Content uploaded by Paul Eling
Author content
All content in this area was uploaded by Paul Eling on Feb 28, 2014
Content may be subject to copyright.
Work stress and attentional difficulties: An initial study on
burnout and cognitive failures
DIMITRI VAN DER LINDEN, GER P. J. KEIJSERS, PAUL ELING, &
RACHEL VAN SCHAIJK
Radboud University Nijmegen, The Netherlands
Abstract
Professional burnout is a stress-related disorder, having mental exhaustion due to work stress as
its most important characteristic. Burned out individuals also often complain about attentional
problems. However, it is currently not clear whether such complaints are based on true cognitive
deficits or whether they merely reflect the way burned out individuals rate their own cognitive
performance. To confirm the cognitive complaints we used a cognitive failure questionnaire (CFQ)
to assess the level of self-reported attentional difficulties in daily life. We also measured performance
on tasks of sustained attention and response inhibition (the SART and the Bourdon-Wiersma Test).
We compared three groups: (1) a group of ‘burned out’ individuals (n
/13) who stopped working
due to their symptoms and sought professional treatment; (2) teachers at a vocational training
institute (n
/16) who reported high levels of burnout symptoms but continued to work; and (3)
teachers from the same institute (n
/14) who reported no burnout symptoms. The level of burnout
symptoms was found to be significantly related to the number of cognitive failures in daily life, and to
inhibition errors and performance variability in the attentional tasks. To our knowledge, explicit tests
of objective cognitive deficits in burned out individuals have not been conducted before.
Consequently, this is the first study to indicate that burnout is associated with difficulties in voluntary
control over attention and that the level of such difficulties varies with the severity of burnout
symptoms.
Keywords: Burnout, work stress, attention deficits, executive control, cognitive failures
Introduction
Professional burnout is a stress-related disorder that is currently receiving much interest
from behavioural researchers and health specialists. Burnout refers to a set of symptoms
that an individual may develop during prolonged exposure to high levels of work stress and
that negatively affects mental and physical fitness, job satisfaction, and perceived
performance (Maslach & Schaufeli, 2000; Schaufeli & Enzmann, 1998). There are many
studies examining the antecedents and prevalence of burnout, showing that symptoms are
associated with high work demands, low levels of work autonomy, and lack of social support
(cf. de Lange, Taris, Kompier, Houtman, & Bongers, 2002; Taris, Schreurs, & Van Iersel-
van Silfhout, 2001; van der Doef & Maes, 1999). On the other hand, to our knowledge,
there are no studies directly addressing the relationship between the burnout syndrome
Correspondence: Dimitri van der Linden, Radboud University Nijmegen, PO Box 9104, 6500 HE Nijmegen, The
Netherlands. Tel:
/31 24 361 2743. Fax: /31 24 361 5937. E-mail: d.vanderlinden@psych.ru.nl
ISSN 0267-8373 print/ISSN 1464-5335 online # 2005 Taylor & Francis Group Ltd
DOI: 10.1080/02678370500065275
Work & Stress, January/March 2005; 19(1): 23 /36
and information processing. Yet, investigating this relationship may lead to additional
insight into the nature of burnout and its behavioural manifestations. Moreover, potentially
it may lead to new ideas for diagnosis and treatment.
The current paper represents an initial effort to study a specific information-processing
aspect of burnout, namely, deficits in attention. More specifically, individuals with burnout
symptoms often complain about difficulties in ‘keeping their mind’ on daily tasks (Maslach
& Schaufeli, 2000; Schaufeli & Enzmann, 1998). Such complaints are reported in clinical
observations but as far as we know, their validity has not yet been tested empirically.
Therefore, the aim of the current study was two-fold. First, we wanted to examine whether
burnout symptoms concur with increased self-reported cognitive failures in daily life; a high
incidence of these failures indicates attentional problems in performing daily tasks. Second,
we tested whether self-reported cognitive failures are associated with objective measures of
attentional deficits in a controlled laboratory setting.
Burnout and cognitive failures: An executive control issue?
The specific concentration problems, as reported by burned out individuals (e.g. inability to
concentrate on reading a newspaper, to keep one’s mind on a complex problem, or to focus
during a conversation) suggest that burnout is particularly accompanied with difficulties in
the voluntary or executive control over attention. Executive control is a term that refers to a
set of cognitive processes underlying voluntary and often effortful regulation of perception
and motor processes, in order to adaptively deal with novel or changing task demands
(cf. Miller & Cohen, 2001; Norman & Shallice, 1986). In general, the cognitive processes
subserving executive control are considered to be involved in the top-down regulation of
behaviour and can be differentiated from the more automatic cognitive processes that
regulate behaviour in a bottom-up way (Norman & Shallice, 1986; Rabbitt, 1997; Shiffrin
& Scheider, 1977).
Several studies have revealed that there are many different ways in which compromised
executive control can become apparent in behaviour. For example, people whose executive
control is impaired, may show deficits in working memory (Miyake et al., 2000), inability to
switch or maintain mental sets (De Jong, Berendsen, & Cools, 1999), problems with dual
task performance (Monsell & Driver, 2000), or response inhibition failures (Barkley, 1997).
There is no consensus yet about which of these deficits are central to executive control
impairment. However, it is widely acknowledged that sustained attention and response
inhibition are two major aspects of executive control (Kimberg & Farah, 1993; Manly,
Robertson, Galloway, & Hawkins, 1999). For example, people whose executive control is
impaired typically show deficits in their ability to sustain attention to ongoing thoughts or
behaviour and to inhibit a prepotent but inappropriate response (Gazzaniga, Ivry, &
Mangun, 1998). Therefore, we examined the relationship between burnout on the one
hand and sustained attention and response inhibition on the other. We expected that,
compared to no-burnout controls, individuals with burnout would perform poorly on
laboratory tasks of attention and inhibition. Moreover, as attentional problems often also
manifest themselves in performing daily tasks, we expected that declined performance in
the laboratory tasks would be related to the level of self-reported cognitive failures in daily
life.
Obviously, measuring these aspects of executive control does not cover the entire range of
possible manifestations. For example, it does not provide direct information about the
ability to maintain information in working memory or to perform multiple tasks at the same
24 D. van der Linden et al.
time. Nevertheless, deficits in attention and inhibition can be considered to be important
indications of compromised executive control (Miller & Cohen, 2001).
Investigating cognitive control processes in burnout is not only of theoretical interest; it is
widely acknowledged that compromised executive control may have far-fetching conse-
quences for almost every aspect of (working) life (Manly et al., 1999; Shallice & Burgess,
1991). For example, problems with sustained attention make it difficult and fatiguing
to uphold performance during an 8-h working day. Moreover, planning and complex
problem- solving seem to critically depend on intact executive control. Finally, declined
response inhibition may increase the likelihood of inappropriate behaviour (e.g. not being
able to ‘keep one’s cool’ during a conflict). Thus, impaired executive control might be
related to work performance during burnout and also to the success of rehabilitation after
burnout.
Self-reported cognitive failures
We formed the following hypothesis:
Hypothesis 1. Burnout will be associated with a high number of self-reported cognitive
difficulties.
We measured cognitive failures in daily life with the Cognitive Failure Questionnaire
(CFQ) of Broadbent, Cooper, FitzGerald, & Parkes (1982). This questionnaire measures
the self-reported frequency of daily cognitive failures, such as saying things that
you immediately regret, not attending to important cues in the environment (e.g. traffic
signs), and forgetting names. The CFQ is based on the assumption that many cognitive
mistakes and slips of actions are manifestations of inappropriate attention to response
selection or ineffective voluntary control over automatic perceptual and motor processes
(Broadbent et al., 1982; Manly et al., 1999; Robertson, Manly, Andrade, Baddeley, &
Yiend, 1997). Thus, control processes, assumed to underlie cognitive failures, fit with
current views of executive control (Miller & Cohen, 2001). Moreover, it has been shown
that people whose executive control is disturbed (e.g. patients with damage to the frontal
lobe) display many of the typical cognitive failures as assessed with the CFQ (Robertson
et al., 1997).
Actual cognitive performance
To determine whether the cognitive complaints of burned out individuals are accompanied
with true cognitive deficits, we assessed the ability to sustain attention to action in order to
voluntarily inhibit an automatic response. West (2001) argued that this ability is transient in
nature and fluctuates in efficiency over time. Consequently, even under optimal conditions,
people may occasionally show lapses in attention, resulting in the execution of unwanted
actions. In individuals with compromised executive control, however, such ‘lapses’ tend to
occur more frequently and to last longer (West, 2001). Moreover, several studies have
shown that mental fatigue, in general, is associated with compromised executive control
(van der Linden, Frese, & Meijman, 2003; van der Linden, Frese, & Sonnentag, 2003).
Thus, we expected that, compared to no-burnout controls, burned out individuals would
show more lapses of attention, which would result in an increased number of failures to
inhibit an automatic response. This expectation was tested with the Sustained Attention to
Response Task (the SART; Robertson et al., 1997). The SART is a simple reaction time
task that is constructed in such a way that it quickly leads to an automatic tendency to
Burnout and cognitive failures 25
respond to every stimulus. However, on few, unexpected occasions, participants have to
inhibit their response.
Robertson et al. (1997) and Manly et al. (1999) argued that successfully inhibiting a
response in the SART, strongly requires sustained attention. They showed that this is
particularly difficult for individuals whose executive control is impaired. Thus, the SART
provides a good measure of executive control functioning. In line with this, we formed the
second hypothesis:
Hypothesis 2. A high level of burnout symptoms will be associated with an increased number
of inhibition failures in the SART.
Another consequence of increased lapses in attention is that performance becomes more
variable (De Jong et al., 1999). During periods of insufficient attention to ongoing
behaviour, there is an increased proneness to occasionally make errors or to react slowly.
For example, West (2001) found that in reaction time tasks, disturbed executive control is
characterized by an increased number of trials having extremely long reaction times. He
subsequently argued that in studies on executive control, performance variability should be
taken into account. Therefore, in this study we also used a pencil-and-paper cancellation
task, in which performance variability is an important outcome measure (the Bourdon-
Wiersma Test; van der Ven & Smit, 1989).
Hypothesis 3. Compared to no-burnout controls, burned out individuals will show increased
performance variability on the Bourdon-Wiersma Test.
Severity of burnout symptoms and cognitive deficits
The obvious design in a first study on cognitive effects of burnout would be to compare a
group of burned out individuals with a no-burnout control group. However, the question
then arises of which burnout group one should use in such a study. More specifically, in the
working population there is a group of people for whom burnout symptoms have such an
impact on their daily functioning that they feel they can no longer uphold their regular work
activities and give up (Schaufeli & van Dierendonck, 1993). These individuals are often
diagnosed as burned out by health and safety services and may therefore be labelled as
‘clinical burnout’ cases. On the other hand, it is widely acknowledged that there are also
employees who experience high levels of burnout symptoms but nevertheless maintain their
regular work activities (Schaufeli & Enzmann, 1998). These individuals are considered to
be in the ‘danger zone’ and some of them might eventually also become clinically burntout.
The difference between these two types of people with burnout symptoms might merely be
one of degree, but some researchers have suggested that there might be some more discrete
differences. For example, it has been suggested that clinical burnout patients have passed
beyond some point in which their body’s stress system shows some protective adaptations
(Keijsers, 2001). These adaptations make it difficult to deal with further demands and
might eventually lead to temporary or permanent absence from work. It would go beyond
the scope of this study to discuss these issues in detail. However, as the possibility exists that
clinical burnout cases differ in some respects from high-burnout employees who are still
working, we decided to compare both groups to a no-burnout control group. Such a three-
group design allows us to examine in an exploratory way whether cognitive deficits are
particularly present in clinical burnouts or can also be found in high (non-clinical) burnout
employees.
26 D. van der Linden et al.
Method
Participants
Three different groups (total n
/43) participated in this study. The group of clinical
burnout cases consisted of individuals (n
/ 13) who volunteered after requests for
participation from the professional (mental) health organizations that provided them with
treatment for their burnout. These participants had various professional backgrounds (for
example, human resource manager, high school teacher, professor, clerk).
Participants in the clinical burnout group (henceforth referred to as the clinical
burnout cases group) were selected, based on their diagnosis of burnout, as established
by official health specialists. The term ‘clinical’ in this group is used for descriptive
purposes and refers to the fact that participants in this group could no longer uphold
their normal work routine and they sought professional help. Selection criteria
for participants in the clinical burnout group also included that those individuals
could be identified as burnout cases, based on the assessment procedures proposed
by Brenninkmeijer and van Yperen (2003). The UBOS (Schaufeli & van Dierendonck,
1993) is the Dutch translation of the Maslach Burnout Inventory, one of the most
widely used instruments in burnout research. Brenninkmeijer and van Yperen (2003)
proposed that for research purposes, burned out individuals can be identified by a
high score on the exhaustion scale of the UBOS (exhaustion cut-off point
/2.67),
combined with a highly unfavourable score on at least one of the two other main
components of burnout, namely cynical attitude (cut-off point
/2.25) and dimi-
nished personal accomplishment (cut-off point B
/3.50). The cut-off scores we used
in the current study to assess burnout symptoms were based on well-validated norm
scores of the UBOS (Brenninkmeijer & van Yperen, 2003; Schaufeli & van Dierendonck,
1993).
Participants in the group of working employees with high burnout scores*
/henceforth
referred to as high (non-clinical) burnout employees*
/and the no-burnout control
group were recruited at a vocational training institute. Two hundred teachers at the
institute were asked to fill in questionnaires about their level of burnout, their level
of depressive symptoms during the previous half a year, and their level of cognitive failures
in daily life (see Materials section for a description of the questionnaires). Eighty-five
teachers returned the questionnaires. Of these 85 teachers, 30 were selected for
participation in the current study and they were assigned to the high (non-clinical) burnout
employees group (n
/16) or to the no-burnout control group (n/ 14), based on their
UBOS scores. Criteria for assigning participants to the high (non-clinical) burnout
employees group were that they scored in the upper range (of the teachers who returned
the questionnaires) of the emotional exhaustion scale of the UBOS (
/75th percentile;
Brenninkmijer & van Yperen, 2003) and on at least one of the two other factors of burnout
(cynical attitude
/75th percentile; diminished personal accomplished B/25th percentile).
A similar assessment procedure was used for assigning participants to the no-burnout
control group (scores within the lowest range of exhaustion and favourable scores on the
two other factors).
A common finding in burnout research is that burnout symptoms clearly coincide with
depressive symptoms, yet burnout is not the same as clinical depression (Glass &
McKnight, 1996). Therefore in this study, we also tested for depressive symptomatology
with a validated depression scale (see Materials ).
Burnout and cognitive failures 27
Materials
Burnout symptoms. Severity of burnout symptoms were assessed with the Dutch version of
the Maslach Burnout Inventory for general professions (UBOS; Schaufeli & van
Dierendonck, 1993). The Maslach Burnout Inventory and its cross-language derivates
are the instruments most often used in burnout research. The UBOS consists of 16 items
with a 7-point Likert scale, assessing the three most important symptoms of burnout:
emotional exhaustion (sample item: ‘I feel mentally exhausted by my job’), cynical attitude/
depersonalization (sample item: ‘I became more cynical about the effects of my work’), and
diminished personal accomplishment (sample item: ‘I think I am really good at my job’).
Internal consistencies (as measured with Cronbach’s a ) were respectively; a
/.93, a / .80,
and a
/ .72.
Self-reported cognitive failures in daily life. These were measured with the Dutch translation of
the Cognitive Failure Questionnaire (CFQ). This questionnaire consists of 25 items in a
5-point Likert format. The questions refer to the level of slips of action, inattentiveness, and
forgetfulness in daily life (Broadbent et al., 1982). Sample items are: ‘Do you say something
and realize afterwards that it might be taken as insulting?’ and ‘Do you find you forget what
you came to the shops to buy?’. Internal consistency of this scale was a
/.93.
Depressive symptoms. These were measured with the Center for Epidemiologic Studies
Depression Scale (CES-D; Radlof, 1977). The CES-D assesses symptoms of depression
during the 3 months previous to filling out the questionnaire. It is a widely used instrument
to measure depressive symptoms. The CES-D contains 20 items, in a 4-point Likert
format, divided over four aspects of depressive symptomatology. These four aspects are
depressed affect (seven items: e.g. ‘I felt depressed’), positive affect (seven items: e.g. ‘I did
not enjoy life’), somatic effects (four items: e.g. ‘My sleep was restless’), and interpersonal
problems (two items: e.g. I felt people disliked me). The CES-D had high internal reliability
(a
/ .92).
Sustained Attention to Response Test (SART). In the SART, digits, ranging from ‘1’ to ‘9’,
were sequentially and in a quasi-random order, presented at the centre of a computer
screen. Digits were presented in Courier letter type with random size (26, 28, 36, or 72
mm). Participants sat approximately 50 cm from the screen. A portable Asus L3800S
computer connected to a 15-inch monitor was used for the SART task. Participants were
instructed to push a button as quickly as possible after the occurrence of a digit, but not
when the digit ‘3’ appeared (which occurred in 11% of the trials). Responses had to be
given on a button box that was placed in front of the participants. The computer recorded
whether or not a response was given and reaction time (RT) in milliseconds (ms). The task
lasted for approximately eight-and-a-half minutes, in which 450 digits were presented. In
each trial, a digit was presented for 250 ms. Inter-stimulus interval was held constant at 900
ms. In the SART, the most important measure is the number of inhibition errors in which a
participant presses a button when a ‘3’ occurs. In addition, we looked at the median
reaction time of all trials in which a correct response was given.
The Bourdon-Wiersma (Sustained Attention) Test. The Bourdon-Wiersma Test consisted of a
scoring form with 50 lines, each line containing 25 groups of dots (consisting of either 3, 4,
or 5 dots, in random order). Participants were instructed to mark all groups of 4 dots
(targets) as quickly and accurately as possible. Lines contain 7 to 12 targets in a random
28 D. van der Linden et al.
order. The Bourdon-Wiersma test is extensively used in clinical practice to assess difficulties
with sustained attention (van der Ven & Smit, 1989).
The experimenter used a digital stopwatch to measure the time participants required for
each line. Performance variability in the Bourdon-Wiersma test was operationalized as the
longest row time minus the shortest row time. Accuracy was measured by the total number
of omissions (forgetting to mark a target group).
Self-reported performance measures. In addition to the Cognitive Failure Questionnaire, as a
measure of self-reported cognitive deficits in daily life, we also added some questions to
assess the level of self-reported cognitive deficits during the laboratory tasks. We introduced
three questions, asking participants to report their perception of their cognitive perfor-
mance during the SART, using a 5-point Likert scale. The questions concerned the level of
distracting thoughts during the task, the difficulty in upholding concentration during the
task, and the perception of how well they did on the tasks. A sample item is ‘I had
difficulties keeping my mind on the task’. Internal consistency of the three questions was
a
/0.81.
Procedure during performance testing
Testing was done on location. Teachers were tested individually in an empty room at the
vocational training institute. Participants in the clinical burnout group were tested in a quiet
room in their homes. Otherwise, all participants followed the same procedure.
Beforehand, participants received only minimal information about the purpose of the
study. They were told (without further specification) that the study involves a test of ‘the
cognitive effects of stress’. During the test session, participants first worked on the SART
and then were requested to fill in the questions about their perceived performance on the
SART. Subsequently, the Bourdon-Wiersma test was presented.
Results
Sample description
The three groups did not differ significantly in age (F(2, 39)
/3.04, p /.05, overall M/47
yrs) or years of work experience (F(2, 39)
/1.47, p / .05, overall M/13 yrs).
Participants in the clinical burnout group did not necessarily have the same type of job as
the participants in the other two groups; however, there were no group significant
differences in average educational level (F (2, 39)
/.11, p/ .05). Thus, as the groups did
not differ on potentially confounding characteristics (age, experience, education), valid
between-group comparisons on task performance could be made.
The mean scores on burnout dimensions are reported in Table I. Not surprisingly,
analysis of variance tests showed that our selection criteria led to significant differences
between the three groups on emotional exhaustion (F(2, 40)
/109.5, p B/.001), and cynical
attitude (F(2, 40)
/63.9. p B/.001). Mean burnout scores of the high (non-clinical)
burnout employees were in between the mean scores of the clinical burnout group and the
no-burnout controls. Post-hoc tests (Tukey’s HSD) showed that all three groups
significantly differed from each other on these measures (all ps B
/.05). The groups did
not differ significantly on the perceived personal accomplishment scale (F (2, 40)
/1.9,
p
/.05).
Burnout and cognitive failures 29
In our sample, a high level of burnout symptoms was also associated with more
depressive symptoms (F (2, 40)
/48.2, pB/ .001). Post-hoc tests showed all three groups to
differ significantly from each other with respect to depressive symptoms (all p sB
/.05).
Cognitive failures in daily life
Analyses of variance, with CFQ scores as dependent variable and Group (clinical cases,
high non-clinical burnout employees, no-burnout controls) as between-subject factor,
revealed a significant effect for Group (F(2, 40)
/43.7, p B/.001, see Table II for means and
SD). Post-hoc tests (Tukey’s HSD) showed that the clinical burnout group reported
significantly more daily cognitive failures than high (non-clinical) burnout employees, who
in turn reported significantly more cognitive failures than the no-burnout controls
(see Table II for the means).
Thus Hypothesis 1, which stated that high levels of burnout symptoms coincide with a
high number of self-reported cognitive difficulties, was confirmed.
SART performance
In the analysis of SART inhibition errors, we controlled for reaction times (RT), because a
general finding with RT tasks is that participants tend to strategically adjust speed or
accuracy to uphold performance. As accuracy (not responding when a target stimulus
occurs) was the most salient performance aspect of the SART, we expected participants to
adjust their reaction time in favour of accuracy. This expectation was supported by the
negative correlation between reaction time and inhibition errors in the SART (r
/ /.31,
p B
/.05). Fishers r-to-z test, showed that the speed-accuracy correlations between the three
groups did not differ (x
2
/2.18, p /.05).
Analysis of covariance of inhibition errors, in which we controlled for RTs, revealed a
significant effect for Group (F (2, 39)
/3.5, p B/.05). Post-hoc ANCOVAs (pair-wise group
comparisons, controlling for RT) showed that, compared to the no-burnout controls, the
clinical burnout group made more inhibition errors (F (1, 24)
/6.1, p B/.05; for means, see
Table II). The high (non-clinical) burnout employees did not significantly differ on
inhibition errors, neither from the no-burnout controls (F (1, 26)
/1.7, p /.05) nor from
the clinical burnout group (F (1, 27)
/2.1, p /.05). The groups did not significantly differ
in RT (F (2, 40)
/.40, p /.05).
Table I. Mean (and SD) burnout and depressive symptoms scores for the different groups.
No-burnout
controls
High (non-clinical)
burnout employees
Clinical burnout
cases group
M (SD) M (SD) M (SD)
Burnout measures
Emotional Exhaustion
a
** 0.5 (0.4) 3.4 (0.8) 4.6 (0.9)
Cynical attitude
a
** 0.4 (0.3) 2.0 (0.8) 3.7 (0.9)
Personal Accomplishment 3.9 (0.4) 4.2 (0.7) 3.7 (1.0)
Depressive symptoms
a
** 0.3 (0.1) 0.6 (0.3) 1.3 (0.6)
a
All groups significantly differed from each other. ** p B/.01.
30 D. van der Linden et al.
Bourdon-Wiersma test
Analysis of variance showed that there was a significant effect for Group on performance
variability (F (2, 40)
/13.3, p B/.001) in the Bourdon-Wiersma test. Post-hoc tests (Tukey’s
HSD) revealed that, compared to no-burnout controls, the clinical burnout group displayed
significantly more variability in performance. No-burnout controls also showed significantly
less variability than the high (non-clinical) burnout employees. Number of omissions
(forgetting to mark target stimuli) did not significantly differ between the groups
(F(2, 40)
/2.5, p /.05). Moreover, accuracy was not significantly correlated with
performance variability (r
/.09, p /.05).
In general, Hypotheses 2 and 3, which stated that high levels of burnout symptoms are
accompanied by inhibition and attention failures, were confirmed.
Relationship between self-reported attention problems and objective performance
In addition to the group comparisons, we correlated CFQ scores with performance
measures of the SART and Bourdon-Wiersma Test. CFQ scores correlated highly with
inhibition errors in the SART (r (43)
/.53, p B/ .001), as well as with performance
variability in the Bourdon-Wiersma Test (r (43)
/.52, p B/.001). There was no significant
correlation between on the one hand, CFQ scores and on the other hand, RTs in the SART
(r (43)
/ /.17, p /.05) and omission errors in the Bourdon-Wiersma Test (r (43) /.14,
p
/.05). None of the reported correlations between CFQ scores and task performance
measures differed significantly between the groups (Fisher’s r-to-z test, all p s
/.05).
Three questions on participants’ perception of (cognitive) performance in the SART were
used as additional tests of the relationship between self-reported cognitive difficulties and
actual performance. Analysis of variance, with the mean scores on the three questions as
dependent variable and Group as between-subject factor, showed significant differences
between the groups on reported cognitive difficulties during the SART (F (2, 40)
/14.6,
p B
/.001). Post-hoc comparisons were all significant (pB/ .05) and showed that in accordance
with their actual performance on the attention tests, the clinical burnout group reported
the most cognitive difficulties (M
/3.5), followed by the high (non-clinical) burnout
employees (M
/2.9), with the no-burnout controls reporting the least difficulties with
cognition (M
/2.5). Moreover, level of self-reported cognitive difficulties during the tasks
Table II. Means (and SD) of cognitive failures and performance measures for the different groups.
No-burnout
controls
High (non-clinical)
burnout employees
Clinical burnout
cases group
M (SD) M (SD) M (SD)
CFQ (total)
a
** 1.0 (0.3) 1.5 (0.5) 2.4 (0.3)
SART
Inhibition errors
b
* 6.4 (4.4) 8.8 (5.5) 11.7 (6.7)
Reaction Time
c
358 (52) 350 (45) 366 (46)
Bourdon-Wiersma
Performance variability
a,d
** 5.3 (0.9) 7.4 (2.0) 8.5 (1.8)
Omissions 15.5 (7.7) 20.2 (18.6) 16.9 (13)
a
All groups significantly differed from each other.
b
Only the no-burnout controls and the clinical
burnout group significantly differed from each other.
c
Milliseconds.
d
Longest minus shortest
line time. CFQ/Cognitive Failure Questionnaire. SART/Sustained Attention to Response Test.
* p B/.05; ** p B/.01.
Burnout and cognitive failures 31
(concentration problems, distracting thoughts) were highly correlated with actual task
performance (with SART inhibition errors r (43)
/.55, p B/.001 and Bourdon-Wiersma
Performance variability r (43)
/.49, p /.001).
Discussion
As burnout symptoms reflect exposure to (work) stress, it can be expected that the
syndrome affects not only work motivation and work-related activities but also general
information processing. Therefore, we examined whether burned-out individuals showed
difficulties in sustained attention and response inhibition. These two abilities are important
aspects of executive control (Duncan et al., 1996; Miller & Cohen, 2001; Norman &
Shallice, 1986).
We found that severe burnout symptoms were associated with many self-reported
cognitive failures (CFQ). This finding supports earlier clinical observations indicating that
burned out individuals have difficulties with attention in daily tasks (Keijsers, 2001;
Schaufeli & Enzmann, 1998). Moreover, in the laboratory tasks of attention, burnout was
associated with poor performance; participants with burnout symptoms did not adequately
allocate attention to action. Subsequently, their behaviour tended to be guided by more
automatic cognitive processes, which led to increased distraction (i.e. performance
variability in the Bourdon Test) and inhibition errors (SART). It is important to note
that compared to the controls, participants with burnout symptoms did not perform
significantly worse on all aspects of the laboratory tasks. For example, reaction times on
the SART and accuracy in the Bourdon-Wiersma did not differ significantly between the
groups. Burnout participants performed particularly badly on those task aspects that in the
research literature are explicitly linked to executive control functioning (Duncan et al.,
1996; Kimberg & Farah, 1993; Norman & Shallice, 1986).
To our knowledge, the current study is the first to describe empirical findings indicating
that burnout is associated with deficits in executive control. As such, this study enhances
knowledge about the range of behavioural manifestations that may accompany burnout.
Moreover, the study also reveals the relationship between self-reported cognitive difficulties
and actual cognitive performance in burnout; Cognitive Failure Questionnaire scores and
self-reported cognitive difficulties during the SART, were strongly associated with actual
cognitive performance on the attention tasks. This is an important finding because such an
association is not always obvious. For example, several studies showed that Chronic Fatigue
Syndrome (CFS) patients perform as well as control groups on many cognitive tasks in the
laboratory, even though they often complain about concentration problems (Wearden &
Appleby, 1997). Several researchers have suggested that complaints of CFS patients reflect
‘misinterpretation’ of their cognitive performance or ‘unrealistic performance standards’,
rather than true cognitive deficits (Fry & Martin, 1996). However, in contrast to these
findings in CFS patients, the current study suggests that people with burnout symptoms
have a fairly realistic insight into their cognitive performance level.
Three different groups were compared in this study, thus we also gained some
preliminary insight into the link between severity of burnout symptoms and level of
cognitive deficits. The high (non-clinical) burnout employees had a level of burnout
symptoms that fell in between the clinical burnout group and the no-burnout controls.
Interestingly, the level of cognitive deficits in the three groups showed the same pattern:
highest in the clinical burnout group, intermediate in the high (non-clinical) burnout
employees, and lowest in the no-burnout control group. This finding indicates two things.
32 D. van der Linden et al.
First, it suggests that cognitive deficits are not only found in the clinical group but also
occur in employees who are still on the job but experience burnout symptoms. Second, it
suggests that the cognitive effects of burnout do not develop in discrete stages but gradually
rise in line with the severity of main burnout symptoms (e.g. exhaustion).
As we found clear indications that burnout is indeed accompanied by attentional deficits,
a future research goal might be to identify the underlying mechanism that link burnout with
cognitive deficits. Several studies in the area of stress and cognition reveal some promising
theoretical underpinning for such a mechanism. It has been proposed that dealing with high
levels of prolonged or uncontrollable (work) stress, initially leads to enhanced release of
stress hormones whereas in the long term it may lead to lowered basal levels of stress
hormones (Deutch & Young, 1995). Interestingly, the levels of stress hormones have also
been linked to cognitive functioning. For example, there is substantial evidence showing
that either too high or too low levels of stress hormones are detrimental for executive
control (Arnsten, 1998; Sullivan & Gratton, 2002). Subsequently, stress may cause brain
areas, underlying executive control, to go ‘off-line’ in favour of more automatic cognitive
processes (Arnsten, 1998; Arnsten & Goldman-Rakic, 1998). Thus, as burnout symptoms
are stress-related, it might be possible that the burnout-cognition link that we found in the
current study is mediated by changes in stress hormones. Future studies should address
these issues in a more direct way (e.g. by also using physiological measures).
Study limitations
Although the current study may advance knowledge about burnout, there were also some
limitations. First, the scope of the cognitive tests that we used was limited. In the
Introduction we argued that executive control involves a diversity of cognitive deficits
ranging from difficulties in working memory to problems with inhibition. In this study we
focused on sustained attention and on response inhibition. Both are important aspects of
executive control. However, in order to more fully understand the relationship between
burnout symptoms and cognition, future studies should also assess other executive control
processes (e.g. working memory functioning, planning, cognitive flexibility, dual task
performance). In addition, future studies should also use tasks that measure automatic
cognitive processes to determine if performance on these tasks stays unimpaired.
A second limitation is the relatively small sample size. In general occupational health
studies it is quite easy to obtain large numbers of participants (e.g. company surveys). In
contrast, it is relatively difficult to find burnout patients who are willing to participate in an
experimental study. Consequently, group sizes in this pilot study ranged from 13 to 16,
which may limit generalizability of the results. On the other hand, finding significant
differences with such relatively small groups suggests that the effects sizes of cognitive
deficits associated with burnout may be substantial. Moreover, our finding that the level of
burnout symptoms in the three groups showed the same pattern as the level of cognitive
deficits, strongly reduces the probability of chance findings. Nevertheless, future studies
might want to use larger samples to examine generalizability of the current results.
Another important limitation of this study relates to the issue of causality. We did not
follow participants over time. Therefore, it is not possible to directly infer a causal
relationship. In future studies examining causal directions will be crucial, because
Broadbent et al. (1982) suggested, already in the early 1980s, that people who experience
many cognitive failures in daily life might also be more susceptible to negative health effects
when exposed to stressors. They argued that cognitive processes, playing a major role in
executive control of behaviour (and thus also in cognitive failures), might also be important
Burnout and cognitive failures 33
for dealing adequately with stressors. Several animal studies support this idea and have
shown that in contrast to healthy animals (either rats or primates), animals with induced
damage to the pre-frontal cortex of the brain did not show adequate coping responses to
stressors and showed prolonged negative physiological stress reactions (Sullivan & Gratton,
2002). The integrity of the pre-frontal cortex of the brain is crucial for executive control.
From the line of reasoning we adopted above one may infer that, when executive control
is sub-optimal, people may be less able to prevent the negative effects of stressors.
Longitudinal studies may provide more insight into causal relationships between work
stress, burnout, and executive control and may establish whether compromised executive
control precedes the negative effects of work stress, accompany work stress, or is
reciprocally related to it.
Potential implications for practice
In general, our study indicates that burnout is accompanied by objective changes in
information processing. This finding also provides some insight into the potential
difficulties that burned out individuals may experience in their professional lives. For
example, when executive control is compromised, burned out individuals might not only
experience more cognitive failures but they might also have problems with planning and
working in a goal-directed way. Trying to uphold (work) performance under such
circumstances may be rather effortful and stressful, and may even lead to a downward
spiral that possibly increases burnout symptoms such as fatigue. In addition, it can be
expected that compromised executive control may also reduce the probability of a
successful return to the job after burnout. In as far as this is the case, treatment
programmes for burnout might want to consider using methods that enhance executive
control. Clues about which methods might be useful can be found in neuropsychological
studies showing that several aids can support the high level control of attention and
behaviour; for example, explicit step-by-step planning of behaviour or external ‘reminders’
to maintain attentional focus. Such aids may also be useful for burned-out individuals who
want to resume work activities, but are not yet fully recovered from the potential cognitive
effects of burnout.
Another potential practical implication of the current study is that one might consider
using cognitive tests in the diagnosis of burnout. It has to be noted that attentional
difficulties are involved in many psychological disorders and in themselves do not
specifically indicate burnout (Gazzaniga et al., 1998). However, future studies may indicate
whether cognitive tests might provide a useful addition to tools for the diagnosis of burnout,
which currently mainly rely on self-report (e.g. interviews and questionnaires).
References
Arnsten, A. F. T. (1998). Development of the cerebral cortex: XIV: Stress impairs prefrontal cortical function.
Journal of American Child and Adolescense Psychiatry, 37 , 1337
/1339.
Arnsten, A. F. T., & Goldman-Rakic, P. S. (1998). Noise stress impairs prefrontal cortical cognitive functions in
monkeys. Archives of General Psychiatry, 55 , 362 /368.
Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and executive functions: Constructing a unified
theory of ADHD. Psychological Bulletin , 121 ,65
/94.
Brenninkmeijer, V., & van Yperen, N. W. (2003). How to conduct research on burnout: Advantages and
disadvantages of a unidimensional approach in burnout research. Occupational & Environmental Medicine , 60 ,
16
/20.
Broadbent, D. E., Cooper, P. F., FitzGerald, P., & Parkes, K. R. (1982). The Cognitive Failures Questionnaire
(CFQ) and its correlates. Journal of Clinical Psychology, 21 ,1/16.
34 D. van der Linden et al.
De Jong, R., Berendsen, E., & Cools, R. (1999). Goal neglect and inhibitory limitations: Dissociable causes of
interference effects in conflict situations. Acta Psychologica , 101 , 379
/394.
de Lange, A. H., Taris, T. W., Kompier, M. A. J., Houtman, I. L. D., & Bongers, P. M. (2002). Effects of stable and
changing demand-control histories on worker health. Scandanavian Journal of Work, Environment, & Health , 28 ,
94
/108.
Deutch, A. Y., & Young, C. D. (1995). A model of the stress-induced activation of prefrontal cortical dopamine
systems: Coping and the development of post-traumatic stress disorder. In M. J. Friedman, D. S. Charney, & A.
Y. Deutch (Eds.), Neurobiological and clinical consequences of stress: From normal adaptation to post-traumatic stress
disorder. New York: Lipincott-Raven.
Duncan, J., Emslie, H., Williams, P., Johnson, R., & Freer, C. (1996). Intelligence and the frontal lobe: The
organization of goal-directed behavior. Cognitive Psychology, 30 , 257
/303.
Fry, A. M., & Martin, M. (1996). Fatigue in the chronic fatigue syndrome: A cognitive phenomenon? Journal of
Psychosomatic Research , 41 , 415
/426.
Gazzaniga, M. S., Ivry, R. B., & Mangun, G. R. (1998). Cognitive neuroscience: The biology of mind . New York:
Norton & Co.
Glass, D. C., & McKnight, J. D. (1996). Perceived control, depressive symptomatology, and professional burnout:
A review of the evidence. Psychology and Health , 11 ,23
/48.
Keijsers, G. P. J. (2001). Neuro-endocrinologische, psychologische en cognitieve processen bij burnout
[Neuroendocrine, psychological, and cognitive processes in burnout]. In C. P. F. v. d. Staak, G. P. J. Keijsers,
& C. A. L. Hoogduin (Eds.), Psychologische interventies bij werkgerelateerde problematiek (pp. 9
/33). Zeist: Cure &
Care Publishers.
Kimberg, D. Y., & Farah, M. J. (1993). A unified account of cognitive impairments following frontal lobe damage:
The role of working memory in complex organized behavior. Journal of Experimental Psychology: General , 122 ,
411
/428.
Manly, T., Robertson, I. H., Galloway, M., & Hawkins, K. (1999). The absent mind: Further investigations of
sustained attention to response. Neuropsychologica , 37 , 661
/670.
Maslach, C., & Schaufeli, W. (2000). Job burnout. Annual Review of Psychology, 52 , 397
/422.
Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of
Neuroscience , 24 , 167
/202.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerte, A., & Wager, T. D. (2000). The unity and
diversity of executive functions and their contribution to complex ‘frontal lobe’ tasks: A latent variable analysis.
Cognitive Psychology, 41 ,49
/100.
Monsell, S., & Driver, J. E. (2000). Control of cognitive processes: Attention and performance XVIII. Cambridge, MA:
MIT Press.
Norman, D. A., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R. J.
Davidson, G. E. Swartz, & D. Shapiro (Eds.), Consciousness and self-regulation: Advances in theory and research
(Vol. 4 (pp. 1
/18). New York: Plenum Press.
Rabbitt, P. (1997). Introduction: Methodologies and models in the study of executive function. In P. Rabbitt (Ed.),
Methodology of frontal and executive function . Hove: Psychology Press.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population.
Applied Psychological Measurement , 1 , 385
/401.
Robertson, I. H., Manly, T., Andrade, J., Baddeley, B., & Yiend, J. (1997). ‘Oops!’: Performance correlates of
everyday attentional failures in traumatic brain injured and normal subjects. Neuropsychologica , 35 , 747
/758.
Schaufeli, W., & Enzmann, D. (1998). The burnout companion to study and practice: A critical analysis . London:
Taylor & Francis.
Schaufeli, W., & van Dierendonck, D. (1993). The construct validity of two burnout measures. Journal of
Organizational Behavior , 14 , 631
/664.
Shallice, T., & Burgess, P. (1991). Higher-order cognitive impairments and frontal lobe lesions in man. In H. S.
Levin, H. M. Eisenberg, & A. L. Benton (Eds.), Frontal lobe function and dysfunction (pp. 125
/138). Oxford:
Oxford University Press.
Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic information processing. II: Perceptual learning,
automatic attending, and a general theory. Psychological Review, 84 , 127
/190.
Sullivan, R. M., & Gratton, A. (2002). Prefrontal cortical regulation of hypothalamic-pituitary-adrenal function in
the rat and implications for psychopathology: Side matters. Psychoneuroendocrinology, 27 ,99
/114.
Taris, T. W., Schreurs, P. J. G., & Van Iersel-van Silfhout, I. J. (2001). Job stress, job strain, and psychological
withdrawal among Dutch university staff: Towards a dual process model for the effects of occupational stress.
Work & Stress , 15 , 283
/296.
Burnout and cognitive failures 35
van der Doef, M., & Maes, S. (1999). The job Demand-Control (-Support) model and psychological well-being: A
review of 20 years of empirical research. Work and Stress , 13 ,87
/114.
van der Linden, D., Frese, M., & Meijman, T. F. (2003). Mental fatigue and the control of cognitive processes:
Effects on perseveration and planning. Acta Psychologica , 113 ,45/65.
van der Linden, D., Frese, M., & Sonnentag, S. (2003). The impact of mental fatigue on exploration in a complex
computer task: Rigidity and loss of systematic strategies. Human Factors , 45 , 483
/494.
Van der Ven, A. H., & Smit, J. C. (1989). Het geheim van de Bourdon-Wiersma test ontraadseld? [The secret of the
Bourdon-Wiersma test unraveled?]. Nederlands tijdschrift voor de psychologie en haar grensgebieden , 44 , 260 /270.
Wearden, A. J., & Appleby, L. (1997). Cognitive performance and complaints of cognitive impairment in chronic
fatigue syndrome. Psychological Medicine , 27 ,81/90.
West, R. (2001). The transient nature of executive control processes in younger and older adults. European Journal
of Cognitive Psychology, 13 ,91
/105.
36 D. van der Linden et al.