Long et al.

BMC Medicine (2018) 16:94

https://doi.org/10.1186/s12916-018-1087-6

OPINION Open Access

Being pragmatic about healthcare

complexity: our experiences applying
complexity theory and pragmatism to
health services research
Katrina M. Long1* , Fiona McDermott2 and Graham N. Meadows1,3,4

Abstract
Background: The healthcare system has proved a challenging environment for innovation, especially in the area of
health services management and research. This is often attributed to the complexity of the healthcare sector,
characterized by intersecting biological, social and political systems spread across geographically disparate
areas. To help make sense of this complexity, researchers are turning towards new methods and frameworks,
including simulation modeling and complexity theory.
Discussion: Herein, we describe our experiences implementing and evaluating a health services innovation in
the form of simulation modeling. We explore the strengths and limitations of complexity theory in evaluating
health service interventions, using our experiences as examples. We then argue for the potential of pragmatism as an
epistemic foundation for the methodological pluralism currently found in complexity research. We discuss the
similarities between complexity theory and pragmatism, and close by revisiting our experiences putting pragmatic
complexity theory into practice.
Conclusion: We found the commonalities between pragmatism and complexity theory to be striking. These included
a sensitivity to research context, a focus on applied research, and the valuing of different forms of knowledge. We
found that, in practice, a pragmatic complexity theory approach provided more flexibility to respond to the rapidly
changing context of health services implementation and evaluation. However, this approach requires a redefinition of
implementation success, away from pre-determined outcomes and process fidelity, to one that embraces the continual
learning, evolution, and emergence that characterized our project.
Keywords: Complexity theory, Pragmatism, Health services research, Epistemology, Methodology, Implementation
science

Background health, and therefore requires a different set of

Complexity theory has become increasingly popular in research methods [1–4]. Indeed, the mismatch be-
healthcare research over the last two decades. Its emer- tween this hypercomplexity [1] and the dominant
gence provides credence to the growing arguments from mechanistic conception of healthcare [5, 6] has been
health services researchers, namely that the healthcare used as a key explanation for the lack of implementa-
system contains a level of complexity qualitatively differ- tion of evidence-based medicine [4, 7] and healthcare
ent to other systems due to the social nature of innovation [6, 8–10].
Herein, we discuss our recent experiences imple-
menting and evaluating healthcare simulation model-
* Correspondence: katrina.long@monash.edu
1
ing in a large Australian health service. We first
Southern Synergy, Department of Psychiatry, School of Clinical Sciences at
Monash Health, Faculty of Medicine, Nursing and Health Sciences, Monash
explore some of the contributions of complexity the-
University, Melbourne, VIC, Australia ory to our understanding of the healthcare context,
Full list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Long et al. BMC Medicine (2018) 16:94 Page 2 of 9

reviewing some of the key debates in this emerging A key feature of the project, as planned, was the involve-
field. We then explore the possibilities of using prag- ment of the SLG throughout the project via regular pre-
matism to provide the missing epistemological foun- sentations and workshops at the existing monthly SLG
dation required to structure the study of social complexity meetings. Involvement of the SLG was intended as essen-
theory in healthcare. Finally, we revisit our case study to tial in generating scenarios for simulation, developing clin-
discuss how we put pragmatic complexity research into ically accurate conceptual models of patient transitions
action as the evaluation framework for a simulation (within the MHS), and validating the model for use by
modeling project. decision-makers. However, about a year into the project,
the MHS underwent a major restructuring after a signifi-
The real world of healthcare implementation and cant number of senior staff left the service. The
evaluation decision-making processes in the organization changed
We begin by describing the initial aims and scope of the substantially so that simulation-related interactions be-
healthcare simulation modeling project, which provides tween researchers and stakeholders became more reliant
the practical case study for this article. We will return to on one-on-one and small group discussions. At the same
the case study throughout the article to demonstrate some time, changes in policy, such as the introduction of the
of our theoretical arguments. The case study discussed Victorian Mental Health Act [12] and National Disability
here presents the researcher experience of the implementa- Insurance Scheme Act [13], and a freeze or contraction in
tion process, which ultimately shaped and motivated the state and federal mental health funding [14, 15], changed
epistemological quest that is the subject of this paper. For the strategic priorities and decision-making scope of the
readers interested in the outcomes of this project, we refer MHS. Consequently, some simulation models in develop-
them to our upcoming evaluation paper, which provides a ment were no longer of immediate relevance to the partic-
more detailed account of the application of the epistemol- ipants, while other issues that came to the fore, such as
ogy, theory, and methods discussed in this article to imple- the redrawing of clinical catchment areas, did so with
mentation evaluation. time-decision horizons not compatible with the develop-
The 3-year research project aimed to develop simula- ment time costs of discrete event simulation. The
tion models to provide strategic decision support for a researchers adapted, changing both their methods and
senior leadership group (SLG) in a large public mental focus to align with the new strategic directions and
health service (MHS) in Australia. The MHS was re- concerns of the service.
sponsible for government-funded inpatient and commu- This experience may be familiar to many implementa-
nity mental health services across the age spectrum, tion scientists and healthcare managers; however, it does
with different but overlapping catchment areas for early pose significant challenges for evaluators. We subse-
in life (under 25 years), adult, and aged (over 65 years) quently outline how, by applying the twin lenses of com-
services. plexity theory and pragmatism, we developed a deeper
The project was designed to consist of four major understanding of the processes of implementation.
phases, namely (1) the development of a conceptual frame-
work for the simulation model, (2) integration with simula- A view from complexity theory
tion software, (3) validation of the model, and (4) A health services research project as a complex adaptive
implementation of the model within the MHS as a decision system (CAS)
support tool. The initial scope included a model of the There is no doubt that the context described above is
whole MHS, with the option of additional scenarios of complex, or in the language of complexity theory, a CAS
specific interest to managers. Three general types of [16]. A CAS is formally defined as “a collection of individ-
scenarios were identified at the outset, namely (1) policy ual agents with freedom to act in ways that are not always
change affecting the structure of services, (2) population totally predictable, and whose actions are interconnected
distribution changes, and (3) organizational innovation in such that one agent’s actions change the context for other
delivery of care models. The planned modeling approach agents” ([5], p. 625). While there is still some disagreement
was discrete-event simulation using the ARENA® software over the terminology, the key features of a CAS generally
package [11]. include embeddedness, nested systems, fuzzy boundaries,
The original project research team included nine in- distributed control, self-organization, emergence, unpre-
vestigators with experience in psychiatric epidemiology, dictability, non-linearity, phase changes, historicism, sensi-
health economics, simulation modeling, health services tivity to initial conditions, non-equilibrium, adaptation,
planning, organizational change management, action and co-evolution (Box 1) [4, 8, 9, 17–19].
research, and qualitative methodologies. Two of these Many of these features of a CAS were found in our ex-
researchers were existing members of the SLG, and bro- periences. The project involved multiple nested systems,
kered research access with the MHS. namely the researcher group, SLG, MHS, and the state
Long et al. BMC Medicine (2018) 16:94 Page 3 of 9

been convinced of its merit. It was only with the support of

Box 1. Key features of complex adaptive systems (CASs)
this individual that the change occurred, representing a
Embeddedness/nested systems: CASs are embedded within a phase change in the organizational context. This started a
wider context and other CASs. period of 2 years of continual staff and role changes within
the SLG, which could not have been foreseen by the indi-
Fuzzy boundaries: System boundaries are permeable and hard
vidual whose support initiated the process. Both the SLG
to define.
and the researchers exhibited adaptation and co-evolution,
Distributed control and self-organization: System patterns changing strategic priorities and approaches based on the
are not created by top-down control; instead, autonomous changes in context. For example, as the state government
agents interact to create outcomes. Thus, organization in a CAS signaled an increased interest in infrastructure planning for
emerges naturally from local rules held by agents. population growth through a series of discussion papers,
Emergence: Interactions between agents create system the researchers refocused their modeling efforts on the
outcomes that are not directly intended and are greater than area. When new mental health funding was released by the
state government in 2017, members of the SLG, aware of
the sum of the individual agent behaviors.
the researcher activity in this area, successfully lobbied for
Unpredictability: The behavior of a CAS cannot be predicted
funding based on this modeling output.
due to its non-linearity, sensitivity to initial conditions, and
historicism. Applying complexity theory to healthcare research
Non-linearity: The magnitude of system input and agent Rather than attempting to control the research context,
interactions is not linearly related to the magnitude of changes complexity theory directs researchers to make it the
in the system. A CAS can react suddenly to minor inputs or fail focus of their study, looking for patterns of interactions
to change despite overwhelming external pressure. within agents, and between agents and the environment
to explain system-level outcomes [17]. In healthcare,
Phase changes: Where a small change in the system inputs
these systems level outcomes represent public health in-
results in a qualitative change in the system’s state.
terests such as the efficiency and effectiveness of health-
Sensitivity to initial conditions and historicism: Future agent care delivery, the population’s quality of life, and rates of
actions are affected by past changes in the system, leading initial disease morbidity and mortality. Embracing this
conditions to exert a strong influence on system behaviors. approach removes the focus from the short-term out-
Non-equilibrium: CASs are characterized by continual change comes of individual interventions (often randomized
and do not reach equilibrium. control trials), which are isolated from the rest of the
Adaptation and co-evolution: Agents and systems evolve healthcare system, and places it on understanding the
complex contextual factors that determine the long-term
together, reacting to changes in the context to ensure optimal
survival of a new healthcare intervention.
functioning and survival.
A classical approach to complexity theory directs re-
searchers to identify rules that govern these behaviors,
and federal governments. Boundaries between systems were attributing them to the agent (local rules) or an environ-
fuzzy, with participants often exerting influence in multiple mental pattern (attractors). In this classical interpret-
systems. For example, in addition to their employment with ation of complexity theory, established research methods
the MHS, a significant number of SLG participants held include agent-based modeling, simulation, and network
roles within university departments, government advisory analysis, where a theory of local rules is built into a
boards, discipline-specific associations (e.g., Royal Austra- mathematical model, which is tested against reality [17,
lian and New Zealand College of Psychiatrists, Australian 19–21]. However, these approaches have had limited
Medical Association, Australian Psychological Society), or success in healthcare, with low rates of modeling imple-
private consulting clinics. Control was distributed, with mentation [22–24] often being attributed to the lack of
members of the SLG holding responsibility and autonomy good data from which to build models [25–28]; the com-
over the running of different programs within the MHS, re- plex social and organizational context of healthcare, with
searchers exerting control over the implementation activ- multiple intersecting and nested stakeholder groups [1,
ities, and politicians, bureaucrats, and senior healthcare 2, 25, 27, 29, 30]; and the high expertise and time costs
managers making policy and funding decisions that affected of creating sufficiently complex, ecologically valid
the operating context of the SLG and researchers. models [25–29, 31–33].
Changes in the SLG were unpredictable and non-linear, Recent applications of complexity theory to healthcare
instead emerging from what may be considered phase have branched out into more qualitative methods, in-
changes in the system. For example, the first restructure of cluding ethnography, case studies, case–comparison or
the SLG did not occur until a key influential member had time-series analyses, and social surveys [10, 18, 20, 34].
Long et al. BMC Medicine (2018) 16:94 Page 4 of 9

These approaches emerged from the seminal work of [19]. This creates a rift in complexity theory’s understand-
Byrne [18], who translated many of the concepts of com- ing of agency. Classic complexity theory defines agency as
plexity theory into the social realm. an agent’s enactment of their local rules, which ensures
their survival, while social complexity theory defines agency
Debates in social complexity theory as the cognitive, motivational, and emotionally driven
Complexity theory has alternatively been defined as a intentional behaviors that agents employ to achieve their
methodology [35], conceptual framework [4, 20, 36], end goal. This end goal is not always individual survival.
metaphor [4, 34], world view [37], frame of reference [34], Given that Byrne et al. [20] identified agency as a key defin-
ontology [34], or as a “loose set of concepts, heuristics, and ing feature of social complexity theory, and a key target for
analytic tools” ([8], p. S31). Different authors have posited research, how are we to study it, considering these defin-
different typologies of complexity science to address this itional issues?
lack of coherence (e.g., [1, 8, 18, 38]), with a clear delinea- We found this to be a key challenge in our project. In
tion emerging between the complexity theory of things order to evaluate the effect of the simulation modeling on
(classic complexity theory, e.g., cells, animals, atomic the decision-making processes of the SLG, we attempted
particles) and that of humans (social complexity theory). to use interviews to establish a baseline picture of the rela-
The multitude of actors, motives, and behaviors animating tionships, mental models, and expectations of the individ-
social complexity theory poses significant challenges to ual participants. However, the experience of interviewing
both theorizing and researching. Below, we outline the key suggested that disclosure levels varied significantly be-
tensions in this emerging field. tween participants, influenced by their perceptions of the
vested interests of the research group, and the existing
Description or explanation? personal relationships between researchers and SLG
In his critique of social complexity theory, Paley states that members. We also faced difficulties in that time and ac-
“complexity is an explanatory concept” ([39], p. 59). Social cess limitations of working with senior managers often
complexity theorists seem to disagree, describing emer- meant that data were not collected when significant deci-
gence as descriptive, not explanatory [40], and arguing sions were made or events occurred. We therefore had to
that the only way to see the outcome of a CAS is to ob- rely on the retrospective recall of participants to piece to-
serve the system as a whole, rather than its component in- gether a picture of the events, and their roles in them.
dividual agents or models [5]. This raises the fundamental This approach meant that our image of individual events
issue of epistemology. The classical complexity theory was often incomplete, preventing us from accurately iden-
focus on explanation aligns with a positivist epistemology, tifying the role of individual agency in the observed inter-
where knowledge is valued if it is generalizable and allows actions and system-level changes.
us to predict, and manipulate, future behavior [25]. This
clearly aligns with the aim of implementation and most Defining social CASs
public health research, which is namely to affect meaning- There are two pervasive issues with defining a social sys-
ful change. The epistemology of social complexity theory, tem, nesting and fuzzy boundaries, both of which are
on the other hand, is unclear. If social complexity theory implicated in, and complicate, complexity research [19].
does represent a purely descriptive epistemology, which In the health system, Byrne et al. [20] identified four
makes no claims to the translation of findings across levels of nested systems, namely the individual, popula-
contexts, then its ability to contribute to implementation tion health, the health service system, and the planetary
science may be minimal. ecosystem. However, several more exist within the health
service system, including general practices, practice net-
The role of agency in self-organization works, hospitals, hospital networks, and national pro-
The redefinition of local rules as human instincts, con- grams [5]. Thus, a key question facing complexity
structs, and mental models has also been subject to debate researchers is which systems should form the core of the
[5, 39]. This is in part due to the inherent problems with analyses, and how many levels of analysis are sufficient
trying to measure internal states, with even qualitative to provide a complete understanding of the system.
methods heavily reliant on individual insight and candor The boundaries of social systems are also harder to de-
[21]. This is also due to the lack of fit between the focus of fine and control than in a classic CAS [21, 34]. As we dis-
classic complexity theory, individual agent survival, and the covered in our efforts to develop simulation models of
postmodern ideas of democracy and collectivism which mental health patients, a patient may pass through mul-
shape the social world. While survival in biological systems tiple different practices, hospitals, and even districts over
can be treated as a key driver and outcome measure, the an episode of care, interacting with scores of individual
survival of organizations, systems of operations, and even agents, each operating in a different context. Likewise, the
ideas are less necessary, or observable, in social systems boundaries of the implementation context proved hard to
Long et al. BMC Medicine (2018) 16:94 Page 5 of 9

define. Despite beginning with a focus on the MHS as the

Box 2: Similarities between social complexity theory
key implementation context and the SLG as the key
and pragmatism
agents, it emerged through the course of the evaluation
that the context of individual researchers (e.g., contract Both:
changes, relocations, life events), researcher team dynam-
ics, and the wider government contexts exerted very sig-  Aim to create ‘useful’ knowledge
nificant influences on the trajectory of the project. Thus,  Reject reductionist science in favor of the study of whole
system boundaries are often arbitrary, with implementa-
systems, in context
tion and evaluation researchers required to balance
 Understand research as a continual learning process
descriptive sufficiency with practicality.
These issues lead us to a key consideration – in light of  Focus on the social consequences of research and
these debates in social complexity theory, how can com- intervention
plexity researchers make transparent and consistent deci-  Value the democratization of knowledge and research,
sions regarding research methodology. While social valuing all stakeholders’ input
complexity theory offers a clear ontology, focusing on  Prioritize understanding over theoretical or methodological
agent interactions and emergent system outcomes [34], it purity, encouraging the use of multiple methods
lacks a clear position on the epistemic contribution of
studying CASs. We suggest that what is needed is a clear
epistemology [4], and we suggest that pragmatism may
provide the epistemological foundations required to struc- In our project, the response to the challenge of work-
ture the study of social complexity theory in healthcare. ing within this particular CAS manifested through the
emergent formulation of two deeply pragmatic research
A contribution from pragmatism questions: How can we (the researchers) help to improve
What is pragmatism? strategic decision-making for mental health services?
We suggest that many healthcare workers would identify What can we learn of value through this process? This
as pragmatists. The everyday use of the term pragma- allowed us, as the context changed, to maintain the
tism implies a focus on the practical and achievable, ra- same focus for the project, but change and expand the
ther than the theoretical or ideal [41]. This idea of evaluation focus from the experiences of the SLG to in-
valuing the applied over the theoretical is mirrored in clude, for instance, adaptations of the researchers to the
the philosophy of Pragmatism. changing stakeholder needs. The same aims were ad-
Pragmatism emerged in the late 1800s in the work of dressed, but using different methods.
Charles Pierce, William James, and John Dewey. At the
center of pragmatism is a rejection of the ‘impossible Continual learning
question’ of philosophy, that of the nature of the mind’s The contextualization of knowledge does not reject the
relationship to reality [42]. Instead, pragmatists judge translation of knowledge between contexts. While prag-
the value of knowledge (and our ways of knowing) by its matism does hold that knowledge is not completely
context-dependent, extrinsic usefulness for addressing generalizable, it also argues that imported knowledge can
practical questions of daily life [43]. Perfect knowledge is play a role in shaping observation and perception and in
not possible, nor required. For pragmatism, knowledge suggesting possible solutions to the current problem [42].
is only meaningful when coupled with action [38]. For implementation science, the merging of complexity
There are many similarities between the arguments of theory’s deep focus on contextual interactions and emer-
social complexity researchers and pragmatists. Below we gent outcomes, coupled with pragmatism’s perspective on
explore key synergies (Box 2). knowledge translation, provides a way of fostering collect-
ive implementation learning [16, 46], without bowing to
Contextualized research the need for research generalizability.
A key feature of pragmatism is the contextualization of For our project, this led us to re-define implementation
knowledge [44, 45]. As contexts change, so too do the cri- success, not as a strict adherence to the project plan or
teria of usefulness for knowledge. Similarly, social com- the achievement of pre-determined outcomes (i.e., the
plexity theory calls for the matching of research approach publication of four simulation models and the use of these
to context and level of environmental complexity [4, 9]. In models to inform decisions), but by the perceived useful-
complexity theory, these contexts could include different ness of the project to the stakeholders and the lessons
nested systems, and different time points [44]. Therefore, learned. As Byrne commented: “The point about complex-
in order to maintain a coherent research agenda in a CAS, ity is that it is useful – it helps us to understand the things
a unifying research question is required. we are trying to understand” ([18], p. 7). Indeed, what we
Long et al. BMC Medicine (2018) 16:94 Page 6 of 9

learnt was that the simulation models themselves seemed participant experience, to one that included the experi-
not to be the main outcome of interest to the SLG; in- ences of the researchers. In the initial design of the
stead, it was the personal insights that members gained evaluation, the CAS of interest was that of the SLG. Our
from the conceptual development discussions and our evaluation was focused on understanding the
presentations of amalgamated patient data. decision-making mental models of these individuals, and
how they negotiated shared group processes and behav-
Research as social action iors based on these individual models. However, the
Another key pillar of pragmatism is the active and social organizational restructure of the SLG affected not only
nature of inquiry. Dewey argued that the primary function access to participants for evaluation data collection, but
of research is to solve societal problems [38]. However, he also affected the researchers’ approach to the simulation
also argues for flexibility in application, proposing “that modeling development and implementation. As men-
policies and proposals for social action be treated as work- tioned above, one way this manifested was as a change
ing hypotheses, not as programs to be rigidly adhered to in engagement with members of the SLG. Researchers
and executed” ([47], pp. 151–2). began using one-on-one interactions with engaged SLG
These sentiments are echoed in social complexity members to develop new scenarios directly related to
theory: the SLG members’ portfolio. Therefore, the experiences
“Complexity/chaos offers the possibility of an engaged and reflections of the researchers became pivotal in un-
science not founded in pride, in the assertion of an abso- derstanding the project’s implementation after the
lute knowledge as the basis for social programs, but ra- organizational restructure.
ther in a humility about the complexity of the world Both pragmatism and complexity theory also encourage a
coupled with a hopeful belief in the potential of human focus on the interactions of knowledge systems, and the
beings for doing something about it.” ([18], p. 45). study of how these intersections are negotiated [4, 44, 48].
Not only does pragmatism argue for a problem-solving For us, this manifested as multiple themes emerging from a
approach to inquiry, but also to an action-based one. All grounded theory approach to the implementation evalu-
modes of experience, including research, are treated as ation, including participant-researcher communication (fre-
interventions [42]. Research success within a pragmatic quency, modality, content), understanding and expectations
epistemology is measured by consequences, whether of the modeling methodology, and different outcome prior-
they be predicted or emergent. This aligns with the hol- ities between the researchers and participants. The case
istic system view of complexity theory, where outcomes study approach of the evaluation, supported by interviews
are not pre-determined, but emergent [36]. Thus, com- and unstructured observation, allowed these themes to
plexity theory provides a way of operationalizing the emerge, but there remains a challenge for creating more
study of emergent consequences, while pragmatism pro- targeted research designs and methods capable of captur-
vides the impetus for change by measuring research ing, measuring, and interpreting these interactive and
quality with respect to its impact on social change. emergent processes.

Valuing of different knowledge Support for mixed methods research

The usefulness of knowledge metric also creates a A key theme in the development of social complexity re-
democratization of scientific endeavor. Scientific know- search is the call for mixed methods research [8, 34].
ledge is treated not as a qualitatively different form of However, there is a risk of method choice being driven by
knowledge, but simply as a more formalized version of the ‘what works’ maxim [50]. As one of the key epistem-
everyday human inquiry [48]. Science therefore becomes a ologies for mixed method research, pragmatism offers a
social pursuit, within anyone’s reach. This idea of intuitive more structured approach to mixed methods research
inquiry aligns with a theme, advanced by many scholars [42]. Pragmatism calls for choices of research questions
advocating for complexity theory in healthcare, that social and methods to be driven by the social purpose of the re-
actors already have an intuitive sense of complexity, which search, not the other way around [42, 45, 51].
can be refined by the framework of complexity theory [4, Another of the risks identified by complexity theorists is
9]. Social complexity theorists also argue for a natural fit the pre-emptive labelling of a system as complex [40]; a
between complexity approaches and participatory pragmatic approach does not require such a priori as-
research, where participant and researcher frames of refer- sumptions. Rather, it allows for the flexible use of multiple
ence are treated as equally important to inquiry [20], fail- methods to capture insights in a complex environment,
ure is tolerated and expected [49], and innovation is which may later be interpreted using a range of frame-
allowed to emerge from any part of the system [9]. works. Therefore, our pluralism of evaluation methods
In our project, this led to a fundamental shift in the (i.e., interviews, questionnaires, document analysis, obser-
implementation evaluation from a focus purely on the vations) provides us with multiple perspectives to be
Long et al. BMC Medicine (2018) 16:94 Page 7 of 9

explored and structured in different ways in order to ul- approaches to the difficulties in evaluating such imple-
timately build an understanding of the process of mentations. The commonalities between pragmatism and
implementation. complexity theory are striking, and include a sensitivity to
Pragmatism also encourages reflection and experimen- research context, a focus on applied research, and the
tation, allowing for the evolution of interventions and valuing of different forms of knowledge. For implementa-
evaluation in a similar fashion to a CAS [7, 42, 45]. tion and evaluation, this fusion of approaches has signifi-
Therefore, our shift in evaluation from the quantitative cant implications:
analysis of participant questionnaire responses to a
grounded theory case study of research adaptation is not  A focus on researcher and stakeholder agency, in
only consistent with complexity theory, but predicted by shaping the direction and outcomes of interventions.
it, as a co-evolution of the researchers in context. Thus,  A re-definition of implementation success, not as a
rather than rejecting the reductionist approach of classic strict adherence to the project plan, or the achieve-
complexity theory [20], pragmatism allows for the con- ment of pre-determined outcomes, but as the emer-
tribution of both quantitative and qualitative methods in gent outcomes of the project and lessons learned, as
addressing the research question. It also allows for dif- identified by all stakeholders.
ferent definitions of complexity theory. Complexity the-  A flexibility in implementation and evaluation
ory can be both an ontology for quantitative approaches methods, encouraging the reflexive use of mixed
and a metaphor for qualitative approaches. methods to capture and adapt to the changing
research context.
The case study revisited  A rejection of the description-explanation divide, fo-
Our case study illustrates how a pragmatic epistemology cusing instead on continual, collective learning,
can support, and broaden, the application of complexity where case studies provide starting points, not the-
theory to healthcare implementation and evaluation. ories, for future research.
By starting from a pragmatic epistemology, we allowed
our focus to be drawn to the most relevant ontology and However, our recommendations are not without limita-
methodologies for the study of this implementation. Com- tions. There are other epistemic options for complexity the-
plexity theory emerged as a relevant theory and ontology for ory, including nested theories [34], an eclectic use of
the analysis; however, we do not hold that it is the only pos- middle-range theories [37], or a pluralistic ontology of levels
sible lens through which to evaluate the implementation. A supported by emergence [26]. One of the more promising
pragmatic frame encouraged us to embrace different types alternatives comes from Byrne et al.’s [20] application of
of inquiry and data collection methods, using questionnaire, complex realism to complexity theory. At face value, the ar-
interview, observation, and document analysis approaches. guments of complex realism seem not incommensurate
As the implementation progressed, we included new partici- with pragmatism [42]; however, we will leave a detailed
pants (i.e., researchers), and expanded our frame of data col- comparison of these two approaches to future scholars. Al-
lection to include government policy and funding changes. ternatively, complexity theorists may entirely reject our sug-
By doing so, we overcame one of the key challenges in social gestion of the need for an epistemology. Another limitation
complexity theory – defining the CAS of interest. is posed by the theoretically agnostic position of pragma-
In our evaluation, we pragmatically allowed implemen- tism, as outlined above. It is highly likely that a pragmatic
tation success to be defined by the collection of stake- approach will not always support the application of com-
holders, honoring the multitude of different expectations plexity theory in healthcare implementation research. While
held by the research funding body, the academic commu- we believe this is a strength in the use of pragmatism in
nity, and individual members of the SLG and research healthcare implementation, it may limit the uptake of prag-
team. We then began the data analysis with a critical inci- matism by researchers who specialize in complexity theory.
dent approach to identify turning points in the system, The application of complexity theory to social science, in-
which were investigated further with thematic analysis. It cluding healthcare, is still in its infancy. So too is the
was only when the emerging themes resonated with a formalization of pragmatism as a school of philosophy [43].
complexity theory interpretation of the project that we la- However, we agree with Talisse and Aikin, in that discus-
beled our case study as a healthcare implementation CAS. sions such as those presented in this article are a positive
sign, “a mark of …vitality, an indication that it is a living
Conclusions philosophy rather than a historical relic.” ([43], p. 3). We
Herein, we described a too-familiar experience in health present this article in that spirit and hope that our contri-
services implementation – a constantly changing imple- bution sparks further discussion about the potential collab-
mentation context– followed by a discussion of how com- oration of pragmatism and complexity theory in informing
plexity theory and pragmatism provide complementary implementation science and health services research.
Long et al. BMC Medicine (2018) 16:94 Page 8 of 9

Abbreviations Received: 31 January 2018 Accepted: 25 May 2018

CAS: complex adaptive system; MHS: mental health service; SLG: senior
leadership group

References
Discontinuing participants 1. Klein JH, Young T. Health care: a case of hypercomplexity? Health Systems.
All participants are advised at the beginning of participation in the study 2015;4(2):104–10.
and in the consent form that they can discontinue participation at any time. 2. Eldabi T. Implementation issues of modeling healthcare problems:
Data already collected will still be used in the analysis unless the misconceptions and lessons. Proceedings of the 2009 winter simulation
discontinuing participant specifically requests that it be removed. conference (WSC). IEEE. 2009; https://doi.org/10.1109/WSC.2009.5429192.
3. Tako AA. Robinson S. Is simulation in health different? J Oper Res Soc. 2015;
66(4):602–14.
Funding 4. Kernick D. Wanted – new methodologies for health service research. Is
The preparation of this article was in part supported by an Australian complexity theory the answer? Fam Pract. 2006;23(3):385–90.
Postgraduate Award funded by the Australian Research Council’s Linkage 5. Plsek PE, Greenhalgh T. The challenge of complexity in health care. Br Med
Projects scheme (LP110200061) entitled “Improving management decisions J. 2001;323:625–8.
in mental health care through applications of advanced simulation 6. Plsek PE, Wilson T. Complexity, leadership, and management in healthcare
modelling”. Additional research funding was provided by the Department of organisations. Br Med J. 2001;323:746–9.
Psychiatry, Monash University, and the University of Calgary. The views, 7. Sanderson I. Intelligent policy making for a complex world: pragmatism,
analyses, interpretations, and conclusions expressed in the article are those evidence and learning. Political Studies. 2009;57(4):699–719.
of the authors and not of the Australian Research Council, Monash 8. Litaker D, Tomolo A, Liberatore V, Stange KC, Aron D. Using complexity
University, or University of Calgary. theory to build interventions that improve health care delivery in primary
care. J Gen Intern Med. 2006;21(S2):S30–4.
9. Plsek PE. Complexity and the adoption of innovation in health care. In:
Availability of data and materials accelerating quality improvement in health care: strategies to accelerate the
Only investigators and approved researchers added by ethics approval will diffusion of evidence-based innovations. Washington, DC: National Institute
have access to the dataset. for healthcare management foundation and National Committee for.
Quality in Health Care. 2003;
10. Anderson RA, Crabtree BF, Steele DJ, McDaniel RR. Case study research: the
Authors’ contributions view from complexity science. Qual Health Res. 2005;15(5):669–85.
KL was responsible for the development of the ideas expressed, and drafting 11. Rockwell Automation. ARENA. In. In: 13.90 edn; 2010.
of the article. FM contributed expertise in using complexity theory in 12. Mental Health Act 2014. Government of Victoria. No 26 of 2014. http://
qualitative research. GM was responsible for the design of the simulation www8.austlii.edu.au/cgi-bin/viewdb/au/legis/vic/num_act/
modeling intervention featured in the case study. All authors read and mha201426o2014174/. Accessed 23 May 2018.
approved the final manuscript. 13. National Disability Insurance Scheme Act 2013. Parliament of Australia, No
20, 2013. https://www.legislation.gov.au/Details/C2016C00894. Accessed 23
May 2018.
Ethics approval and consent to participate 14. Cordery A. Innovative mental health program falling victim to funding
The project was approved by the Human Research Ethics Committee of the freeze. The Sydney Morning Herald. 2015; https://www.smh.com.au/
partner Mental Health Service, with approval from 5 December 2013 until 9 healthcare/innovative-mental-health-program-falling-victim-to-funding-
January 2019. freeze-20150423-1ms0wm.html. Accessed 22 May 2018
The simulation modeling component of the project received reciprocal 15. Victorian Healthcare Association. State budget submission 2015–16.
approval by the Human Research Ethics Committee at Monash University Melbourne: VHA; 2015.
(CF14/1072–2,014,000,460). The evaluation component of the project 16. McDermott F. Complexity theory, trans-disciplinary working and reflective
received reciprocal approval by the Human Research Ethics Committee at practice. In: Applying complexity theory. Pycroft a, Bartollas C. Bristol: policy
Monash University (CF14/48–2,013,001,967). press; 2014. p. 181–98.
All protocol amendments were submitted to, and approved by, the Human 17. Holland JH. Complexity: a very short introduction. Oxford: Oxford University
Research Ethics Committee of the partner Mental Health Service, using the Press; 2014.
appropriate ethics amendment forms. 18. Byrne DS. Complexity theory and the social sciences: an introduction. New
Signed consent is obtained from all participants during their first in-person York, NY: Routledge; 1998.
contact with the study. Any information gained in connection with this research 19. Manson SM. Simplifying complexity: a review of complexity theory.
project that can identify individuals will remain confidential. All information will Geoforum. 2001;32(3):405–14.
be stored in password-protected files and folders on password-protected 20. Byrne D, Callaghan G, Winter T. Complexity theory and the social sciences:
computers, that can only be accessed by the research staff the state of the art. Abingdon: Taylor and Francis; 2013.
21. McKelvey B. Complexity theory in organization science: seizing the promise
or becoming a fad? Emergence. 1999;1(1):5–32.
Competing interests
22. Long KM, Meadows G. Simulation modelling in mental health: a systematic
The authors declare that they have no competing interests.
review. J Simulation. 2018;12(1):76–85.
23. Fone D, Hollinghurst S, Temple M, Round A, Lester N, Weightman A,
Roberts K, Coyle E, Bevan G, Palmer S. Systematic review of the use and
Publisher’s Note value of computer simulation modelling in population health and health
Springer Nature remains neutral with regard to jurisdictional claims in care delivery. J Public Health Med. 2003;25(4):325–35.
published maps and institutional affiliations. 24. Brailsford SC, Harper PR, Patel B, Pitt M. An analysis of the academic
literature on simulation and modelling in health care. J Simulation. 2009;
Author details 3(3):130–40.
1
Southern Synergy, Department of Psychiatry, School of Clinical Sciences at 25. Brailsford SC, Bolt TB, Bucci G, Chaussalet TM, Connell NA, Harper PR, Klein
Monash Health, Faculty of Medicine, Nursing and Health Sciences, Monash JH, Pitt M, Taylor M. Overcoming the barriers: a qualitative study of
University, Melbourne, VIC, Australia. 2Department of Social Work, School of simulation adoption in the NHS. J Oper Res Soc. 2013;64(2):157–68.
Primary and Allied Health Care, Faculty of Medicine, Nursing and Health 26. Brailsford SC. Overcoming the barriers to implementation of operations
Sciences, Monash University, Melbourne, VIC, Australia. 3Melbourne School of research simulation models in healthcare. Clin Investig Med. 2005;28(6):312–5.
Population and Global Health, University of Melbourne, Melbourne, VIC, 27. Robinson S, Pidd M. Provider and customer expectations of successful
Australia. 4Monash Health, Melbourne, VIC, Australia. simulation projects. J Operat Res Soc. 1998;49(3):200–9.
Long et al. BMC Medicine (2018) 16:94 Page 9 of 9

28. van Lent WA, VanBerkel P, van Harten WH. A review on the relation
between simulation and improvement in hospitals. BMC Med Inform Decis
Mak. 2012;12:18.
29. Brailsford SC, Bolt T, Connell C, Klein JH, Patel B. Stakeholder engagement in
health care simulation. Proceedings of the 2009 Winter Simulation
Conference. doi:https://doi.org/10.1109/WSC.2009.5429190.
30. Kirchhof P, Meseth N. A survey on the use of simulation in German
healthcare. Proceedings of the 2012 Winter Simulation Conference. doi:
https://doi.org/10.1109/WSC.2012.6465026.
31. Tunnicliffe-Wilson JC. Implementation of computer simulation projects in
health care. J Operat Res Soc. 1981;32(9):825–32.
32. Lane DC, Monefeldt C, Husemann E. Client involvement in simulation
model building: hints and insights from a case study in a London hospital.
Health Care Manage Sci. 2003;6:105–16.
33. Barnes CD, Benson C, Quiason JL, McGuiness D. Success Stories in
Simulation in Health Care. Proceedings of the 29th Conference on Winter
Simulation 1997. Atlanta, GA: IEEE Computer Society; 1997. p. 1280–1285.
34. Walton M. Applying complexity theory: a review to inform evaluation
design. Eval Program Plann. 2014;45:119–26.
35. Haynes P. Complexity theory and evaluation in public management. Public
Manage Rev. 2008;10(3):401–19.
36. Arndt M, Bigelow B. Commentary: the potential of chaos theory and
complexity theory for health services management. Health Care Manag Rev.
2000;25(1):35–8.
37. Greenhalgh T, Plsek P, Wilson T, Fraser S, Holt T. Response to ‘the
appropriation of complexity theory in health care’. J Health Serv Res Policy.
2010;15(2):115–7.
38. Grayling AC, Baggini J, Fricker M. Pragmatism. In: In Our Time. Bragg M. BBC
Radio 4; 2005. https://www.bbc.co.uk/programmes/p003k9f5. Accessed 22
May 2018.
39. Paley J. The appropriation of complexity theory in health care. J Health Serv
Res Policy. 2010;15(1):59–61.
40. Goldstein J. Emergence as a construct: history and issues. Emergence. 1999;
1(1):49–72.
41. Oxford English Dictionary: "pragmatic, adj. and n.": Oxford University Press.
42. Biesta G. Pragmatism and the philosophical foundations of mixed methods
research. In: SAGE handbook of mixed methods in social and behavioral
research. 2nd edn. Los Angeles, CA: SAGE Publications; 2010. p. 95–118.
43. Talisse RB, Aikin SF. Pragmatism: a guide for the perplexed. London:
Continuum. 2008;
44. Ruwhiu D, Cone M. Advancing a pragmatist epistemology in organisational
research. Qualit Res Organ Manage. 2010;5(2):108–26.
45. Rylander A, Pragmatism and Design Research - An Overview. Designfakultetens
kunskapssammanställningar. In: Stockholm; 2012. http://www.designfakulteten.
kth.se/sites/default/files/designfpragdesignrapport_18.4.pdf .
46. Richardson K, Cilliers P. What is complexity science? A view from different
directions. Emergence Complex Organ. 2001;3(1):5–23.
47. Dewey J. The public and its problems: an essay in political inquiry.
University Park: Pennsylvania State University Press; 2012.
48. Morgan DL. Paradigms lost and pragmatism regained: methodological
implications of combining qualitative and quantitative methods. J Mixed
Methods Res. 2007;1(1):48–76.
49. Little A. Political action, error and failure: the epistemological limits of
complexity. Political Studies. 2011;60(1):3–19.
50. Hathcoat JD, Meixner C. Pragmatism, factor analysis, and the conditional
incompatibility thesis in mixed methods research. J Mixed Methods Res.
2015;11(4):433–49.
51. Greene JC, Hall JN. Dialetics and pragmatism: being of consequence. In:
SAGE handbook of mixed methods in social and behavioral research. 2nd
edn. Los Angeles, CA: SAGE Publications; 2010. p. 119–44.