Content uploaded by Grant Munro
Author content
All content in this area was uploaded by Grant Munro on Feb 10, 2018
Content may be subject to copyright.
A realist review of digital cognitive bias modification
Viewpoint
G. M. Munro (PhD Candidate)1, R. Maddison (Professor)1, K. Ball (Professor)1
1 Institute for Physical Activity and Nutrition, Deakin University,
221 Burwood Hwy, Burwood VIC 3125, Australia.
email: gmmunro@deakin.edu.au
tel: +64 223 332 663.
A realist review of digital cognitive bias modification
interventions: current challenges and future
prospects
Highlights
•mHealth are complex interventions that defy prediction or control.
•User engagement in complex intervention depends on dynamic interactions
between personality traits and VUCA contexts.
•‘Traditional’ systematic reviews fail to identify such contexts.
•Realist synthesis of emergent models that combine trait and complexity theory
may help guide future E-Rehabilitation research.
Abstract
Background: Cardiovascular disease (CVD) is the leading cause of morbidity and
mortality globally. While the proliferation of mobile health (mHealth) apps promises
to revolutionise delivery of care, their ability to foster health behaviour change
remains a challenge. Research has suggested digital cognitive bias modification
(dCBM) interventions may improve outcomes.
Objective: To review evidence supporting the current use of dCBM.
Methods: A realist review method was chosen as it is a theory-driven approach that
seeks to explore complex mechanisms underlying program effectiveness. A scoping
literature search in CINAHL, PubMed, and Scopus within ten years (2007 to 2017)
identified 5 articles of dBCM strategies for people living with chronic disease..
Narrative content analysis methods were applied to extract reported correlations
between behaviour change models and health behaviour changes related to diet,
exercise, alcohol and drug use, medication adherence, tracking physiological
symptoms, and improved psychosocial management of stress.
Results: Findings showed limited use of behaviour change theory in dBCIs for CVD.
In the few apps that drew on theory, most used Bandura’s Social Cognitive Theory
1
A realist review of digital cognitive bias modification
(SCT) and associated concept of self-efficacy (i.e., one’s self-confidence to execute
requisite actions to satisfy situational demands). However, scant reporting of dose-
dependent links between theories and health outcomes prevented generalisations to be
made. Further, ubiquitous use of SCT-based models was problematic, due to growing
evidence that behaviour change is part of a nonlinear complex adaptive system that
defies prediction or control.
Conclusion: Despite proliferation of dBCIs to support CR, current research fails to
effectively integrate evidence-based behaviour change theories and respond to change
as a fundamentally unstable and unsteady process. An alternative dynamic behaviour
change model is presented to help researchers identify change potentials, and
intertemporal touchpoints from which trait-tailored exposures are most likely to
benefit people seeking CR.
Keywords: behaviour change; cardiovascular disease; complex adaptive systems,
digital intervention; realist review/synthesis; theory.
Introduction
Background
Cardiovascular disease (CVD) is the leading cause of death and disability globally –
with low and middle income countries experiencing the highest mortality rates (i.e.,
80%) (World Health Organization 2016). Ubiquitous use of mobile phones has
prompted healthcare providers to develop mobile health (mHealth) systems to connect
people to services and help them self-manage conditions (World Health Organization
2016). This shift toward mHealth is evidenced by the rapid proliferation of large-scale
digital behaviour change interventions (dBCIs). dBCIs are defined as any Web- or
smartphone-based application (app) accessed and taking input from people (Lathia et
al. 2013). The main advantage of dBCIs over previous centre- or home-based
rehabilitation treatments is that they can be scaled to reach large numbers of people.
Thereby resulting is significant cost savings to healthcare providers and contextually-
relevant behavioural prompts to users seeking change (Lathia et al. 2013). Despite the
potential of mHealth and dBCIs, most studies report limited adherence and small
effect sizes. Research has also suffered from an overemphasis on text or short-
message service (SMS) (Hall et al. 2015) and diabetes (Free et al. 2013). CVD
prevention in particular has been overlooked due to concerns that older people (65
years+) with CVD are less likely to use mobile phones (Widmer et al. 2015). Limited
effectiveness has prompted many researchers to argue that future dBCIs may be more
effective if grounded in appropriate behaviour change theory (Michie et al. 2011).
Objectives
Recent systematic reviews have examined the impact of behaviour change techniques
(BCTs) and related theories in dBCIs for CVD prevention. All the reviews concluded
dBCIs apply BCTs and related models in various ways to target modifiable CVD risk
factors (Burke et al. 2015, Neubeck et al. 2015, Goodwin et al. 2016, Pfaeffli Dale et
al. 2016, Winter et al. 2016). One key challenge for dBCIs and mHealth generally is
linking theoretical components to health outcomes (Goodwin et al. 2016, Pfaeffli Dale
et al. 2016). Researchers are also well aware of the heterogeneity in dBCIs, and the
fact that people experience different risk factors differently. The inability of current
studies to explain such heterogeneity should thus prompt future behaviour change
2
A realist review of digital cognitive bias modification
researchers to examine contextual mediators of outcomes as well as the active
components (Bouton 2014, Widmer et al. 2015, Goodwin et al. 2016, Pfaeffli Dale et
al. 2016). Currently, it is unclear how and why certain dBCIs are effective and in what
contexts they likely to be most successful (Levati et al. 2016). Future reviews would
thus benefit from examining the underlying theoretical rationale guiding behaviour
change models and techniques in the field. This review aims to develop such a
theoretical perspective. The authors thus contribute to the current literature gap by
adopting a theory-focused approach to understanding dBCIs in CVD groups. This
realist review addresses the following key research questions: What active
components in dBCIs are responsible for mediating health behaviour change? What
are the unmet challenges of the components? Can a new dynamic theory of behaviour
change be articulated and refined, and if so, what are the prospects for program
implementation? To answer the above, the authors develop a theory-based realist
review of the literature {Citation}.
Methods
Table : Summary of database search terms, limiters, and inclusion criteria
Databases Scopus and PubMed
Search terms “future self” OR “future self-continuity” OR “self-continuity”
AND behavior OR delinquency OR procrastination
Search limiters Subject: human; Language: English; Date: 2007-2017
Inclusion criteria Any digital intervention delivered via web- or phone-based
application (app), accessed and taking input from people in
support of cognitive bias modification.
Reasons for exclusions
•Study not identified as
intervention
•Study focused on other disease
Excluded from
Database searches in Scopus (n
= 102) PubMed (n = 68)
Records after initial screening
Full-text articles
assessed for eligibility
Excluded from
Figure 1: Article search flow diagram
3
A realist review of digital cognitive bias modification
Rationale for using realist review
The realist review is a new method of systematic review designed for complex
interventions, guided by three social principles of realism (i.e., causal explanations are
achievable; social reality is an interpretative reality of social actors; and social actors
evaluate their social reality) (Gerard 2005). Rationale for adopting such methods is to
‘…articulate underlying programme theories and then to interrogate the existing
evidence to find out whether and where these theories are pertinent and productive’
(Pawson 2006). Such methods appear to be highly responsive to the complexities of
policy-based interventions (Pawson et al. 2005). While systematic in its approach, the
realist method is also uniquely iterative in the sense that it draws on diverse forms of
evidence, involves broad stakeholder collaboration, and targets theory development.
To achieve such objectives, this research draws on Pawson’s 5-stage realist approach
to: (i) clarify scope; (ii) search evidence; (iii) appraise and extract data; (iv) draw
conclusions; and (iv) disseminate narrative (Pawson et al. 2005).
Table : Reviews of behavioural theories underlying dBCIs for CVD
Author Design Location Review Intervention Behavioural models
Burke et
al. 2015 11 RCTs USA Literature
review mHealth for
CVD prevention - Social Cognitive
Theory (9%)
Goodwin
et al.
2016
22 RCTs UK Systematic
review and
meta-analysis
dBCIs for CHD
prevention - Social Cognitive
Theory (28.5%)
- Trans-theoretical
Model/Stages of
Change (4.5%)
- Health Belief Model
(4.5%)
- Control Theory
(4.5%)
- Stress Theory (4.5%)
- Self-regulation
Theory (4.5%)
Neubeck
et al.
2015
4 mixed
method
designs
Australia Literature
review Mobile apps for
CVD prevention - Social Cognitive
Theory (25%)
- Social Influence
Theory (25%)
- Operant
Conditioning, Self-
Recognition (25%)
Pfaeffli
et al.
2015
7 RCTs New
Zealand Systematic
review mHealth for
CVD prevention - Social Cognitive
Theory (28.5%)
- Trans-theoretical
Model/Stages of
Change (14.2%)
Winter et
al. 2016 240
mixed
method
designs
USA Comprehensive
review dBCIs for CVD
prevention - Social Cognitive
Theory (21.6%)
- Trans-theoretical
Model/Stages of
Change (9.5%)
- Theory of Planned
Behavior/Reasoned
Action (6.2%)
4
A realist review of digital cognitive bias modification
Scoping the literature
A scoping review was performed on PubMed and Scopus in October 2015 and later
updated in December 2016 (Figure 1). The search was limited to studies published in
English over ten years (i.e. 2006 to 2016). Search terms were identified through a
preliminary literature review in consultation with health informatics specialists at the
National Institute of Health Informatics (NIHI), University of Auckland. Search
strategies were refined and a more comprehensive search was conducted, including a
broader range subject headings and terms (Table 1). The final search was limited to
published reviews of dBCIs for CVD prevention (unpublished grey literature was not
considered). Two independent reviewers screened titles and abstracts to determine
eligibility (Table 2). A total of 117 articles were identified. After the removal of 106
articles from reading titles/abstracts, 11 were included for the final screening. Of the
11 articles initially identified for review, only five met all three inclusion criteria.
Results
Study characteristics
All 5 studies were published recently (2015-2016) and focused on the use of
behaviour change models to guide dBCI development in CVD groups: 2 were
literature reviews while the remaining 3 were either a systematic review,
comprehensive review, or meta-analysis. Studies were conducted across a range of
locations, including the USA (n=2), the UK (n=1), Australia (n=1) and New Zealand
(n=1). More than half of the studies examined RCT designs (n=3) while the remaining
2 analysed a variety of study methods. In general, studies reported small effect sizes
and inconsistencies due to poor method designs and ineffective measures of app
components (Burke et al. 2015, Neubeck et al. 2015, Goodwin et al. 2016, Pfaeffli
Dale et al. 2016, Winter et al. 2016). For example, Neubeck et al. (2015) found that
while some apps supported promotion of CVD self-management, data sets were
limited and long term outcomes unclear. Burke et al. (2015) found that while
behaviour theories were often mentioned; most mHealth programs had limited
theoretical or empirical basis. Limitations meant healthcare providers had scant
understanding of which apps they should recommend and in what context (Burke et
al. 2015).
Winter (2016) found only 52% of studies used behavioural theories to guide
intervention. The most frequently used was Social Cognitive Theory (Bandura 1986),
followed by the Trans-Theoretical Model (Prochaska and Velicer 1997), and Theory
of Planned Behaviour (Ajzen 1991). While each behaviour change model applied
various approaches to behaviour change, all adopted the core assumption outlined in
Bandura’s (1986) SCT model. More than half of the studies (n=3) applied Michie’s
BCT taxonomy (Michie et al. 2013) to identify the techniques underlying behaviour
change models/theories (Goodwin et al. 2016, Pfaeffli Dale et al. 2016, Winter et al.
2016). Goodwin et al. (2016) reviewed 22 papers and found the most commonly used
BCTs were automated text messaging, provision of information, and goal setting. A
review of 7 RCTs by Pfaeffli Dale et al. (2016) found that while CR programs were
partially delivered by text messaging, only 3 were effective at improving adherence to
medication and 2 increased physical activity. No positive effects were observed on
dietary behaviour or smoking. Simple text messaging interventions appeared to be
most effective; however, no dose-dependent relations between duration or targeted
BCTs were found (Pfaeffli Dale et al. 2016).
5
A realist review of digital cognitive bias modification
There was also consensus that while Michie’s BCT taxonomy helped researchers
understand which BCTs were most commonly used, critical information about their
active components were unclear (Goodwin et al. 2016, Pfaeffli Dale et al. 2016,
Winter et al. 2016). Lack of theoretical clarity combined with a heterogeneity of
reported studies suggest future research may need to: (1) examine theoretical
constructs of behaviour change more critically, and (2) develop more effective
research methods to identify dose-dependent relations between active components and
behaviour change outcomes (Widmer et al. 2015, Goodwin et al. 2016, Pfaeffli Dale
et al. 2016).
Narrative synthesis: semi-predictable behaviour patterns
Despite the heterogeneity of studies and lack of dose-dependent reports, the review
identified a common theme across the literature that health promotion is directly
related to the presence or absence of user engagement (Short et al. 2015). Variant
meanings aside, user engagement typically refers to a process and product of system
interaction which evolves over time, depending on the ability of key aesthetic and
usability features to foster evaluation and experience (O’Brien and Toms 2008).
Further examination of the literature revealed three semi-predictable patterns that
determine user engagement: (i) the presence and impact of self-evaluation personality
traits; (ii) generic use of BCTs and models; and (iii) personalised computer-tailoring
that attempt to adapt to changes in self-evaluation over time (Short et al. 2015).
Pattern I: Health-supportive self-evaluation traits
A narrative synthesis of findings confirmed previous conclusions from user
engagement studies that not all systems engage users equally (O’Brien and Toms
2010). This is simply because user engagement typically depends on differences in
personal identity traits (Vinciarelli and Mohammadi 2014, Tkalčič et al. 2016).
Indeed, such findings are not new. Rather, they confirm eight decades of well-
established lexical analysis that differences in people’s self-evaluation profiles
critically affect how confident they feel about their present and future selves (Judge
and Bono 2001, Rushton and Irwing 2011, Cloninger 2013).
This self-evaluation metatrait domain, commonly referred to as the General Factor of
Personality (GFP) or more recently ‘creative’ profile appears at the apex of an
emergent trait hierarchy (Cloninger and Cloninger 2013, Saucier and Srivastava
2015). Variant definitions aside, this multidimensional creative profile (henceforth
Creativity) is a unique synthesis of: Dynamism or self-expression (i.e., boldness,
receptivity); Stability or self-regulation (i.e., empathy, constraint); and Emotionality
or self-awareness (i.e., psychological resilience, calmness) (Judge and Bono 2001,
Rushton and Irwing 2011, Cloninger 2013). Researchers argue that the key advantage
of Creativity in health promotion may lie in its ability to harness certain ‘creative’
evolutionary processes to ensure optimal reproduction, survival, and growth (Musek
2007, Rushton and Irwing 2011). Creativity during adolescence predicts positive real-
world consequences that continue throughout life (Trzesniewski et al. 2006).
Figure 2: Emergent trait hierarchy, from Saucier & Srivastava (2015)
6
A realist review of digital cognitive bias modification
As a result, high Creativity cohorts tend to be ranked as more socially desirable and
socially effective than low scorers (Erdle et al. 2010, Dunkel and Van der Linden
2014). In addition, low Creativity groups tend to adopt unhealthier behaviours such as
poor diet, physical inactivity, and drug use (Belcher et al. 2014, Israel et al. 2014).
Habits that often lead to metabolic syndrome (Friedman and Martin 2011), heart
disease (Steptoe and Molloy 2007, Denollet et al. 2010) and CVD (Jokela et al.
2014). Findings are not based purely on lexical analysis. For example, studies of
heart-rate variability (HRV) found Creativity predicted improved autonomic balance
as measured by the ratio of low frequency (sympathetic) to high frequency
(parasympathetic) activity (Zohar et al. 2013). A similar study correlated high ECG
values to improvements in Stability and Dynamism (i.e., high Agreeableness and
Extraversion) and low ECG values with low Emotionality (i.e., high Neuroticism) and
low Dynamism (i.e., low Extraversion) (Koelsch et al. 2012).
Health-supportive traits are not set in plaster as theorists once argued (Costa and
McCrae 2006) but change continually, either as a result of natural hormonal changes
or due to life experiences (Roberts and Mroczek 2008). For instance, brain damage
from stroke events can result in decreased Dynamism (Stone 2004, Jokela et al. 2014)
whereas stress-evoked disruption of the hypothalamic-pituitary-adrenal axis (HPA)
can decrease Stability and Emotionality, rendering patients hostile and neurotic to
changes (Sher 2005, Denollet et al. 2010). These changes in core self-evaluation traits
(as represented in the GFP of Creativity) can seriously impair an individual’s sense of
self over time (i.e., future self-continuity), including their cognitive ability to delay
gratification (i.e., temporal discounting). For example, people high in Dynamism (i.e.,
Extraversion) are more sensitive to positive emotions and thus more susceptible to
immediate rewards over larger future gains. Contrastingly, high Emotionality (i.e.,
psychological resilience) scorers exhibit increased self-awareness and thus less
temporal discounting (Hirsh et al. 2008). However, if Extraversion is accompanied by
elevated Openness/Intellect traits, the individual tends to exhibit more self-control and
can thus more effectively override instinctual urges. Similarly, an inability to control
7
A realist review of digital cognitive bias modification
urges, may be impaired by low Emotionality and increased trait anxiety (i.e.,
Neuroticism) and undermine the delayed gratification.
While self-reported changes in Creativity may be subject to confirmation bias
(Shephard 2003), such changes can be empirically validated using event-related
functional magnetic resonance imaging (fMRI). For example, two studies found
greater differences in the medial prefontal cortex (mPFC) and rostral anterior
cingulate (rACC) activation for self-judgments compared to other-judgments. In
others words, positive correlations were found between the extent to which the future
self ‘looked’ like another person on a neural level, and the degree to which
participants chose to consume smaller present rewards over larger future rewards
(Kelley et al. 2002, Hershfield et al. 2009). Evidence from controlled lab studies and
real world experiments thus confirm the idea that people with core self-evaluation
traits (i.e., Creativity) experience greater future self-continuity and are more willing to
wait for future rewards. Reiterating the central role future self-continuity plays in both
mediating poor decision-making and mitigating self-defeating behaviours.
Interventions promoting healthy trait profiles
Interestingly, health traits can also be promoted via various exposure techniques and
intervention – (e.g., fasting, exercise, and mindfulness). The benefits of which involve
initiating a healthy stress response (SR) (i.e., eustress) on the individual (Selye 1976).
This ‘hormetic’ eustress process activates a cascade of reactions that drives neural
plasticity, thus enabling individuals to adapt to environmental changes (Mattson and
Calabrese 2010). However, too much stimulus renders the SR maladaptive, producing
insulin resistance and inflammation (Figure 3). First, intermittent fasting that aims to
restrict caloric intake has been linked to improvements in Stability (i.e., self-
regulation). Clinical trials of people following a prescribed therapeutic fasting
protocol reported increased vigilance and mood stability (Fond et al. 2013).
Brandhorst et al. (2015) found people on the fasting-mimicking-diet (FMD) reported
increased cognitive performance and mood stability and additional benefits in weight,
cholesterol, blood pressure, and CVD risk (Brandhorst et al. 2015, Weiss et al. 2016).
Second, physical activity (PA) or indeed any exercise intervention that raises energy
expenditure as a result of bodily movement has been linked to positive changes in
Creativity (i.e., self-evaluation) and CVD risk. A large population-based study in the
Netherlands found adults who participated in a minimum of 60 min of exercise
weekly reported higher Creativity scales and lower rates of depression and anxiety
than non-exercisers (De Moor et al. 2006). Findings have been replicated in two
longitudinal studies which also found exercise helps people maintain more resilient
traits (i.e., less declines in Dynamism, Stability, and Emotionality) (Stephan et al.
2014). Third, mindfulness-based cognitive therapy (MBCT) (Kabat-Zinn and Hanh
2009) which incorporates elements of cognitive-behavioural therapy with
mindfulness-based stress reduction is found to improve Emotionality trait facets (i.e.,
anxiety and psychological resilience) (Baer 2003) while reducing observed CVD risk
factors such as elevated blood pressure and weight gain (Parswani et al. 2013).
Figure 3: Biology of the hormetic dose, from Mattson and Calabrese (2010)
8
A realist review of digital cognitive bias modification
Pattern II: generic use of behaviour change models
The ubiquitous use of SCT-guided BCTs and related models suggests SCT has
received broad clinical acceptance among dBCI researchers and operates as the
default theoretical structure to guide practice (Michie et al. 2011) (Figure 4). Model
variants notwithstanding, all BCTs and related models converge on the idea that self-
regulation and behaviour change depend on a person’s conscious ability to self-
monitor schema, evaluate whether conditions reflect their desired goals, and take
appropriate action across timelines and contexts when necessary (Bandura 1986).
Implicit within SCT is the concept of self-efficacy – i.e., one’s self-confidence to
execute requisite actions to satisfy situational demands (Bandura 1977). SCT
proponents argue self-efficacy can be increased via the practice of social cognitive
tasks: (i) performance accomplishments (learning by doing); (ii) vicarious learning
(observing others); (iii) informational persuasion (influence of experts); and (iv)
limiting emotional arousal (anxiety reduction). SCT thus contends the successful
performance of social cognitive tasks affords mastery over difficult situations,
boosting one’s self-efficacy and providing people with powerful sources of efficacy
expectations and outcomes.
Despite clinical acceptance and broad use, the linear rationale of SCT-based models
contradict evidence that behaviour is fundamentally a complex adaptive system
(CAS) (Resnicow and Vaughan 2006, Resnicow and Page 2008) situated within
volatile, unpredictable, complex, and ambiguous (VUCA) worlds (Rouse 2008, The
Health Foundation 2010, Kannampallil et al. 2011, Marchal et al. 2014). Indeed,
healthcare operates as a CAS from many perspectives, including biology (Noble
2006), precision medicine (Hood and Flores 2012), primary care (Miller et al. 1998),
nursing (Anderson et al. 2003), management (Plsek and Wilson 2001), addiction
(Randle et al. 2015), research (Marchal et al. 2014), and behaviour change (Resnicow
and Vaughan 2006). The notion that behaviour change is a CAS is perhaps not
surprising given the evidence that motivating and sustaining change is an inherently
unstable and unsteady process (Bouton 2014). Such instability is represented in at
least two findings across the literature.
9
A realist review of digital cognitive bias modification
Figure 4: Social cognitive models of behaviour change
First, despite the proliferation of interventions to promote health behaviours, many
people (40-50%) abstain or enter relapse as a result of ‘sudden gains’ or ‘mini-
epiphanies’ (Tang and DeRubeis 1999, Kelly et al. 2005). Second, despite major
advances in personalisation and computer-tailoring, most dBCIs have been
surprisingly resistant to change (Short et al. 2015), as indicated by small effect sizes,
low user engagement, and high drop-out rates (Brouwer et al. 2011, Davies et al.
2012, Kelders et al. 2012, Free et al. 2013). Research attempts to either predict on
control change also appear to be flawed (Kurtz and Snowden 2003, Resnicow and
Vaughan 2006), simply because the same antecedent variables have multiple
outcomes (i.e., multifinality), whereas different antecedent variables can have the
same outcome (i.e., equifinality) (Cloninger et al. 1997). Such findings imply change
is a more dynamic and nuanced process than previously assumed (Hughes et al.
2014).
Pattern III: personalised tailoring that adapts to changes over time
In response to the complexities, developers have turned towards computer-tailoring
materials as a way of personalising systems to individual needs (Dijkstra 2016).
Tailored materials use one or all of the following 3 elements: (i) personalisation (i.e.,
using an person’s name or other identifiable aspects; (ii) adaptation (i.e., presenting
information relevant to users’ characteristics such as age or gender), and (iii) feedback
(i.e., responding to an user’s psychological or behavioural state) (Dijkstra 2008).
While casual mechanisms are not well understood, studies show computer-tailored
materials have superior influence on users’ information processing and behaviour
change that non-tailored data (Kreuter et al. 2013). Despite this evidence, key
engagement features such as aesthetics, expectations, and skills have been routinely
overlooked (Short et al. 2015). Such oversights are likely the confluence of two
factors.
First, user data is inherently difficult to measure and interpret, even with access to
server data measures (e.g., log-ins, time spent, pages viewed). For example, a dBCI
may require users to access material in a prescribed order (no skipping), yet some may
skim while others read material thoroughly, creating different levels of engagement
(Schubart et al. 2011). Second, decision-making and behaviour change is a complex
process that is virtually impossible to predict or control (Kurtz and Snowden 2003,
Resnicow and Vaughan 2006, Resnicow and Page 2008). People behave differently
10
A realist review of digital cognitive bias modification
and express multiple identities over time and space (e.g., either as children, siblings,
spouses, or parents). People also rarely follow prescribed rules but impose structure
on their interactions either by conforming to social norms or acting on individual
decisions of free will. Finally, people are aware of large-scale patterns due to their
ability to communicate concepts via language and interact with people remotely
(Kurtz and Snowden 2003). Findings thus confirm previous user engagement studies
that not all systems engage users equally (O’Brien and Toms 2010) because
engagement depends on key personal identity differences (Vinciarelli and
Mohammadi 2014, Tkalčič et al. 2016).
Dissemination: Emergent Behaviour Change (EBC) model
VUCA contexts, emergent behaviours
Findings identified in this review have prompted some researchers to develop
dynamic behaviour change models to help guide intervention designs within VUCA
contexts (Penprase and Norris 2005, Nahum-Shani et al. 2015, Munro 2016). While
broad in concept, dynamic models converge on the idea that change is an adaptive,
emergent, spontaneous, and social process that occurs at the edge of chaos (i.e., point
of maximum complexity) (Penprase and Norris 2005). Models also focus on how
factors and effects that occur at a higher level (e.g., a year) can affect factors/effects
that occur later, at a lower level (e.g., a minute), either directly, or by facilitating
change in trait-related factors (i.e., behaviours and disease symptoms) (Nahum-Shani
et al. 2015, Munro 2016).
To disseminate the two key findings from this review to behaviour change research
communities, authors synthesise well-evidenced models such as the emergent trait
hierarchy (Saucier and Srivastava 2015) with VUCA models of the adaptive cycle
(Gunderson and Holling 2001). The proposed unified framework, referred to hereafter
as the Emergent Behaviour Change (EBC) model (Figure 5) aims to inform
researchers when ‘hard’ evidence can be applied to decision-making processes, and
when change predictions are futile, i.e., contexts where emergent flexibility of new
ideas and behaviours are most likely to thrive (Penprase and Norris 2005, Resnicow
and Page 2008, Munro 2016).
Unlike SCT-based models such as the COM-B (Michie et al. 2011) which overlook
intertemporal factors; the EBC model explicitly focuses on how temporal changes
determine two fundamental research issues central to behaviour change: (i) where a
problem or situation is positioned within the context of sense-making frames (e.g.,
simple-complicated-complex-chaotic) (Kurtz and Snowden 2003); and (ii) the most
appropriate design development tools that reflect that situation (Marchal et al. 2014).
The EBC model thus points out where and when ‘hard’ evidence can be applied in
decision-making, and when researchers need to accept that evidence is unlikely to
predict or control outcomes (Resnicow and Vaughan 2006, Resnicow and Page 2008).
Figure 5: Nonlinear model of emergent behaviour change
11
A realist review of digital cognitive bias modification
Conclusions
Potential of dynamic behaviour change models
The current proposal for a dynamic behaviour change model (EBC) is prescient in the
sense that it better responds to behaviour change within VUCA worlds. By fusing trait
theory and adaptive cycle models into a holistic framework, EBC can help researchers
identify when and where predictable change mechanisms are likely to be effective and
deliver trait-targeted exposures accordingly. Tailored exposures (e.g., fasting,
exercise, mindfulness) are likely to initiate a ‘perfect storm’ of intrapsychic events
that encourage patients to be pushed to the edge of chaos, a place where flexibility of
new ideas and behaviours can flourish. A point of maximum complexity that promotes
creative capacity in motivated individuals (i.e., high adaptivity, inventiveness, and
resilience). The main benefit of adopting a bottom-up EBC model may lie in its ability
to adapt to spatiotemporal changes affecting users, rather than specifically targeting
changes of clinical interest (i.e., a clinician-centric top-down approach). As a result,
the bottom-up emergent change model outlined in this review (EBC) is fundamentally
person-centric in the sense that it allows key change processes to become personally
relevant to users as they move through different contexts and timelines. Reported
efficacy of such bottom-up methods suggest similar tailored exposures may be easily
transferrable to mHealth programs that target CVD prevention.
Challenges
Despite the potential of the EBC model to inform development of trait-tailored
exposures (e.g., fasting, exercise, mindfulness) and creative capacity for health
behaviour change, key limitations should be noted. First, current 3-meal-a-day eating
habits combined with high daily calorie intake of large amounts of simple sugars and
saturated fats are so socially engrained that integration of fasting cycles may be
difficult. Second, physical exercise is no longer required for most professions, and
combined with ease of transportation, has typically rendered fitness an elective chore
for many. Third, providing cognitive training to boost trait profiles in people with
CVD may be difficult, especially among patients with cognitive impairment. In light
of these challenges, implementing trait-tailored exposures to surmount
aforementioned barriers will be a Herculean task (Mattson 2015). That being said,
significant CVD risk reductions of up to 75% have been reported in countries such as
France and Finland as result of targeted prevention, suggesting dramatic systemic
changes are indeed possible when correctly targeted. The authors thus argue that
12
A realist review of digital cognitive bias modification
dynamic trait-based models such as the ETM (Figure 5) present researchers with
novel ways to foster core self-evaluation traits and behaviour changes within VUCA
worlds (Resnicow and Vaughan 2006, Munro 2015). Such frameworks may also help
researchers generate cardiac personality signatures (CPS) to diagnose cardiac risk and
co-morbid personality disorders (e.g., anxiety, depression) due to ANS disruption
(Koelsch et al. 2012). Signatures which may lead to more efficient risk-stratification,
risk-prevention, and pre-clinical diagnostics (Koelsch et al. 2012). Adoption of such
personalised methods may also help counter some of the excessive diagnostic co-
occurrence and lack of discriminant validity that currently impacts general practice
(Wakefield, 2015), and CVD healthcare specifically (Mensah and Collins 2015).
Summary
Researchers argue that future success of digital CR programs will likely depend on
rigorous integration of behaviour change techniques and related models. However,
this realist review found no evidence for this assumption. Conclusions are based on
three key findings within the literature. First, most CR apps are not rooted in
behaviour change theories. Second, of the few apps that do apply such models, dose-
dependent links between components and outcomes are unclear. Third, and potentially
most problematic is the broad clinical acceptance and use of current SCT-based
behaviour change models. This is concerning because such linear models fail to
respond to the nonlinear nature of health behaviour change as a complex adaptive
system that defies prediction or control. In response to these challenges, researchers
are developing dynamic change models that better respond to intertemporal change
within VUCA contexts. This shift is based on evidence that the most consistent
predictor of health is the multidimensional metatrait of Creativity. A core self-
evaluation trait that combines Dynamism (i.e., boldness, receptivity), Stability (i.e.,
empathy, constraint), and Emotionality (i.e., psychological resilience). Proposals are
thus presented for an alternative change model that synthesises trait theory with
VUCA models of the adaptive cycle. The proposed model aims to inform researchers
when ‘hard’ evidence can be applied to decision-making, and when predictions are
impossible, i.e., contexts where emergent flexibility of new ideas and behaviours are
most likely to thrive.
Acknowledgements and disclosure statement
The authors report no conflicts of interest related to this review. Funding source for
the review was solely the responsibility of the primary author.
Conflicts of Interest
None declared.
Abbreviations
CAS: Complex adaptive system
CNS: Central nervous system
COM-B: Capability, opportunity, motivation, and behaviour model
CR: Cardiac rehabilitation
CVD: Cardiovascular disease
dBCI: Digital behaviour change intervention
EBC: Emergent behaviour change
GFP: General factor of personality
HRV: Heart rate variability
13
A realist review of digital cognitive bias modification
HPA: Hypothalamic-pituitary-adrenal axis
RCT: Randomized controlled trial
SCT: Social cognitive theory
References
Ajzen, I., 1991. Theories of Cognitive Self-RegulationThe theory of planned behavior.
Organizational Behavior and Human Decision Processes, 50 (2), 179–211.
Anderson, R.A., Issel, L.M., and McDaniel Jr, R.R., 2003. Nursing homes as complex
adaptive systems: relationship between management practice and resident
outcomes. Nursing research, 52 (1), 12.
Baer, R.A., 2003. Mindfulness training as a clinical intervention: a conceptual and
empirical review. Clinical Psychology: Science and Practice, 10, 125–143.
Bandura, A., 1977. Self-efficacy: Toward a unifying theory of behavioral change.
Psychological review, 84, 191.
Bandura, A., 1986. Social foundations of thought and action: A social cognitive
theory. PrenticeHall series in social learning theory, 1, 617.
Belcher, A.M., Volkow, N.D., Moeller, F., and Ferré, S., 2014. Personality traits and
vulnerability or resilience to substance use disorders. Trends in Cognitive
Sciences, 18, 211–7.
Bouton, M.E., 2014. Why behavior change is difficult to sustain. Prev Med, 68, 29–
36.
Brandhorst, S., Choi, I.Y., Wei, M., Cheng, C.W., Sedrakyan, S., Navarrete, G.,
Dubeau, L., Yap, L.P., Park, R., Vinciguerra, M., Di Biase, S., Mirzaei, H.,
Mirisola, M.G., Childress, P., Ji, L., Groshen, S., Penna, F., Odetti, P., Perin,
L., Conti, P.S., Ikeno, Y., Kennedy, B.K., Cohen, P., Morgan, T.E., Dorff, T.B.,
and Longo, V.D., 2015. A periodic diet that mimics fasting promotes multi-
system regeneration, enhanced cognitive performance, and healthspan. Cell
Metab, 22 (1), 86–99.
Brouwer, W., Kroeze, W., Crutzen, R., de Nooijer, J., de Vries, N.K., Brug, J., and
Oenema, A., 2011. Which intervention characteristics are related to more
exposure to internet-delivered healthy lifestyle promotion interventions? A
systematic review. Journal of Medical Internet Research, 13 (1), e2.
Burke, L.E., Ma, J., Azar, K.M., Bennett, G.G., Peterson, E.D., Zheng, Y., Riley, W.,
Stephens, J., Shah, S.H., Suffoletto, B., and others, 2015. Current science on
consumer use of mobile health for cardiovascular disease prevention a
scientific statement from the American Heart Association. Circulation, 132
(12), 1157–1213.
Cloninger, C.R., 2013. What makes people healthy, happy, and fulfilled in the face of
current world challenges? Mens Sana Monographs, 11 (1), 16.
Cloninger, C.R. and Cloninger, K.M., 2013. People create health: effective health
promotion is a creative process. International journal of person centered
medicine, 3 (2), 114.
Cloninger, C.R., Svrakic, N.M., and Svrakic, D.M., 1997. Role of personality self-
organization in development of mental order and disorder. Development and
Psychopathology, 9 (881–906).
Costa, P.T. and McCrae, R.R., 2006. Age changes in personality and their origins:
comment on Roberts, Walton, and Viechtbauer (2006). Psychological Bulletin,
132, 26–28.
14
A realist review of digital cognitive bias modification
Davies, C.A., Spence, J.C., Vandelanotte, C., Caperchione, C.M., and Mummery,
W.K., 2012. Meta-analysis of internet-delivered interventions to increase
physical activity levels. International Journal of Behavioral Nutrition and
Physical Activity, 9 (1), 1.
De Moor, M.H.M., Beem, A.L., Stubbe, J.H., Boomsma, D.I., and De Geus, E.J.C.,
2006. Regular exercise, anxiety, depression and personality: a population-
based study. Preventive Medicine, 42 (4), 273–279.
Denollet, J., Schiffer, A.A., and Spek, V., 2010. A general propensity to psychological
distress affects cardiovascular outcomes evidence from research on the Type D
(distressed) personality profile. Circulation: Cardiovascular Quality and
Outcomes, 3 (5), 546–557.
Dijkstra, A., 2008. The psychology of tailoring-ingredients in computer-tailored
persuasion. Social and Personality Psychology Compass, 2 (2), 765–784.
Dijkstra, A., 2016. Personalization/computer-tailoring in persuasive technology:
tailoring ingredients target psychological processes. In: Proceedings of the
International Workshop on Personalization in Persuasive Technology
(PPT’16), Salzburg, Austria.
Dunkel, C.S. and Van der Linden, D., 2014. Evidence for the general factor of
personality as social-effectiveness. Personality and Individual Differences, 64,
147–151.
Erdle, S., Irwing, P., Rushton, J.P., and Park, J., 2010. The general factor of
personality and its relation to self-esteem in 628,640 Internet respondents.
Personality and Individual Differences, 48 (3), 343–346.
Fond, G., Macgregor, A., Leboyer, M., and Michalsen, A., 2013. Fasting in mood
disorders: neurobiology and effectiveness. A review of the literature.
Psychiatry Res, 209 (3), 253–8.
Free, C., Phillips, G., Galli, L., Watson, L., Felix, L., Edwards, P., Patel, V., and
Haines, A., 2013. The effectiveness of mobile-health technology-based health
behaviour change or disease management interventions for health care
consumers: a systematic review. PLoS Medicine, 10.
Friedman, H. and Martin, L.R., 2011. The Longevity Project: surprising discoveries
for health and long life from the landmark eight decade study. Hay House,
Incorporated.
Gerard, D., 2005. Social Science. McGraw-Hill Education (UK).
Goodwin, L., Ostuzzi, G., Khan, N., Hotopf, M.H., and Moss-Morris, R., 2016. Can
we identify the active ingredients of behaviour change interventions for
coronary heart disease patients? A systematic review and meta-analysis. PLOS
ONE, 11 (4), e0153271.
Gunderson, L.H. and Holling, C.S., 2001. Panarchy: understanding transformations
in human and natural systems. Island Press.
Hall, A.K., Cole-Lewis, H., and Bernhardt, J.M., 2015. Mobile text messaging for
health: a systematic review of reviews. Annual Review of Public Health, 36
(1), 393–415.
Hershfield, H.E., Wimmer, G.E., and Knutson, B., 2009. Saving for the future self:
neural measures of future self-continuity predict temporal discounting. Social
Cognitive and Affective Neuroscience, 4 (1), 85–92.
Hirsh, J.B., Morisano, D., and Peterson, J.B., 2008. Delay discounting: interactions
between personality and cognitive ability. Journal of Research in Personality,
42 (6), 1646–1650.
15
A realist review of digital cognitive bias modification
Hood, L. and Flores, M., 2012. A personal view on systems medicine and the
emergence of proactive P4 medicine: predictive, preventive, personalized and
participatory. New Biotechnology, 29 (6), 613–624.
Hughes, J.R., Solomon, L.J., Naud, S., Fingar, J.R., Helzer, J.E., and Callas, P.W.,
2014. Natural history of attempts to stop smoking. Nicotine & Tobacco
Research, 16 (9), 1190–1198.
Israel, S., Moffitt, T.E., Belsky, D.W., Hancox, R.J., Poulton, R., Roberts, B.,
Thomson, W.M., and Caspi, A., 2014. Translating personality psychology to
help personalize preventive medicine for young adult patients. Journal of
Personality and Social Psychology, 106 (3), 484–498.
Jokela, M., Hakulinen, C., Singh-Manoux, A., and Kivimäki, M., 2014. Personality
change associated with chronic diseases: pooled analysis of four prospective
cohort studies. Psychological Medicine, 44 (12), 2629–2640.
Judge, T.A. and Bono, J.E., 2001. Relationship of core self-evaluations traits--self-
esteem, generalized self-efficacy, locus of control, and emotional stability--
with job satisfaction and job performance: a meta-analysis. Journal of Applied
Psychology, 86 (1), 80–92.
Kabat-Zinn, J. and Hanh, T.N., 2009. Full catastrophe living: using the wisdom of
your body and mind to face stress, pain, and illness. Delta.
Kannampallil, T.G., Schauer, G.F., Cohen, T., and Patel, V.L., 2011. Considering
complexity in healthcare systems. Journal of Biomedical Informatics, 44 (6),
943–947.
Kelders, S.M., Kok, R.N., Ossebaard, H.C., and Van Gemert-Pijnen, J.E., 2012.
Persuasive system design does matter: a systematic review of adherence to
web-based interventions. Journal of Medical Internet Research, 14 (6), e152.
Kelley, W.M., Macrae, C.N., Wyland, C.L., Caglar, S., Inati, S., and Heatherton, T.F.,
2002. Finding the self? An event-related fMRI study. Journal of Cognitive
Neuroscience, 14 (5), 785–794.
Kelly, M.A.R., Roberts, J.E., and Ciesla, J.A., 2005. Sudden gains in cognitive
behavioral treatment for depression: when do they occur and do they matter?
Behaviour Research and Therapy, 43 (6), 703–714.
Koelsch, S., Enge, J., and Jentschke, S., 2012. Cardiac signatures of personality. PLoS
ONE, 7 (2), e31441.
Kreuter, M.W., Farrell, D.W., Brennan, L.K., and Olevitch, L.R., 2013. Tailoring
health messages: customizing communication with computer technology.
Routledge.
Kurtz, C.F. and Snowden, D.J., 2003. The new dynamics of strategy: Sense-making in
a complex and complicated world. IBM Journal, 42 (3), 462.
Lathia, N., Pejovic, V., Rachuri, K.K., Mascolo, C., Musolesi, M., and Rentfrow, P.J.,
2013. Smartphones for large-scale behavior change interventions. IEEE
Pervasive Computing, 12 (3), 66–73.
Levati, S., Campbell, P., Frost, R., Dougall, N., Wells, M., Donaldson, C., and Hagen,
S., 2016. Optimisation of complex health interventions prior to a randomised
controlled trial: a scoping review of strategies used. Pilot and Feasibility
Studies, 2 (1).
Marchal, B., Van Belle, S., De Brouwere, V., Witter, S., and Kegels, G., 2014.
Complexiy in Health: consequences for research & evaluation. Antwerp:
Tropical Institute for Medicine.
Mattson, M.P., 2015. Lifelong brain health is a lifelong challenge: from evolutionary
principles to empirical evidence. Ageing Res Rev, 20, 37–45.
16
A realist review of digital cognitive bias modification
Mattson, M.P. and Calabrese, E.J., 2010. Hormesis: a revolution in biology,
toxicology and medicine. New York: Springer.
Mensah, G.A. and Collins, P.Y., 2015. Understanding mental health for the prevention
and control of cardiovascular diseases. Global Heart, 10 (3), 221–224.
Michie, S., Richardson, M., Johnston, M., Abraham, C., Francis, J., Hardeman, W.,
Eccles, M.P., Cane, J., and Wood, C.E., 2013. The Behavior Change Technique
Taxonomy (v1) of 93 hierarchically clustered techniques: building an
international consensus for the reporting of behavior change interventions.
Annals of Behavioral Medicine, 46 (1), 81–95.
Michie, S., van Stralen, M.M., and West, R., 2011. The behaviour change wheel: a
new method for characterising and designing behaviour change interventions.
Implementation Science, 6, 42.
Miller, W.L., Crabtree, B.F., McDaniel, R., and Stange, K.C., 1998. Understanding
change in primary care practice using complexity theory. The Journal of
Family Practice, 46 (5), 369–376.
Munro, G.M., 2015. Personality change interventions: a systematic review of
empirical trials.
Munro, G.M., 2016. Towards a complexity approach to digital health behaviour
change: the role of personality computing and exposure therapy. In: Health
Informatics New Zealand (HiNZ).
Musek, J., 2007. A general factor of personality: evidence for the Big One in the five-
factor model. Journal of Research in Personality, 41 (6), 1213–1233.
Nahum-Shani, I., Hekler, E.B., and Spruijt-Metz, D., 2015. Building health behavior
models to guide the development of just-in-time adaptive interventions: a
pragmatic framework. Health psychology : official journal of the Division of
Health Psychology, American Psychological Association, 34 (0), 1209–1219.
Neubeck, L., Lowres, N., Benjamin, E.J., Freedman, S.B., Coorey, G., and Redfern,
J., 2015. The mobile revolution—using smartphone apps to prevent
cardiovascular disease. Nature Reviews Cardiology, 12 (6), 350–360.
Noble, D., 2006. The music of life: biology beyond genes. Oxford University Press.
O’Brien, H.L. and Toms, E.G., 2008. What is user engagement? A conceptual
framework for defining user engagement with technology. Journal of the
American Society for Information Science and Technology, 59 (6), 938–955.
O’Brien, H.L. and Toms, E.G., 2010. The development and evaluation of a survey to
measure user engagement. Journal of the American Society for Information
Science and Technology, 61 (1), 50–69.
Parswani, M.J., Sharma, M.P., and Iyengar, S., 2013. Mindfulness-based stress
reduction program in coronary heart disease: A randomized control trial.
International Journal of Yoga, 6 (2), 111–117.
Pawson, R., 2006. Evidence-based policy: a realist perspective. Sage publications.
Pawson, R., Greenhalgh, T., Harvey, G., and Walshe, K., 2005. Realist review: a new
method of systematic review designed for complex policy interventions.
Journal of Health Services Research & Policy, 10 (1_suppl), 21–34.
Penprase, B. and Norris, D., 2005. What nurse leaders should know about complex
adaptive systems theory. In: Nursing leadership forum. Springer Publishing
Company, 127.
Pfaeffli Dale, L., Dobson, R., Whittaker, R., and Maddison, R., 2016. The
effectiveness of mobile-health behaviour change interventions for
cardiovascular disease self-management: a systematic review. European
Journal of Preventive Cardiology, 23 (8), 801–817.
17
A realist review of digital cognitive bias modification
Plsek, P.E. and Wilson, T., 2001. Complexity, leadership, and management in
healthcare organisations. British medical journal, 323 (7315), 746.
Prochaska, J.O. and Velicer, W.F., 1997. The transtheoretical model of health behavior
change. American Journal of Health Promotion, 12 (1), 38–48.
Randle, J.M., Stroink, M.L., and Nelson, C.H., 2015. Addiction and the adaptive
cycle: a new focus. Addiction Research & Theory, 23 (1), 81–88.
Resnicow, K. and Page, S.E., 2008. Embracing chaos and complexity: a quantum
change for public health. American journal of public health, 98 (8), 1382–
1389.
Resnicow, K. and Vaughan, R., 2006. A chaotic view of behavior change: a quantum
leap for health promotion. Int J Behav Nutr Phys Act, 3, 25.
Roberts, B.W. and Mroczek, D., 2008. Personality trait change in adulthood. Current
Directions in Psychological Science, 17, 31–35.
Rouse, W.B., 2008. Health care as a complex adaptive system: implications for design
and management. Bridge-Washington-National Academy of Engineering, 38
(1), 17.
Rushton, J.P. and Irwing, P., 2011. The general factor of personality: normal and
abnormal. In: T. Chamorro-Premuzic, S. von Stumm, and A. Furnham, eds.
The Wiley-Blackwell Handbook of Individual Differences. Blackwell
Publishing.
Saucier, G. and Srivastava, S., 2015. What makes a good structural model of
personality? Evaluating the Big Five and alternatives. In: Handbook of
Personality and Social Psychology. American Psychological Association, 283–
305.
Schubart, J.R., Stuckey, H.L., Ganeshamoorthy, A., and Sciamanna, C.N., 2011.
Chronic health conditions and Internet behavioral interventions: a review of
factors to enhance user engagement. CIN: Computers, Informatics, Nursing,
29 (2), 81–92.
Selye, H., 1976. The stress concept. Can Med Assoc J, 115 (8), 718.
Shephard, R.J., 2003. Limits to the measurement of habitual physical activity by
questionnaires. British journal of sports medicine, 37 (3), 197–206.
Sher, L., 2005. Type D personality: the heart, stress, and cortisol. QJM, 98 (5), 323–
329.
Short, C.E., Rebar, A.L., Plotnikoff, R.C., and Vandelanotte, C., 2015. Designing
engaging online behaviour change interventions: a proposed model of user
engagement. European Health Psychologist, 17 (1), 32–38.
Stephan, Y., Sutin, A.R., and Terracciano, A., 2014. Physical activity and personality
development across adulthood and old age: evidence from two longitudinal
studies. Journal of Research in Personality, 49, 1–7.
Steptoe, A. and Molloy, G.J., 2007. Personality and heart disease. Heart, 93 (7), 783–
784.
Stone, J., 2004. Personality change after stroke: some preliminary observations.
Journal of Neurology, Neurosurgery & Psychiatry, 75 (12), 1708–1713.
Tang, T.Z. and DeRubeis, R.J., 1999. Sudden gains and critical sessions in cognitive-
behavioral therapy for depression. Journal of Consulting and Clinical
Psychology, 67 (6), 894–904.
The Health Foundation, 2010. Complex adaptive systems. The Health Foundation.
Tkalčič, M., De Carolis, B., de Gemmis, M., Odić, A., and Košir, A., eds., 2016.
Emotions and personality in personalized services. Cham: Springer
International Publishing.
18
A realist review of digital cognitive bias modification
Trzesniewski, K.H., Donnellan, M.B., Moffitt, T.E., Robins, R.W., Poulton, R., and
Caspi, A., 2006. Low self-esteem during adolescence predicts poor health,
criminal behavior, and limited economic prospects during adulthood.
Developmental Psychology, 42 (2), 381–390.
Vinciarelli, A. and Mohammadi, G., 2014. A survey of personality computing. IEEE
Transactions on Affective Computing, 5 (3), 273–291.
Weiss, E.P., Albert, S.G., Reeds, D.N., Kress, K.S., McDaniel, J.L., Klein, S., and
Villareal, D.T., 2016. Effects of matched weight loss from calorie restriction,
exercise, or both on cardiovascular disease risk factors: a randomized
intervention trial. The American Journal of Clinical Nutrition, 104 (3), 576–
586.
Widmer, R.J., Collins, N.M., Collins, C.S., West, C.P., Lerman, L.O., and Lerman, A.,
2015. Digital health interventions for the prevention of cardiovascular disease:
a systematic review and meta-analysis. Mayo Clinic Proceedings, 90 (4), 469–
480.
Winter, S.J., Sheats, J.L., and King, A.C., 2016. The use of behavior change
techniques and theory in technologies for cardiovascular disease prevention
and treatment in adults: a comprehensive review. Progress in Cardiovascular
Diseases, 58 (6), 605–612.
World Health Organization, 2016. Cardiovascular diseases (CVDs): fact sheet
[online]. Available from: http://www.who.int/mediacentre/factsheets/fs317/en/
[Accessed 4 Oct 2016].
Zohar, A.H., Cloninger, C.R., and McCraty, R., 2013. Personality and heart rate
variability: exploring pathways from personality to cardiac coherence and
health. Open Journal of Social Sciences, 1 (6), 32–39.
19
A realist review of digital cognitive bias modification
20