. Handbook of stress and burnout in health care. Hauppauge, NY: 33. Bianchi R, Schonfeld IS, Laurent E. Physician burnout is better
Nova Science Publishers; 2008. p. 65–78. conceptualised as depression. Lancet. 2017;389:1397–8.
5. Kristensen TS, Borritz M, Villadsen E, Christensen KB. The Copenhagen 34. Bianchi R, Schonfeld IS, Laurent E. Burnout or depression: both individual
burnout inventory: a new tool for the assessment of burnout. Work Stress. and social issue. Lancet. 2017;390:230.
2005;19:192–207. 35. Bianchi R, Schonfeld IS, Vandel P, Laurent E. On the depressive nature of the
6. Maslach C, Schaufeli WB, Leiter MP. Job burnout. Annu Rev Psychol. “burnout syndrome”: a clarification. Eur Psychiatry. 2017;41:109–10.
2001;52:397–422. 36. Bianchi R, Schonfeld IS, Laurent E. Burnout does not help predict depression
7. Shirom A, Melamed S. A comparison of the construct validity of two among French schoolteachers. Scand J Work Environ Health. 2015b;41:565–
burnout measures in two groups of professionals. Int J Stress Manag. 2006; 8. https://doi.org/10.5271/sjweh.3522.
13:176–200. 37. Maslach C, Leiter MP. Understanding the burnout experience: recent research
8. Huibers MJ, Beurskens AJ, Prins JB, Kant IJ, Bazelmans E, van Schayck CP, and its implications for psychiatry. World Psychiatry. 2016;15:103–11.
Knottnerus JA, Bleijenberg G. Fatigue, burnout, and chronic fatigue 38. Schaufeli WB, Bakker AB, Hoogduin K, Schaap C, Kladler A. On the clinical
syndrome among employees on sick leave: do attributions make the validity of the Maslach burnout inventory and the burnout measure.
difference? Occup Environ Med. 2003;60:26–31. Psychol Health. 2001;16:565–82.
9. Ahola K, Hakanen J, Perhoniemi R, Mutanen P. Relationship between 39. Schaufeli WB, Enzmann D. The burnout companion to study and practice: a
burnout and depressive symptoms: a study using the person-centred critical analysis. London: Taylor and Francis; 1998.
approach. Burnout Res. 2014;1:29–37. 40. Leiter MP, Durup J. The discriminant validity of burnout and depression: a
10. Bianchi R, Boffy C, Hingray C, Truchot D, Laurent E. Comparative confirmatory factor analytic study. Anxiety Stress Coping. 1994;7:357–73.
symptomatology of burnout and depression. J Health Psychol. 2013;18:782–7. 41. Raedeke TD, Arce C, De Francisco C, Seoane G, Ferraces MJ. The construct
11. Bianchi R, Schonfeld IS, Laurent E. Burnout-depression overlap: a review. Clin validity of the Spanish version of the ABQ using a multi-trait/multi-method
Psychol Rev. 2015;36:28–41. approach. Anales de Psi-cologia. 2012;29:693–700.
12. Bianchi R, Schonfeld IS, Laurent E. Is burnout a depressive disorder? A 42. Ahola K, Hakanen J. Job strain, burnout, and depressive symptoms: a
reexamination with special focus on atypical depression. Int J Stress Manag. prospective study among dentists. J Affect Disord. 2007;104:103–10.
2014;21:307–24. 43. McKnight JD, Glass DC. Perceptions of control, burnout, and depressive
13. Leiter MP, Maslach C. Latent burnout profiles: a new approach to symptomatology: a replication and extension. J Consult Clin Psychol.
understanding the burnout experience. Burnout Res. 2016;3:89–100. 1995;63:490–4.
14. Maslach C, Leiter M. Early predictors of job burnout and engagement. J 44. Toker S, Biron M. Job burnout and depression: unraveling their temporal
Appl Psychol. 2008;93:498–512. relationship and considering the role of physical activity. J Appl Psychol.
15. Mäkikangas A, Kinnunen U. The person-oriented approach to burnout: a 2012;97:699–710.
systematic review. Burnout Res. 2016;3:11–23. 45. Ahola K, Honkonen T, Isometsä E, Kalimo R, Nykyri E, Aromaa A, Lönnqvist J. The
16. Golembiewski RT, Munzenrider RF, Stevenson JG. Phases of burnout: relationship between job-related burnout and depressive disorders—results from
developments in concepts and applications. New York: Praeger; 1986. the Finnish health 2000 study. J Affect Disord. 2005;88:55–62.
17. Friedman IA. Multiple pathways to burnout: cognitive and emotional 46. Armon G, Melamed S, Toker S, Berliner S, Shapira I. Joint effect of chronic
scenarios in teacher burnout. Anxiety Stress Coping. 1996;9:245–59. medical illness and burnout on depressive symptoms among employed
18. Leiter MP, Maslach C. The impact of interpersonal environment of burnout adults. Health Psychol. 2014;33:264–72.
and organizational commitment. J Organ Behav. 1988;9:297–308. 47. Hakanen JJ, Schaufeli WB. Do burnout and work engagement predict
19. Taris TW, Le Blanc PM, Schaufeli WB, Schreurs PJ. Are there causal depressive symptoms and life satisfaction? A three-wave seven-year
relationships between the dimensions of the Maslach burnout inventory? A prospective study. J Affect Disord. 2012;141:415–24.
review and two longitudinal studies. Work Stress. 2005;19:238–55. 48. Hakanen JJ, Schaufeli WB, Ahola K. The job demands–resources model: a
20. Te Brake H, Smits N, Wicherts JM, Gorter RC, Hoogstraten J. Burnout development three-year cross-lagged study of burnout, depression, commitment, and
among dentists: a longitudinal study. Eur J Oral Sci. 2008;116:545–51. work engagement. Work Stress. 2008;22:224–41.
21. van Dierendonck R, Schaufeli WB, Buunk BP. Toward a process model 49. Salmela-Aro K, Savolainen H, Holopainen L. Depressive symptoms and
of burnout: results from a secondary analysis. Eur J Work Organ Psy. school burnout during adolescence: evidence from two cross-lagged
2001;10:41–52. longitudinal studies. J Youth Adolesc. 2009;38:1316–27.
22. Bergman LR, Trost K. The person-oriented versus the variable-oriented 50. Shin H, Noh H, Jang Y, Park YM, Lee SM. A longitudinal examination of the
approach: are they complementary, opposites, or exploring different relationship between teacher burnout and depression. J Employ Couns.
worlds? Merrill Palmer Q. 2006;52:601–32. 2013;50:124–37.
23. Boersma K, Lindblom K. Stability and change in burnout profiles over time: 51. Armon G, Shirom A, Melamed S. The big five personality factors as predictors
a prospective study in the working population. Work Stress. 2009;23:264–83. of changes across time in burnout and its facets. J Pers. 2012;80:403–27.
24. Demerouti E, Verbeke WJ, Bakker AB. Exploring the relationship between a 52. Campbell J, Prochazka AV, Yamashita T, Gopal R. Predictors of persistent
multidimensional and multifaceted burnout concept and self-rated burnout in internal medicine residents: a prospective cohort study. Acad
performance. J Manage. 2005;31:186–209. Med. 2010;85:1630–4.
Bauernhofer et al. BMC Psychiatry (2018) 18:10 Page 13 of 13

53. Salmela-Aro K, Aunola K, Nurmi JE. Trajectories of depressive symptoms 77. Westermann C, Kozak A, Harling M, Nienhaus A. Burnout intervention
during emerging adulthood: antecedents and consequences. Eur J Dev studies for inpatient elderly care nursing staff: systematic literature review.
Psychol. 2008;5:439–65. Int J Nurs Stud. 2014;51:63–71.
54. van Dam A. Subgroup analysis in burnout: relations between fatigue, 78. Maricutoiu LP, Sava FA, Butta O. The effectiveness of controlled
anxiety, and depression. Front Psychol. 2016; doi: https://doi.org/10.3389/ interventions on employees’ burnout: a meta-analysis. J Occup Organ
fpsyg.2016.00090. Psychol. 2016;89:1–27.
55. Oosterholt BG, Maes JH, van der Linden D, Verbraak MJ, Kompier MA. 79. Cameron IM, Reid IC, MacGillivray SA. Efficacy and tolerability of
Getting better, but not well: a 1.5 year follow-up of cognitive antidepressants for sub-threshold depression and for mild major depressive
performance and cortisol levels in clinical and non-clinical burnout. Biol disorder. J Affect Disord. 2014;166:48–58.
Psychol. 2016;117:89–99. 80. Cuijpers P, Koole SL, van Dijke A, Roca M, Li J, Reynolds CF. Psychotherapy
56. Hätinen M, Kinnunen U, Mäkikangas A, Kalimo R, Tolvanen A, Pekkonen M. for subclinical depression: meta-analysis. Br J Psychiatry. 2014;205:268–74.
Burnout during a long-term rehabilitation: comparing low burnout, high 81. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E,
burnout – benefited and high burnout – not benefited trajectories. Anxiety Herqueta T, Baker R, Dunbar GC. The Mini-international neuropsychiatric
Stress Coping. 2009;22:341–60. interview (M.I.N.I.): the development and validation of a structured
57. Demerouti E, Bakker AB, Nachreiner F, Schaufeli WB. The job demands- diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry.
resources model of burnout. J Appl Psychol. 2001;86:499–512. 1998;59:22–33.
58. Jiménez P, Dunkl A. The buffering effect of workplace resources on the
relationship between the areas of worklife and burnout. Front Psychol.
2017; doi:https://doi.org/10.3389/fpsyg.2017.00012.
59. Bakker AB, Demerouti E, Taris T, Schaufeli WB, Schreurs PJ. A multigroup
analysis of the job demands-resources model in four home care
organizations. Int J Stress Manag. 2003;10:16–38.
60. Lee RT, Ashforth BE. A meta-analytic examination of the correlates of the
three dimensions of job burnout. J Appl Psychol. 1996;2:123–33.
61. Schaufeli WB, Bakker AB. Job demands, job resources, and their
relationship with burnout and engagement: a multi-sample study. J
Organ Behav. 2004;25:293–315.
62. Jiménez P, Dunkl A, Kallus KW. Recovery-stress questionnaire for work. In:
Kallus KW, Kellmann M, editors. The recovery-stress questionnaire. User
manual. Frankfurt/Main: Pearson; 2016. p. 158–87.
63. Kallus KW. Stress and recovery: an overview. In: Kallus KW, Kellmann M,
editors. The recovery-stress questionnaire. User manual. Frankfurt/Main:
Pearson; 2016. p. 27–48.
64. Kallus KW, Kellmann M. The recovery-stress questionnaire. User manual.
Frankfurt/Main: Pearson; 2016.
65. Oosterholt BG, van der Linden D, Maes JH, Verbraak MJ, Kompier MA.
Burned out cognition – cognitive functioning of burnout patients before
and after a period with psychological treatment. Scand J Work Environ
Health. 2012;38:358–69.
66. Büssing A, Glaser J. Managerial stress und burnout. A collaborative
international study (CISMS). Die deutsche Untersuchung. [managerial stress
and burnout. The German study]. Bericht nr. 44 aus dem Lehrstuhl für
Psychologie. München: Technische Universität, Lehrstuhl für Psychologie; 1998.
67. Hautzinger M, Bailer M, Worall H, Keller F. Beck-Depressions-Inventar (BDI).
Bearbeitung der deutschen Ausgabe. Testhandbuch. Bern, Göttingen,
Toronto, Seattle: Huber; 1994.
68. Beck AT, Steer RA, Carbin MG. Psychometric properties of the Beck
depression inventory: twenty-five years of evaluation. Clin Psychol Rev.
1988;8:77–100.
69. First MB, Spitzer RL, Gibbon M, Williams JB. User’s guide for the structured
clinical inter-view for DSM-IV interview for DSM-IV Axis I disorders.
Washington, DC: Ameri-can Psychiatric Press; 1997.
70. Cohen J, Cohen P, West SG, Aiken LS. Applied multiple regression/
correlation analysis for the behavioral sciences. 3rd ed. Mahwah, NJ:
Earlbaum; 2003.
71. Everitt BS, Landau S, Leese M, Stahl D. Cluster analysis. 5th ed. Chichester,
West Sussex: Wiley; 2011.
72. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Submit your next manuscript to BioMed Central
Hillsdale, NJ: Lawrence Erlbaum; 1988. and we will help you at every step:
73. Endler NS, Deisoff E, Rutherford A. Anxiety and depression: evidence for the
differentiation of commonly cooccurring constructs. J Psychopathol Behav • We accept pre-submission inquiries
Assess. 1998;20:149–71. • Our selector tool helps you to find the most relevant journal
74. Awa WL, Plaumann M, Walter U. Burnout prevention: a review of
• We provide round the clock customer support
intervention programs. Patient Educ Couns. 2010;78:184–90.
75. Ruotsalainen J, Serra C, Marine A, Verbeek J. Systematic review of • Convenient online submission
interventions for reducing occupational stress in health care workers. Scand • Thorough peer review
J Work Environ Health. 2008;34:169–78.
• Inclusion in PubMed and all major indexing services
76. Walter U, Krugmann CS, Plaumann M. Preventing burnout? A systematic
review of effectiveness of individual and comorbid approaches [German]. • Maximum visibility for your research
Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz. 2012;55:
172–82. https://doi.org/10.1007/s00103-011-1412-0. Submit your manuscript at
www.biomedcentral.com/submit