RESEARCH NOTES
Design Blueprint for Stress-Sensitive Adaptive Enterprise Systems
Marc T. P. Adam
•
Henner Gimpel
•
Alexander Maedche
•
Rene
´
Riedl
Received: 16 April 2014 / Accepted: 4 February 2016 / Published online: 5 September 2016
Springer Fachmedien Wiesbaden 2016
Abstract Stress is a major problem in the human society,
impairing the well-being, health, performance, and pro-
ductivity of many people worldwide. Most notably, people
increasingly experience stress during human-computer
interactions because of the ubiquity of and permanent
connection to information and communication technolo-
gies. This phenomenon is referred to as technostress.
Enterprise systems, designed to improve the productivity of
organizations, frequently contribute to this technostress and
thereby counteract their objective. Based on theoretical
foundations and input from exploratory interviews and
focus group discussions, the paper presents a design blue-
print for stress-sensitive adaptive enterprise systems
(SSAESes). A major characteristic of SSAESes is that bio-
signals (e.g., heart rate or skin conductance) are integrated
as real-time stress measures, with the goal that systems
automatically adapt to the users’ stress levels, thereby
improving human-computer interactions. Various design
interventions on the individual, technological, and organi-
zational levels promise to directly affect stressors or
moderate the impact of stressors on important negative
effects (e.g., health or performance). However, designing
and deploying SSAESes pose significant challenges with
respect to technical feasibility, social and ethical accept-
ability, as well as adoption and use. Considering these
challenges, the paper proposes a 4-stage step-by-step
implementation approach. With this Research Note on
technostress in organizations, the authors seek to stimulate
the discussion about a timely and important phenomenon,
particularly from a design science research perspective.
Keywords Adaptive automation Affective computing
Enterprise systems Biofeedback NeuroIS Stress
Technostress Design science research
Accepted after three revisions by Prof. Dr. Karagiannis.
Electronic supplementary material The online version of this
article (doi:10.1007/s12599-016-0451-3) contains supplementary
material, which is available to authorized users.
Dr. M. T. P. Adam (&)
The University of Newcastle, University Drive, Callaghan,
NSW 2308, Australia
e-mail: marc.adam@newcastle.edu.au
URL: http://www.newcastle.edu.au/profile/marc-adam
Prof. Dr. H. Gimpel
University of Augsburg, 86135 Augsburg, Germany
e-mail: henner.gimpel@fim-rc.de
URL: http://gimpel.fim-rc.de
Prof. Dr. H. Gimpel
Project Group Business & Information Systems Engineering,
Fraunhofer Institute for Applied Information Technology FIT,
Sankt Augustin, Germany
Prof. Dr. A. Maedche
Institute of Information Systems and Marketing (IISM) and
Karlsruhe Service Research Institute (KSRI), Karlsruhe Institute
of Technology (KIT), 76131 Karlsruhe, Germany
e-mail: alexander.maedche@kit.edu
URL: http://issd.iism.kit.edu/236.php
Prof. Dr. R. Riedl
University of Applied Sciences Upper Austria, 4400 Steyr,
Austria
e-mail: rene.riedl@fh-steyr.at
URL: http://research.fh-ooe.at/en/staff/30128
Prof. Dr. R. Riedl
University of Linz, Linz, Austria
123
Bus Inf Syst Eng 59(4):277–291 (2017)
DOI 10.1007/s12599-016-0451-3
1 Introduction
While the tremendous advances in the field of information
and communication technology (ICT) have resulted in
significant benefits for the human society, growing evi-
dence shows the ‘‘dark side’’ of ICT for individual users
and organizations (e.g., Salanova et al. 2013; Tarafdar et al.
2013). Technostress (TS), defined as ‘‘stress experienced
by end users in organizations as a result of their use of
ICTs’’ (Ragu-Nathan et al. 2008, pp. 417–418), is one
major ‘‘dark side’’ of ICT. While TS is not a new phe-
nomenon (Brod 1984), it has gained significant momentum
during the past few years, primarily because of ICT’s high
penetration of the human society. Today, at least in Wes-
tern countries, it is difficult to imagine people without a
personal computer, smartphone, or tablet; similarly, in
business life, it is difficult to imagine companies without
packaged application systems, including groupware,
enterprise resource planning (ERP), or business intelli-
gence and analytics (BI&A).
Organizational TS contributes to a general trend of an
increasing stress perception in human society with detri-
mental effects on human health and performance (Riedl
2013). The pervasiveness of ICT in firms, along with daily
incidents of computer hassles (e.g., system breakdown,
waiting times, or printer problems) can negatively affect
employees’ psychological and physiological conditions
(e.g., Ayyagari et al. 2011; Maier et al. 2015b; Riedl et al.
2012). Motivated by an increasing number of staff com-
plaints and by empirical research (e.g., Barley et al. 2011),
enterprises have already started to counteract TS and its
negative consequences by implementing interventions. In
2011, for instance, the automobile manufacturer
Volkswagen agreed to stop mail servers when employees
are off-shift in order to reduce stress levels (BBC News
Technology 2012).
However, increasing penetration of organizational tasks
and business processes with ICT, along with the intensive
use of devices and packaged application systems, does not
necessarily constitute a negative development for the
human society. The trade-off between maximizing the
benefits of ICT and minimizing TS levels and its negative
consequences needs to be taken care of with design and
intervention measures. In this context, Riedl (2013, p. 44)
wrote recently: ‘‘Design science researchers could con-
tribute to the development of information systems, which
use bio-signals as real-time system input in order to make
human-computer interactions less stressful.’’
In this Research Note, we intend to promote a discussion
among scholars, managers, and engineers by proposing a
design blueprint for how enterprise systems (ESes) can use
bio-signals in order to mitigate stress by means of inter-
ventions on the individual, technological, and
organizational levels. We refer to such systems as stress-
sensitive adaptive enterprise systems (SSAESes). The
blueprint addresses the following objective:
Design Objective: Support humans via information
systems (1) in managing stress in an enterprise context,
and (2) in reducing stress in order to increase well-being
and health, performance and productivity, and user
satisfaction.
Design is an iterative search process (Hevner et al.
2004). In order to structure this process, we followed the
objective-centered design science research process (Peffers
et al. 2007) and used a hybrid approach that balances
deductive, inductive, and abductive reasoning (Gregor
2009; Gregory and Muntermann 2011). In particular, we
integrate a body of highly fragmented theoretical and
empirical literature from various disciplines (deduction),
conduct a series of qualitative exploratory expert inter-
views and focus groups (induction), and propose a design
blueprint consisting of a set of design guidelines, an
architecture, a roadmap for implementation, and a plan for
empirical evaluation (abduction). Several interviewees
emphasized the relevance of this research. The occupa-
tional health and safety officer at Private Customer E (see
the Appendix; available online via http://link.springer.
com), for example, stated: ‘‘We believe that productivity
and innovation arise from a ‘healthy corporate climate’
and, thus, have detailed corporate policies to foster well-
being. In these policies, ‘well-being’ reflects a holistic
approach including prevention of accidents, workplace
ergonomics, nutritional and physical activity choices, as
well as stress management. […] Your approach to stress
management is very innovative and promising. We would
like to explore it in our organization.’’ Other potential users
of SSAESes supported this perspective in our interviews.
In total, we conducted 71 interviews with 39 different
experts (30 practitioners and 9 academics) from 25 dif-
ferent organizations and two focus groups to capture the
multiple realms of creativity, insight, and knowledge
required in designing SSAESes. We followed an interdis-
ciplinary approach by involving technology-oriented
practitioners who demonstrated subject matter expertise
(e.g., through innovative technology development or con-
sulting experience) as well as application-oriented practi-
tioners dealing with the challenges of technostress from a
company, labor organization or government point of view.
Furthermore, academics from the fields of electrical engi-
neering, psychology, computer science, and information
systems were interviewed. By doing so, we were able to
gather insights of subject matter experts from a diverse set
of industries (high-tech industry, manufacturing firms,
consulting services) as well as non-profit organizations
(trade unions, governmental regulatory agencies, research
organizations). Please see the Appendix for details on our
123
278 M. T. P. Adam et al.: Design Blueprint for Stress-Sensitive Adaptive Enterprise Systems, Bus Inf Syst Eng 59(4):277–291 (2017)
sample. Interviews varied in focus, length, and format
depending on the varying needs during the design process.
Early interviews were non-directive in-depth informant
interviews. During the design process, interviews gradually
turned towards semi-structured, directed participant inter-
views (Easterby-Smith et al. 2002, Ch. 5). These interviews
featured graphical and textual descriptions of the current
state of the design blueprint for participants (1) to evaluate
them descriptively (informed arguments and scenarios) and
analytically (static analysis and architectural analysis), and
(2) to suggest specific refinements. Expert sampling and
interview analysis followed the general notions of theo-
retical sampling, constant comparison, and theoretical
saturation as introduced by Glaser and Strauss (1967) in the
context of grounded theory. However, we did not transcribe
and formally code the interviews and did not apply the full
methodological toolbox associated with grounded theory.
Expert input rather served as inspiration as well as a con-
tribution to problem awareness and understanding. The
integration of theoretical and empirical literature from
various disciplines finally constituted the foundation to
compile a design blueprint for SSAESes.
The remainder of this article is organized as follows:
Sect. 2 discusses the theoretical basis of the design blue-
print, Sect. 3 presents the design blueprint, and Sect. 4
outlines the limitations and discusses directions for future
IS design science research in the area of TS.
2 Theoretical Foundations
This section conceptualizes the theoretical foundations and
building blocks of SSAESes. The lower part of Fig. 1
outlines a simplified model of TS in organizations, while
the upper part sketches how the individual, technological,
and organizational dimensions of ESes can interfere at
different stages with the process of stress elicitation. Our
model of TS is based on the Transactional Model of Stress
developed by Lazarus and Folkman (1984), which is one of
the most influential frameworks to study stress perceptions
and coping mechanisms in psychological and in IS research
(e.g., Ayyagari et al. 2011; Ragu-Nathan et al. 2008).
Moreover, our model is informed by a stress model
developed by Hancock and Warm (1989) and a TS model
developed by Riedl (2013). Our model of TS is intended to
be used as a guideline for investigating stress in the context
of ES, for identifying possible interventions, and for
informing the design and evaluation of SSAESes.
2.1 Transactional Model of Stress
Based on a number of empirical investigations by Richard
Lazarus at the nexus of physiology and psychology in the
1960s and 1970s, Lazarus and Folkman (1984) presented a
seminal theory to explain human stress reactions. The
major characteristic of this Transactional Model of Stress is
Job Characteristics
Performance and
Productivity
User
Satisfaction
Well-Being
and Health
Physiology
MODEL OF TECHNOSTRESS IN ORGANIZATIONS
Stressors
(secondary appraisal)
Stress reaction
(strain)
Consequences
POSSIBLE INTERVENTIONS
(ENTERPRISE SYSTEM)
Technological Environment
Organizational Environment
Social Environment
Emotion
Cognition
Behavior
Stimuli
(primary
appraisal)
TechnologyIndividual
Organization
Fig. 1 Model of technostress in organizations and possible interventions in enterprise systems
123
M. T. P. Adam et al.: Design Blueprint for Stress-Sensitive Adaptive Enterprise Systems, Bus Inf Syst Eng 59(4):277–291 (2017) 279
that stress is not solely conceptualized as a biological
phenomenon, but as a complex construct that results from
the interplay between an individual and the environment
(hence, the term ‘‘transactional’’). In particular, the theory
states that stress (1) emerges from an imbalance between
demands from the environment and an individual’s
resources, and (2) is subject to the meaning of a stimulus to
the perceiver, implying that the same stimulus may dif-
ferently affect the stress of different individuals.
According to the seminal stress theory by Lazarus and
Folkman (1984), the underlying rationale is that when
faced with stimuli (see Fig. 1), an individual evaluates
whether they are irrelevant, benign-positive, or stressful
(primary appraisal). In the latter case, another evaluation
process takes place (secondary appraisal). Here, the indi-
vidual assesses whether he/she can cope with the stimulus
(stressor) by using the available resources (e.g., institu-
tional, personal, and social). Two outcomes are possible:
the resources are either sufficient or they are not. In the
latter case, stress reactions are possible on four levels:
physiology, emotion, cognition, and behavior [see ‘‘Stress
reaction’’ in Fig. 1; note that the term ‘‘strain’’ is used as a
synonym in seminal research on organizational stress, see
Hakanen et al. (2006, p. 496), and this is consistent with
Lazarus and Folkman (1984, p. 4), who equate stress with
strain and explicitly define the latter, based on Wolff
(1953), as ‘‘a disturbed state of the body’’; also note that
consistent with our model in Fig. 1 organizational TS
research grouped strain into different types, such as phys-
ical, emotional, and cognitive/mental, see Boucsein and
Thum (1997)]. Next, to mitigate these stress reactions, an
individual applies different coping strategies, which can be
either problem-focused or emotion-focused (Hudiburg and
Necessary 1996; Lazarus and Folkman 1984). The former
strategy has the goal to actively change the person-envi-
ronment realities related to a stressful situation (e.g., by
increasing the amount or quality of resources), while the
latter seeks to reduce negative feelings by changing the
primary and/or secondary appraisal of a given stressful
situation.
Applying the rationale of the Transactional Model of
Stress in organizational settings, we find that stress is
generated as a dynamic process that is triggered by a set of
acute and chronic stressors (i.e., stress-creating factors and
conditions), and involves individual stress reactions,
which, in turn, have a number of consequences on well-
being and health, performance and productivity, and user
satisfaction (see ‘‘Consequences’’ in Fig. 1) (Hancock and
Warm 1989; Lazarus 1991; Riedl 2013). This dynamic
process includes conscious changes in perception; how-
ever, there are also unconscious changes in body physiol-
ogy that usually set in before conscious stress perception
(e.g., Riedl 2013; Tams et al. 2014b). This includes, for
example, the release of the stress hormones adrenaline
(Johannsson and Aronsson 1984), noradrenaline (Korunka
et al. 1996), and cortisol (Riedl et al. 2012) and other
chemical substances related to stress such as alpha-amylase
(Tams et al. 2014b), as well as changes in heart rate
(Trimmel et al. 2003), heart rate variability (Hjortskov
et al. 2004), blood pressure (Boucsein 2009), muscle ten-
sion (Emurian 1993; Hazlett and Benedek 2007), pupil
dilation (Partala and Surakka 2003; Buettner et al.
2013),
and skin conductance (Le
´
ger et al. 2010; Riedl et al. 2013).
Importantly, it needs to be emphasized that there is more
to the cognitive side than perception alone. Users can
cognitively intervene at an earlier stage of the process. As
explained by Lazarus and Folkman (1984), the elicitation
of stress is subject to the users’ appraisal of the overall
situation, availability of resources, and coping strategies. In
this vein, users can apply, for example, information
avoidance, stress management, and other coping strategies
in order to mitigate the elicitation of stress and its negative
consequences (Denson et al. 2009; Bostock et al. 2011).
Thus, the impact of stressors heavily depends on the users’
individual capabilities and stress-coping strategies.
2.2 Stressors
In an ES context, a large number of stressors exist. One
interviewee from a workers union indicated: ‘‘Many factors
are relevant for workplace stress, for example, sufficient
staffing levels, leadership culture, corporate culture, and
certainly also individual stress coping strategies.’’ In order
to facilitate stress interventions, we introduce a high-level
distinction among stressors related to (1) job characteris-
tics, (2) technological environment, (3) organizational
environment, and (4) social environment. These categories
have recently been described as crucial in organizational
TS (Fischer and Riedl 2015). These stressor types can
induce stress reactions in the users, both individually and
collectively. The categorization is useful for our goal of
SSAESes, as the context and usage data of the individual
users help to better understand their current situation and
trigger context-sensitive interventions at the level of the
current task, the technology, the organization, or the social
environment.
In the context of a specific task (i.e., job characteristics),
possible stressors are, for example, task monotony, task
complexity, and multi-tasking (Friend 1982; Tarafdar et al.
2011; Riedl 2013). As for technology-related stressors,
Tarafdar et al. (2007, 2011, pp. 116–117) identified five
different TS creators: techno-overload (i.e., too much, ICT
forces users to work faster and do more work than they can
handle), techno-invasion (i.e., always connected, blurring
boundaries between private life and work), techno-com-
plexity (i.e., devices have many features and their usage is
123
280 M. T. P. Adam et al.: Design Blueprint for Stress-Sensitive Adaptive Enterprise Systems, Bus Inf Syst Eng 59(4):277–291 (2017)
difficult to learn), techno-insecurity (i.e., fear of being
replaced by users with better ICT knowledge), and techno-
uncertainty (i.e., constant software and hardware changes).
We add techno-unreliability (i.e., system malfunctions and
other IT hassles) to the category of technology-related
stressors (Fischer and Riedl 2015). Organizational stressors
refer to the potential causes of stress originating in the
organizational structure. This covers, for example, role
overload (i.e., level of difficulty or amount of work
exceeding capacity) and role conflict (i.e., contradictory
and incongruent role requirements) (Rizzo et al. 1970;
Tarafdar et al. 2007). The social environment may also
affect the employees’ stress levels (e.g., social pressure to
use specific system features) (Edwards 1998).
While stressors from all four categories can each induce
stress in the users, it is important to highlight that tech-
nology-related stressors have been found to exacerbate the
others. For example, Tarafdar et al. (2011) found that due
to the pervasiveness of ICT in human society, techno-in-
vasion amplifies task-related and organizational stressors
resulting in increased round-the-clock stress levels – at
work and at home. More generally, packaged application
systems (as a major part of the technological environment,
see Fig. 1) may constitute a source of stress for six main
reasons [see the sources cited in Fischer and Riedl (2015,
p. 1462), Ragu-Nathan et al. (2008) and Tarafdar et al.
(2007, 2011)], namely techno-overload, techno-invasion,
techno-complexity, techno-insecurity, techno-uncertainty,
and techno-unreliability. It is important to note that the first
five factors were derived by Ragu-Nathan, Tarafdar, and
colleagues, and this list of five factors was later comple-
mented by a sixth factor, namely techno-unreliability (de-
fined as ‘‘users face system malfunctions and other IT
hassles’’, see Fischer and Riedl 2015, p. 1462). It is further
of importance that this sixth factor, while (surprisingly) not
conceptualized as TS creator in Ragu-Nathan et al. (2008)
and Tarafdar et al. (2007, 2011), has been shown to con-
stitute a major source of TS throughout the entire history of
TS research. Specifically, since the 1980s up until today
overwhelming evidence in the TS literature [see, for
example, Brod (1984, p. 43), Weil and Rosen (1997, p. 5),
or Ayyagari et al. (2011), and several studies reviewed in
Riedl (2013)] has conceptually substantiated and/or
empirically shown that techno-unreliability may lead to
notable stress reactions in users.
2.3 Stress
Stress manifests in neurophysiological changes in the body,
which usually set on before conscious stress perception
(e.g., Riedl 2013; Tams et al. 2014b). In the context of
SSAESes, a variety of these changes can be measured to
assess stress correlates. Cortisol, a stress hormone released
by the adrenal glands in response to stimulation by the
hypothalamus, plays a critical role in internal stress pro-
cesses. Cortisol can be assessed by means of saliva samples
and provides a well-established measure for increased stress
levels (Dickerson and Kemeny 2004). For example, Riedl
et al. (2012) showed that cortisol levels significantly
increased in response to system breakdown. Therefore, in
combination with context data, it is a valuable measure for
assessing the users’ stress levels in offline analysis. More-
over, a number of other neurophysiological stress parame-
ters can be assessed with online measurements. These
include pupil dilation, heart rate, heart rate variability,
mouse pressure, muscle tension, pulse transit time, and skin
conductance, which, based on modern sensor technology
and advances in battery technology, can be continuously
and unobtrusively assessed over periods of several days
(Hancock and Szalma 2007; Schaaff et al. 2012; Riedl
2013; Zhai and Barreto 2006). Therefore, an assessment of
these parameters allows enterprise-system-context data to
be mapped with neurophysiological stress data in order to
make systems stress sensitive and to trigger context-sensi-
tive interventions (cf. Guideline 3 in the next section).
2.4 Consequences
When investigating stress, it is critical to take into account
its possible consequences (Hancock and Warm 1989);
therefore, assessing specific indicators for consequences
should also play an important role in SSAESes (see Fig. 1,
right). First, according to the Yerkes and Dodson (1908)
law, the relationship between physiological arousal and
performance resembles an inverted U-shaped curve,
whereat arousal levels that are too low or too high impair
performance. Therefore, in order to maximize perfor-
mance, humans should aim at reaching a task-dependent
midrange arousal level that balances detrimental influences
of under-arousal and over-arousal (Kaufman 1999; Han-
cock and Szalma 2007).
Moreover, research on users’ physiological stress reac-
tions suggest, or even directly show (e.g., Arnetz and Berg
1996), that TS may have detrimental effects on well-being
and health (e.g., Arnetz and Wiholm 1997; Boucsein 2009;
Maier et al. 2015b; Riedl et al. 2012). For example, stress
has been linked to chronic headaches, burnout, obesity,
stroke, and cardiovascular diseases (e.g., Bakker et al.
2005; De Kloet et al. 2005; McEwen 2006). SSAESes can
help the user to link private health issues to his/her moni-
tored stress data.
In addition to these severe health issues, stress has been
shown to decrease performance and productivity, which in
turn has detrimental effects on the organizations’ overall
success (e.g., Tarafdar et al. 2007). In particular, several
studies have shown that self-reported stress perceptions
123
M. T. P. Adam et al.: Design Blueprint for Stress-Sensitive Adaptive Enterprise Systems, Bus Inf Syst Eng 59(4):277–291 (2017) 281
