Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/271856985
AnEmergingScienceofImprovementinHealth
Care
ARTICLEinQUALITYENGINEERING·JANUARY2015
ImpactFactor:0.36·DOI:10.1080/08982112.2015.968042
CITATION
1
4AUTHORS:
BoBergman
AndreasHellström
ChalmersUniversityofTechnology
ChalmersUniversityofTechnology
94PUBLICATIONS899CITATIONS
16PUBLICATIONS47CITATIONS
SEEPROFILE
SEEPROFILE
SvanteLifvergren
SusanneGustavsson
ChalmersUniversityofTechnology
ChalmersUniversityofTechnology
19PUBLICATIONS33CITATIONS
3PUBLICATIONS2CITATIONS
SEEPROFILE
SEEPROFILE
Availablefrom:SvanteLifvergren
Retrievedon:21August2015
This article was downloaded by: [Svante Lifvergren]
On: 25 December 2014, At: 09:50
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Quality Engineering
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/lqen20
An Emerging Science of Improvement in Health Care
ab
a
Bo Bergman , Andreas Hellström , Svante Lifvergren
ac
& Susanne M. Gustavsson
ac
a
Centre for Healthcare Improvement, Chalmers University of Technology, Gothenburg,
Sweden
b
Meiji University, Tokyo, Japan
c
Skaraborg Hospital Group, Västra Götaland Region, Sweden
Published online: 21 Dec 2014.
Click for updates
To cite this article: Bo Bergman, Andreas Hellström, Svante Lifvergren & Susanne M. Gustavsson (2015) An Emerging Science
of Improvement in Health Care, Quality Engineering, 27:1, 17-34, DOI: 10.1080/08982112.2015.968042
To link to this article: http://dx.doi.org/10.1080/08982112.2015.968042
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Quality Engineering, 27:17–34, 2015
Copyright Ó Taylor & Francis Group, LLC
ISSN: 0898-2112 print / 1532-4222 online
DOI: 10.1080/08982112.2015.968042
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
An Emerging Science of Improvement in
Health Care
Bo Bergman 1, 2,
€ m 1,
Andreas Hellstro
Svante Lifvergren 1, 3,
Susanne M. Gustavsson 1, 3
1
Centre for Healthcare
Improvement, Chalmers
University of Technology,
Gothenburg, Sweden
2
Meiji University, Tokyo, Japan
3
Skaraborg Hospital Group,
€taland Region, Sweden
V€
astra Go
ABSTRACT The purpose of this article is to describe the emerging science of
improvement in health care and to add a perspective from the industrial quality
improvement movement, the use of data from quality registers, and to give
some personal reflections and suggestions. Furthermore, we want to introduce
to the broader quality management community what is happening in health
care with respect to quality improvement. We will discuss some of the
challenges of the health care system and the current status of a science of
improvement and give some suggestions for further improvements to the area.
We discuss a possible extension of improvement knowledge and the
theoretical and practical arsenal of a science of improvement, in particular,
understanding variation and implications for the use of, for, example control
charts. In addition, the normative side of a science of improvement is
discussed. The article ends with some brief reflections of use for future research
agendas.
KEYWORDS health care, quality improvement, improvement science, science of
improvement, profound knowledge
INTRODUCTION
This article was presented at the
Second Stu Hunter Research
Conference in Tempe, Arizona,
March 2014.
Address correspondence to Bo
Bergman, Chalmers University of
Technology, Gothenburg SE-412 96,
Sweden. E-mail: [email protected]
Color versions of one or more of the
figures in the article can be found
online at www.tandfonline.com/lqen.
Health care is increasingly facing a lot of challenges. The demands are
increasing and reports about health care shortcomings have become increasingly common. Thus, there is a need to introduce new ways of working. Health
care of today is said to be evidence driven with high requirements for data from
randomized controlled trials and from careful observational studies. However,
even though huge amounts of data are collected in quality registers and by other
means, clinical practice is often not governed utilizing these data for learning,
improvement, and innovation. Thus, even though individual health care workers are providing excellent and dedicated work, the health care system might be
flawed and patient outcome might be less than optimal, perhaps unsatisfactory,
or even unacceptable without anybody noticing it, more than individual
patients.
During the last decades there has been a movement toward a view that clinics
and health care systems should not only have a systematic way of improving
17
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
medical treatment and nursing utilizing evidence from
research but also have a systematic way of improving
how health care is delivered to the patients—how the
health care system supports its patients to better health
without risking any harm. An attempt to provide a
basis for such initiatives is called improvement science or
a science1 of improvement. The purpose of this article
is to describe this emerging field and to add a perspective from the industrial quality improvement movement, the use of data from quality registers, and to give
some reflections and suggestions. In this introductory
section we will first discuss some challenges of the
health care system and give a short outline of the rest
of the article where the current status of a science of
improvement is discussed as well as some suggestions
for further developments of the area.
The increasing demand on the health care system is
due to, first, the growing population of elderly people
suffering from multiple diseases. At the same time, the
productive part of the population has become relatively smaller and therefore health care funding may be
jeopardized (Weisz et al. 2013). A second reason is the
appearance of new and much improved possibilities to
give life supporting cure and care. Some of these possibilities are very costly, and even if many may indeed
reduce the total costs to society, the cost reductions do
not always benefit those that have the cost burdens
from new treatments. Systems handling this unbalance
are not common. Furthermore, the rate of innovations
in health care is increasing—it has been argued (DixonWoods, Alamberti et al. 2011; Sjohania et al. 2007)
that the half-life of knowledge is 5.5 years. A third
argument is an increased focus on health and wellbeing in the population—how does the health care system in a productive way take care of an increased health
literacy and requirements on patient autonomy and
patient empowerment? (Anderson and Funnell 2005;
Reach 2014).
Another kind of challenge is the stream of reports on
the shortcomings of health care systems. In the UK, for
example, the Bristol heart scandal (Weick and Sutcliff
2001) at the Bristol Royal Infirmary a number of years
ago revealed unacceptable shortcomings and later the
Mid Staffordshire scandal attracted a lot of attention,
with many investigations and government reports—
similar shortcomings were reported in these inquiries:
No clear visions and goals, no culture of patient centeredness, boards and top management not taking the
lead on these questions, insufficient feedback loops on
what is really going on in the organization with respect
to quality and safety, etc. In a recent study (DixonWoods et al. 2014) of the National Health Service
(NHS) many good things were reported, specifically
the ambitions and goodwill of the health care professionals to provide good care; on the negative side, however, it reported that the above-mentioned
shortcomings might be more widespread than so far
revealed. We will return to this study later.
It is not only in the UK that these types of problems
are frequent. In Swedish newspapers we read almost
every day about mistreatments, long waiting times, and
scarcity of personnel. Just to mention one recent story
revealed by a Swedish TV program, the deteriorating
screening procedures of cervix cancer samples may
have caused malign tissue samples to remain undiscovered. Even though results were fed into a database,
there was no feedback to the person performing the
screening. In all work correct feedback is important—without correct feedback any work process may
deteriorate.
The high cost of the U.S. health care system2 is, of
course, also well known. For example, the U.S. health
care expenditure in relation to the gross national product is almost double that of Sweden and the medical
results are not better; in many respects, in fact, they are
much worse. Furthermore, in a recent paper (James
2013) it was reported that as many as 200,000 and perhaps even 400,000 deaths per year in the United States
may be caused by avoidable harm.
Of course, there are no silver bullets to solve these
problems (Shojania and Grimshaw 2004). However, as
one part of a solution, a focus on quality and quality
improvement seems necessary. As illustrated by the
study of the NHS, a mindset, close to what in industrial settings has been known as total quality management
(Bergman and Klefsj€
o 2010), should be strived for—
perhaps a new management paradigm in health care. In
this direction we find the emerging science of improvement in health care.
In this article, we will first discuss the findings of
the NHS study mentioned above and relate these to
2
The concept science might be considered problematic; see, for
example, Wilmott (1997).
1
18
Based on Oganization for Economic Cooperation and
Development data comparisons between the cost and medical
outcomes were compared in a report from Swedish Association of
Local Authorities and Regions (SALAR).
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
the early discussions on quality in health care by Ernest
Amory Codman3 and Avedis Donabedian (1992,
2002)4 but also to the early discourse on quality
improvement in industry. Based on these discussions
we will inquire into the scope of a science of improvement. During the course of the article we will come
back to the NHS study mentioned above for illustrative purposes. Illustrations from the Swedish health
care will also be used. As part of a brief review, some
earlier discussions on the emerging science of improvement will be given. Then, in the main part of the article, the core of a science of improvement will be
discussed. It should, however, be emphasized that a science of improvement in health care really is in a preparadigmatic stage, to use the words of Kuhn (1962). Thus,
many interpretations and outlines have been given, perhaps under different headings, and probably many
more will come—and should come. Indeed, this article
will give our interpretations and suggestions, which
might go further in scope than many other alternatives.
A science of improvement must, of course, address
how improvement issues are identified and elaborated
upon, eventually leading to successful interventions
with even better outcomes for the patients. It must also
address how these improvements are evaluated and
how lessons are learned from them. But it is equally
important to foster an environment where improvements are seen as normal and inspiring parts of the
daily life in an organization. This environment is a pivotal requirement for a transformed health care system.
How is that achieved? These latter questions are much
harder and more pressing than the operating work with
improvements themselves, however important they
are. Of course, these aspects—that is, the operational
improvement interventions and the improvement of
the organization’s readiness (capability and capacity) to
improve and innovate—should go hand in hand, and
no simple recipe can be expected.
THE SCOPE OF A SCIENCE OF
IMPROVEMENT
The NHS study mentioned above was a multimethod study of the NHS during the years 2010–2012
with interviews, surveys, board protocols, and direct
3
For a short biography of Codman, see Neuhauser (2002).
4
For a short biography about Donabedian, see Best and Neuhauser
(2004).
Improvement in Health Care
observations together with data from 2007 to 2011 on
staff experiences, patient experiences, and patient mortality rates. A large number of findings in this study
corroborate our view of what is important in a science
of improvement. Some findings were highly encouraging; for example, a consistent focus from front line
people to do their best for their patients to create the
best possible care.
However, in some organizations there were no clear
goals, no vision, and no culture for continually refining
patient quality and safety. In such organizations the
front line staff observed severe system deficiencies but
felt powerless to deal with them, whereas management
perceived problems with quality and safety as caused
by the front line people. That is, the staff and the management had no common understanding about how to
solve their problems and to improve patient quality
and safety. In addition, there was a lack of intelligence—that is, poor follow-up of both their own results
and results from similar units—in order to find
improvement opportunities. Interestingly, the researchers were able to observe that
. . . hospital standardised mortality ratios were inversely associated with positive and supportive organisational climates.
Higher levels of staff engagement and health and wellbeing
were associated with lower levels of mortality, as were staff
reporting support from line managers, well structured appraisals (e.g., agreeing objectives, ensuring the individual feels valued, respected and supported), and opportunities to
influence and contribute to improvements at work. NSS
[NHS Staff Survey] data also showed that staff perceptions
of the supportiveness of their immediate managers, the extent
of staff positive feeling, staff satisfaction and staff commitment were associated with other important outcomes, including
patient satisfaction. (Dixon-Woods et al. 2014, p. 112)
Let us compare these results with the simplified
model of health care that was suggested by Avedis
Donabedian, who observed that health care (1) outcomes are produced in (2) processes that are imbedded
in systems and their (3) structures. According to Donabedian (2003), “structure” is meant “to design the conditions under which care is provided” (p. 46).
Examples include material and human resources, the
presence of teaching and research, performance
reviews, etc. In our interpretation aspects like governance, leadership, learning mechanisms, organizational
climate, organizational design, intelligence, attractors,
incentives, norms, values, and similar “soft” issues
should be included. Process signifies “the activities that
constitute health care – including diagnosis, treatment,
19
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
prevention, and patient education. . . .” (Bergman et al.
2011, p. 41). Finally, outcome measures are the resulting
states from care processes, both technical (for instance,
absence of complications) and interpersonal outcomes
(for instance, patient satisfaction).
However, we also want to add how individuals in
the organization would react to different change interventions; that is, individual predispositions.5 It should
also be noted that we interpret the concept structure
rather freely, not as it is often interpreted as something
that is but rather as something that is in a constant
becoming and flux; see, for example, Sztompka (1991).
Indeed, this view on structure is pivotal to our view on
the health care system.6
With a serious ambition to improve outcome for
patients—that is, quality and safety—a science of
improvement has to address improvements on all levels
of the Donabedian model. It is not enough to make
improvements on the process level only. In addition,
soft aspects of the environment of the process (and its
microsystems) have to be addressed: How is the climate
for change generally and how are the individuals’ predispositions for change? How is patient centeredness
supported? How is intrinsic motivation of the staff supported? What are the norms and values? How is reflection and learning supported? It is only at such levels
that we can assure some sort of long-term survivability
of improvement efforts, see e.g. Docherty et al. (2003)
and Docherty and Shani (2008).
From this discussion, the following definition of
quality improvement suggested by Batalden and
Davidoff (2007, p. 2) is adequate; that is, quality
improvement is seen
. . . as the combined and unceasing efforts of everyone—health care professionals, patients and their families,
researchers, payers, planners and educators—to make the
changes that will lead to better patient outcomes (health),
better system performance (care) and better professional
development (learning).
When talking about system performance in the
above definition we should also include the predispositions for change (or readiness for change) of the
5
Individual predispositions were suggested by Walker et al. (2007)
as an important factor for the effect of change interventions in an
organization; see also Moore and Buchanan (2013).
6
In the language of researchers on change we have a process view;
see, for example, Pettigrew (1987) and Walker et al. (2007).
However, here we have a somewhat conflicting usage of the word
process; thus, we keep the concept structure and interpret it in a
more open way than it its often done in social sciences.
20
individuals within the organization. Depending on
past experiences, these predispositions could look very
different in different organizations. The organization’s
drive for change (i.e., improvements including innovations) should be a very important target of improvement efforts.
A SCIENCE OF IMPROVEMENT: A
BRIEF REVIEW
An important starting point of the quality movement was the seminal books by Shewhart (1931, 1939);
Shewhart was very much influenced by pragmatic philosophy (Lewis 1929), eventually leading to the much
used Plan-Do-Study-Act (PDSA) cycle and his thoughts
about predictability and corresponding criteria have
been central. Later, the experiences from the Japanese
wonder and the reawakening (Cole 1999) of the American industry were important. In health care, Ernest
Amory Codman was already a pioneer in advocating
the reporting of failures in order to encourage improvement before WWI. Later, after WWII, Avedis Donabedian developed the model referred to previously.
However, inspiration from the industrial quality movement hit health care in the 1980s, first with quality
circles and later in a much broader scope; see, for example, Berwick (1989), Berwick et al. (1990), and Batalden
and Stoltz (1993). In the late 1980s, the Institute of
Healthcare Improvement was founded by Don Berwick
and a number of his colleagues. Many more initiatives
started in the beginning of the 1990s.
Some important milestones were the roundtable
around quality that was arranged by the Institute of Medicine; see Chassin and Garvin (1998) and the subsequent
publications To Err Is Human: Building a Safer Health System (Institute of Medicine 1999) and Crossing the Quality
Chasm: A New Health System for the 21st Century (Institute
of Medicine 2001). These publications have had a great
impact; for example, in Sweden they changed the earlier dominant punitive7 view on health care errors and
quality dimensions of health care were identified as
those suggested by the Institute of Medicine (Swedish
National Board of Health and Welfare 2006).
In Sweden, M€
olndals Hospital (Al€ange and Steiber
2011) was a forerunner focusing on quality improvement of its processes; a clinic at Link€
oping University
7
Of course, there were also earlier many proponents for a
nonpunitive view on health care errors, but these views had not
reached the governments, legislation, and media.
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Hospital (Bergman and Klefsj€
o 2010) was one of the
first organizations, in competition with many otherwise
successful manufacturing companies, to win the Swedish Quality Award; and in J€
onk€
oping County they initiated a journey on quality improvement that is still
going on (Andersson-G€are and Neuhauser 2007).
Deming (1994) identified that for leaders to work
firmly in what he called the “New Economy” they have
to have some understanding of a number of important
interrelated knowledge areas. He emphasized “a system
of profound knowledge” and “appreciation for a system; knowledge about variation; theory of knowledge;
psychology” (p. 96). These knowledge areas and their
extensions will be discussed more in the next section.
Here we only mention that these areas were introduced
to the health care community in a paper by Batalden
and Stoltz (1993) under the concept improvement knowledge; see Figure 1. In a seminal book (Langley et al.
1996) on health care improvement the name improvement science seems to have been introduced for the first
time (Perla et al. 2013; see also Berwick 2008).
They also suggested, as illustrated in Figure 2, how
improvement of performance as experienced by
patients (3) is achieved by the application of generalizable scientific knowledge (1) to the special context
based on careful planning (4) and then a careful execution of the plan (5).
In Batalden and Stoltz’s approach, eight domains of
interest with associated tools and methods are of
importance. These domains are health care as processes
within systems; variation and measurement; customer/
beneficiary knowledge; leading, following, and making
changes;
collaboration;
social
context
and
accountability; and developing new knowledge. They
are similar to but formulated differently than the principles for organizing we present in the section The Normative Side of a Science of Improvement. In Sweden,
professional societies (nurses, doctors, nutritionists,
physiotherapists, work therapists, etc.) have joined
forces to create a common ground for professional education on improvement knowledge.
During April 12–16, 2010, a colloquium8 on the
epistemology of improving quality convened at Cliveden (see, e.g., Batalden et al. 2011). One important
outcome of this colloquium was the awareness and
importance of taking many different perspectives on
quality improvement in health care and to base training
programs on multiple epistemologies informing health
care improvement. Not only does traditional evidencebased medicine, resting on a natural science foundation, have to be called upon but also social science and
deeper understandings of social change processes.
Recently, Perla et al. (2013) suggested some theoretical and philosophical foundations on which “the science of improvement” rests. For example, learning and
the development of actionable knowledge is stressed
and they refer back to philosophical pragmatism, especially that of Lewis (1929) and his conceptualistic
pragmatism and subsequent interpretations and translations of Shewhart and Deming.
Concerns that improvement efforts too often are
lacking in a thorough scientific basis has been aired by,
for example, Marshall (2011) and Marshall et al.
(2013). There is also a tendency that too much focus
has been on implementation of solutions and on operational process improvement rather than on improvement and innovation of the health care system itself;
that is, improvement of the organizational abilities and
transformational changes of the health care system.
Another striking observation is the relative lack of a
patient’s perspective. Not that the patient is forgotten,
but quite a lot of the literature is more operations
focused than really patient focused. Works on the
patient perspective (Bergman et al. 2011), patient
empowerment (for a review, see Aujoulat et al. 2007),
patient involvement in health care design and improvement (Bate and Robert 2006; Gustavsson 2013) should,
we think, be much more in focus.
FIGURE 1 This figure is an enhanced version of the illustra-
8
tion by Batalden and Stoltz (1993). It illustrates how improvement
knowledge should be integrated with professional knowledge
for the improvement of outcomes.
Improvement in Health Care
The Vin McLoughlin Colloquium on the Epistemology of
Improving Quality, resulting in 23 papers downloadable from
http://qualitysafety.bmj.com/content/20/Suppl_1/i99.full.html#reflist-1
21
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
FIGURE 2 The “equation” for improvement as suggested by Batalden and Davidoff (2007) (adapted from Batalden and Davidoff 2007).
It is quite natural for an emerging field that its theories are discussed quite extensively and also that results
close to the very aim of the field are discussed. Educational aspects are, of course, also very important. What
would be needed, though, are more empirical papers
contributing to our understanding of how an organization’s capability and capacity9 for improvement and
innovation are enhanced.10 The NHS study described
earlier is a good illustration. A Ph.D. thesis on a similar
theme but limited to one hospital group including the
surrounding municipalities and the primary care centers has been presented by Lifvergren (2013).
IMPROVEMENT KNOWLEDGE
As emphasized in Figure 1 based on Deming’s “profound knowledge,” improvement knowledge takes a
central position in most descriptions of a science of
improvement. However, the different knowledge
domains have deepened since the days of Deming and
the need to broaden these areas has become obvious.
In the following we will discuss each one of the basic
knowledge domains separately, emphasizing their deepening and possible ways to extend them. However, we
will leave most of the “understanding variation”
domain until the next section.
Appreciation for a System
Some basic features of the understanding that
Deming drew on—for example the insight that enlarging the system boundary may give opportunities to better solutions to pressing issues—are still important and,
unfortunately, in these respects health care still has a
lot to do to integrate their different parts as well as to
integrate with other organizations outside health care
for the benefit of their patients (see the section
Understanding Variation). This is also an important
part of what Deming called win–win solutions.
However, our understanding of systems has
increased substantially since the days of Deming. The
theories of complexity (see, for example, Waldrop
[1992] and Stacey [2003]) and complex systems had
just started in the beginning of the 1990s. Recent
knowledge from complex adaptive systems research
shows that the complexity increases11 and thus the
manageability and predictability decreases (e.g., Axelrod and Cohen 1999; Stacey 2003). In complex systems
the concept of attractors—that is, what attracts the system to develop toward a certain set of states—is also
important for our understanding of how change efforts
may work. In order to try to understand social systems
better we also have to draw much more on social sciences in our basic improvement knowledge—we will
come back to that later in this section, giving the
knowledge area psychology a much wider definition than
in the original one suggested by Deming.
Understanding Variation
Deming emphasized reduction of variation (see also
Bergman 2003); however, there is much more to it. We
will come back to that later but let us first make a note
on the control chart, suitable for finding assignable or
(as Deming put it) special cause variation.12 Shewhart
emphasized the need to have predictable processes,
Shewhart’s (1939) definition of such processes is the
same (see Bergman [2009] and references cited therein)
as that of exchangeability used by de Finetti (1974).
Shewhart (1931) gave five criteria for such processes.13
One of these is based on the control chart—the one
that is most well known and for which Shewhart is the
11
It could, of course, be argued whether it is the complexity that
increases or our understanding thereof.
9
Note that an organization might have the capability to make
improvements but due to resources restrictions their capacity is
limited.
12
10
13
This is also emphasized by Marshall et al. (2013); see also Parry
et al. (2013).
22
The use of control charts in health care is discussed in, for example,
Langley et al. (1996), Benneyan et al. (2003), Woodall (2006),
Neuhauser et al. (2011), and Woodall and Montgomery (2014).
Interestingly, these have been graphically explained by Ishikawa
(1982, 1985) in the seven quality control tools.
B. Bergman et al.
most famous. Violations of these criteria may be used
to identify assignable causes. Specifically, if we have
made an intervention hoping for an improvement,
then that should result in an assignable cause indicating
a positive result. That used to be one of the cornerstones in the philosophy of Joseph Juran.14 We will discuss variation more in the next section.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Knowledge Theory
This theory originated in the conceptualistic pragmatic philosophy of Clarence I. Lewis, professor of
philosophy at Harvard University. It provided a background for the PDSA cycle (see, for example, Mauleon
and Bergman 2009; Peterson 1999; Wilcox 2004). In
addition, it is emphasized that theory is the basis for
prediction—prediction is the basis for action and if the
action does not lead to the predicted result there is a
need for reflection and revision of the theory (i.e.,
learning) in order to perform better actions in the
future. Lewis emphasized that knowledge is for action
(Lewis 1929). Indeed, this is what Dixon-Woods, Bosk
et al. (2011) wrote in their reflections on the Michigan
case. Furthermore, when different humans observe the
same phenomenon, they may experience it very differently due to their a priori understanding. This understanding sometimes leads to conflicts in a group if it is
not understood that the differences depend on different mental models, as Peter Senge (1990) would state it.
Closely related is how we as individuals create meaning—and even more important how individuals transform their ways of creating meaning—something that is
important in many quality improvement efforts, especially on the structural and system levels in the Donabedian model. When there is a need for second-order
type of learning, how is that achieved? We have to call
in much more from current research around meaning
creation, transformative learning (Mezirow & Associates 2000), and learning to learn (Argyris 2003; Visser
2003). Note that first-order learning means that we
learn something within our earlier understanding of
reality, and second-order learning means to change our
way of understanding reality, to get a new way of creating meaning. Third-order learning means achieving a
way of reflecting on self and question the current way
of creating meaning. In this learning, reflection is an
important process: to use a statistical metaphor, the
Bayesian learn more and more about the current model
when the given likelihood function changes due to an
increasing amount of data (single loop learning). However, to change the shape of the likelihood function or
recognizing that no likelihood function exists due to
assignable causes of variation (not in themselves under
statistical control; see Bergman and Chakhunashvili
2007) is an illustration of a kind of second-order learning. Within this metaphor, third-order learning should
mean to have a systematic approach to find out whether
the current assumed likelihood function really is relevant
for the problem at hand or not. Usually, the above forms
of learning are taken as individual learning; however, in
an organization it is important to attach these types
of learning to the organization. Learning is the achievement of knowledge—knowledge is for action—and common knowledge is for common action.15
Psychology
What Deming emphasized here was the importance
of intrinsic motivation. Currently, there is much more
evidence about the importance of intrinsic motivation
and how that has been put into systematic use in some
organizations. A popular book by Dan Pink (2010)
illustrates this nicely. Closely related to research results
on intrinsic motivation and self-determination theory
(see the introductory chapter of Deci and Ryan 2002)
is the new research field called positive psychology (see,
for example, Csikszentmihalyi and Nakamura 2011)
with an interesting forerunner, Mihaly Csikszentmihalyi (1990), and his research on creativity and the concept of optimal experiences, also known as flow.
Under the heading psychology, Deming emphasized
not only individual psychology but also interactions
between people. This should lead us into a much
broader research field to integrate into a science of
improvement. Indeed, a lot of social science should be
incorporated here; there might be a need for a new
heading for this broad knowledge field, which has been
proven to be of great interest to a science of improvement (see, e.g., Dixon-Woods, Bosk et al. 2011). For
example, an important research tradition with much in
common with the quality movement is action research;
15
14
This is called the Juran trilogy: design–control–improvement; see
Juran (1986).
Improvement in Health Care
This is a reformulation of a comment in Lewis (1929); an
interesting framework connecting the individual with the
common is given by Crossan et al. (1999).
23
see, for example, Lewin (1948) and Coghlan and Brannick (2010). For some recent applications within health
care see, for example, Lifvergren, Gremyr et al. (2010),
Lifvergren, Docherty et al. (2010), Lifvergren et al.
(2011), Lifvergren et al. (2012), Gustavsson (2014), and
Lifvergren (2013).
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
A System of Knowledge
Deming emphasized that profound knowledge is a
system of knowledge—the different parts are interrelated. Even though Dixon-Woods, Bosk et al. (2011) do
not explicitly use this conceptualization in their study
of the Michigan project, it is an excellent illustration.
Even though Deming’s profound knowledge serves
well as a starting point, we now need a broader set of
basic scientific knowledge—perhaps under different
headings. Indeed, this is underway even if the headings
have remained the same. However, it is important to
understand that the theories we use in our interpretations are important for our continual learning how to
better achieve our organizational goals; as Deming
(1994) said, “Experience by itself teaches nothing. . . .
Without theory, experience has no meaning. Without
theory, one has no questions to ask. Hence, without
theory, there is no learning” (p. 106).
UNDERSTANDING VARIATION
In the Beginning There Was Variation
Already with the Big Bang variation was there—if it
had not, we would probably still have a homogenous
universe—and no “we”! And so it goes on—heavier
and heavier atoms, the creation of stars and galaxies,
and eventually some things that replicate, vary, interact, and compete for energy—the organic life has
begun.16 And eventually intelligent creatures who create ideas and thereby a new evolution starts. And here
we are—the results of a long chain of results from variation, selection, and interaction.17 Indeed, variation is a
central part of our lives.
In discussions of variation it is important to see differences and similarities between two different kinds of
variation: variation between units when time is not in
16
Richard Dawkins has written many popular books on this theme,
starting with The Selfish Gene in 1976.
17
Some interesting references are Dawkins (1976); Jablonka and
Lamb (2005) on evolution in four dimensions, and Dennett (1995).
24
FIGURE 3 Number of sudden infant death syndrome cases; a
reasonable assumption is that these deaths occur as in a Poisson
process.
focus, or synchronic variation, and variation over time,
or diachronic variation. Too often, only the synchronic
variation is discussed, though in many cases diachronic
variation is the most interesting—how is it possible to
predict future outcomes given the data? Did an intervention give the intended result? Much medical
research in terms of randomized controlled trials
emphasizes synchronic variation. A stronger emphasis
on diachronic variation would be an important contribution to medical research. And, of course, diachronic
variation is of great importance for the improvement
of health care. An illustration where a quality improvement intervention seems to have had a very important
effect is illustrated in Figure 3. For some further reflections see the following subsection.
Variation in Health Care
We are all different—patients are all different. Most
often health care professionals have a good sense for
that—meeting each patient here and now on the conditions of this patient. Thus, on an individual level there
is often a deep understanding of variation. However,
when it comes to clinical research, variation has often
been kept under the carpet—what have been counted
are averages—mean values. At least, that has very often
been the case in randomized controlled clinical studies,
known as the golden standard of clinical research. Variation is used mostly through the standard error of the
mean to judge whether mean values are statistically significant or not. If it is not mean values it is often percentages of individuals reaching a certain target; HbA1c
is a typical case18 where changes have been made a
18
HbA1c measures the long-range blood sugar level and is an
important measure in diabetes care.
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
number of times—not only target values but also measurement methods. This is, of course, very problematic—target values change and thus operating definitions
change and possibilities to compare with history are
lost. This may partly be due to a lack of insight concerning diachronic variation.
Nowadays, with the insights and the possibilities for
genetic testing, an increasing understanding for variation and that randomized testing is not always the best
answer is emerging. But genetic variation is just one
source of variation—also important are epigenetic variation, behavioral variation, and perhaps even cultural
variation (Jablonka and Lamb 2005). These types of
variation must be taken into account. That is often not
done in the current clinical research paradigm. As
many interventions and improvement activities have to
be adopted in a social system, the variation in contexts
is very important and can explain sometimes confusing
results and large variations; see, for example, Pettigrew
and Whipp (1991), Bate (2014), and the Health Foundation (2014). Today, there is quite a lot of focus on
quality registers and thereby a potential to evaluate
new interventions. From the registers random selections taking different factors into account—the use of
design of experiments based on registers and follow-up
through the registers and then subsequent analysis of
results—seems to be a very interesting possibility not
yet utilized.
Quality Registers
The world’s first national quality register was created
by Ove Guldberg Høegh to register patients suffering
from leprosy in Norway 1856. In the beginning of the
20th century, Ernest Amory Codman created a register
on shoulder surgery. However, it was not until the end
of the 20th century that registers for a large number of
diseases became common. In Sweden, we have around
100 such registers today—each recording a number of
different diagnoses and corresponding process and outcome measures. They were first created by enthusiastic
physicians to get better hold of their specialities as well
as for research purposes. An illustration is the first register in Sweden, a knee replacement register, initiated in
1975 and shortly after followed by a hip replacement
register. Due to the registers it has been possible to
increase reliability of these types of replacements with
great benefits both for patients and for replacement
Improvement in Health Care
costs. The situation in Sweden, where all citizens have
a unique personal number, is especially beneficial for
research. It is, for example, possible to combine results
from different types of registers. Of course, efforts are
needed to assure that integrity aspects are taken into
account.
Registers are increasingly thought of as a huge possibility for improvement activities. However, it has still
very much stayed at a rhetoric level and much has to
be done to educate health care professionals to utilize
these registers in their daily improvement work (see,
e.g., Bergman 2013). Among the best registers to promote improvement work is the diabetes register, where
today it is possible to get information online concerning different hospitals and their results with respect to
smoking cessation, blood pressure, HbA1c, patient
recorded outcome measures, etc.
Reflections on Variation, Statistics,
and the Theory of Probability
As we observed in an earlier section, variation is a
basic concept in our understanding of the universe. In
this section, we want to explore further some of the
basic aspects of knowledge theory and variation. When
observing the world we find variation, and that gives
rise to uncertainty when we try to make predictions
about the future—and, in a way, such predictions lie
behind our decisions and subsequent actions, even if
not always as rational as we would like to think; see, for
example, Kahneman (2011). In this way of looking at
things, which is very much in line with the ideas of C.
I. Lewis that influenced Deming and Shewhart, probabilities describing our uncertainty exist in our theories
about the world, not in the world itself. In fact, that
leads us to a subjective (or Bayesian) interpretation of
probabilities as promoted early by Ramsay (1926), de
Finetti (1974), and Savage (1954). Bayesian statistics
has gained a lot of attention in medical statistics lately
(see, e.g., Ashby 2006).
The Subjectivist Interpretation of
Probabilities and Conceptualistic
Pragmatism
Lewis (1929), in his conceptualistic pragmatism,
emphasized that theories are a priori, which however,
we are free to choose. He also emphasized that it is
only in the light of this a priori that we experience what
25
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
happens in the world. According to Lewis, experiences
are an interpretation of the sensuously given based on
the a priori. It is indeed also a basic ingredient in our
possibility to reflect upon and learn from our experiences. We find a striking similarity between the thinking
of Lewis and that of Bayesians.
There is another interesting connection between
Bayesian statistics and traditional quality improvement
a la Shewhart and Deming. The control chart makes an
interesting bridge investigated by Bergman (2009).
When Shewhart (1939) made a formal definition of a
process under statistical control—that is, about predictability—he used a criterion that in the Bayesian/
subjectivist literature is called exchangeability.19 It was
introduced by de Finetti (1974) as a prerequisite for a
predictive distribution to exist. However, in the Bayesian/subjectivist literature the judgment of a process
being exchangeable is left to the decision maker and no
tools are provided. However, here the Shewhart control
chart comes in handy. It is designed exactly as a tool for
judging predictability of a process. We will come back
to this in the next section. First, however, we will make
an important note that is sometimes forgotten when
discussing the contributions of Shewhart. He not only
suggested the control chart to be used for judging if a
process is under statistical control, but he also pointed
to four more criteria to be fulfilled for a process to be
free from assignable causes of variation.20 Four of these
five criteria of Shewhart’s were later transformed into a
graphical form by Kaoru Ishikawa (1982) to form some
of the well-known seven QC tools: histogram, stratification, scatter diagram, and control chart.
Control Charts
When Shewhart introduced the control chart in his
seminal book from 1931 he recognized that not all processes were predictable in a statistical sense—sometimes assignable causes of variation interfered and
predictability was lost. It is important to identify such
assignable causes of variation and to eliminate them in
order to get a predictable process. But not all predictable processes give an acceptable result. We have to
make more drastic improvements in the process, not
19
It could also be called permutability—each permutation of the
stochastic variables has the same distribution.
20
Note that sometimes the process may be predictable even
though assignable causes exist.
26
FIGURE 4 The quality trilogy (Planning, Control, Improve)
suggested by Juran (1986) with a more refined division in phases
I-IV.
only to remove assignable causes of variation but to
more radical changes in the process/system. This was
illustrated in what Juran called the Juran quality trilogy (Juran 1986): planning a process, controlling it
(and take away assignable causes of variation), and
then improving the process. After a successful
improvement we go into a control phase again; see
Figure 4.
As seen in Figure 4, the control chart works in three
different ways. First, to judge whether a process is predictable, take away assignable causes, estimate parameters, etc. (phase I); second, to survey the process (phase
II) to guard against new assignable causes that hopefully will be eliminated to give a predictable process;
and thirdly, to work as a confirmation that an intervention, a change of some kind, really was an
improvement (phase III). It should be mentioned
that phase III is usually not mentioned in a systematic way in the literature on control charts even
though it should be regarded as a very important
aspect of the use of control charts. In Figure 3 on
sudden infant death syndrome we saw that a substantial improvement in the process had occurred.
The occurrence of an assignable cause of variation
of a positive kind could be taken as a strong indication that the intervention (in this case a campaign)
really was successful. Of course, it is important to
make sure that no other changes have been active
that could have affected the process.
In health care today, both process measures and outcome measures are documented in quality registers.
This provides us with an excellent possibility to evaluate improvement interventions—if the relevant measures come from a process under statistical control it is
possible to conclude whether an intervention really is
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
an improvement or not because it should show up as a
positive assignable cause. In many respects, such a strategy for evaluating an intervention has many advantages
over the usual gold standard of clinical medical
research; however, we will not go further into this
debate here.
When reflecting on the ideas behind Figure 4 it
becomes obvious that the usual phase I and phase
II descriptions of the work with a control chart are
not quite enough. A phase III should be added
when improvements are in focus. The special assignable cause that might have occurred due to an intervention usually takes some time to mature (if
positive as intended). How long do we need to run
the process until we can judge predictability (and
the new level)? A possible phase IV should be very
similar to phase II. We will come back to this later
in this section.
A Bayesian Control Chart
In the literature on Bayesian control charts we do
not see the type of control charts that Shewhart advocated; that is, a chart for judging whether a process is
under statistical control or not. Rather, the state under
statistical control is determined from the beginning
and an out-of-control state is defined. The argument
for the name Bayesian is because these states are given
subjective probabilities and if the out-of-control state
gets a high enough probability an out-of-control situation is declared. In fact, the out-of-control state is in
itself assumed to be a process under statistical control.
Depending on the purpose, this might be perfectly all
right. However, we just want to determine whether the
process is under statistical control or not; that is,
whether it is reasonable to judge it predictable or not.
Based on the nature of the process, some sort of model
might be judged if the process is under statistical
control.21
When we observe the process, we also increase our
knowledge about the process and increase our possibility to predict what will happen in the future of the
process—at least if it is under statistical control; that is,
still predictable. But how do we judge that? A simple
way would be that the observation at time t C 1 is
21
Such an assumption could be based on a judgment concerning
statistics that are judged to be sufficient.
Improvement in Health Care
within what could be called the predictive interval as
judged from observations up to time t. And that should
also be the case for all of the past observations due to
the fact that the process under statistical control is
exchangeable.22 The only thing to find out is how the
predictive interval should be calculated. In the spirit of
Shewhart it may be taken as the 3s limits of the predictive distribution. However, just as Shewhart indicated,
another choice could be relevant depending on the situation. It should be noted that it is necessary to go
back also to old observations because in the beginning
of the observation period the knowledge about the process is not as strong as later on. That means that an
observation at time s might be within the predictive
interval based on observations up to time s ¡ 1 but
when time has passed to time t >> s we know much
more and therefore the observation at time s may be
outside the new predictive interval and judged to be
due to an assignable cause of variation. This is an interpretation of what Lewis (1934) meant by “[. . .] knowing begins and ends in experience; but it does not end
in the experience in which it begins” (p. 134). Based on
more knowledge, we may reinterpret the history that
gave us the knowledge we have.
We still have a task to think about how the process
should be handled after an improvement has been
made and found effective as described by the Juran trilogy. In fact, we could proceed just as before—we have
an improvement and we have to reformulate a priori in
order to predict the future. When starting to observe
the process it was natural to assume a noninformative a
priori distribution. However, after observing a positive
effect of an intervention it is not quite clear how a
renewed a priori would look like. A pessimistic choice
would probably be to use a noninformative a priori
before the observation that has just been seen to be
outside the predictive limits—most probably due to
the intervention made—and then start anew from
there. As is often the case, the process has not yet
reached a steady state and thus it would be natural to
find the next point outside its predictive interval. The
procedure goes on until a new steady state may be
concluded.
22
A small problem is that the prediction interval for an observation
at time s < t is based on all observations including the one we want
to have a predictive interval for. If a recalculation should be made
for the predictive interval for the observation at time s, utilizing all
observations but that at time s is a matter of being purist or not.
27
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Evidence-Based Medicine
The gold standard of clinical research has been the
randomized controlled trial for a long time. However,
some problems are surfacing. Too much focus has
been on differences in averages (or median), though of
course variation is also very important, as well as factors
underlying that variation. The Plavix case is an interesting illustration—due to genomics, 30% of patients do
not have the possibility to transform the drug into its
active component. Emphasis is also almost always on
one-factor-at-a-time experimentation. Furthermore, the
population used for the test was different from the one
in which the results will be utilized, and the controls
may be different and not adequate for the application
area. For a long time treatments that have been shown
to have efficacy for one type of population have been
used for a very different population (for example, the
elderly or small children). Thus, much more emphasis
should be put on methods utilized in industrial settings: design of experiments, control charts, data taken
from the treatment processes, etc.
Control charts are an excellent illustration: Assume
that an intervention to a process under statistical control has been performed (after Phase II in Figure 4),
and that after the intervention the process again is
under statistical control but on a better level (better
mean, lower standard deviation etc.). Then there are
good reasons to assume that the intervention had a positive effect on the process, at least if no other changes
to the process are found.
We will not go further into these interesting areas
here, but there is a great potential in letting improvement methodologies stimulate the development of
complementary medical research methodology.
formulated as a value or a principle for organizing. As a
support to this main principle some other principles
have been promoted (see, e.g., Anderson et al. 1994;
Bergman and Klefsj€
o 2010; Cole and Scott 2010; Dean
and Bowen 1994; Hackman and Wageman 1995; ISO9000, International Organization for Standardisation
(ISO) 2000). As we emphasized earlier, it has to address
operational improvements not only at the operational
process level but on the organization as well as on the
overall health system level. An illustration is given in
Figure 5. The different principles are related to those
presented in Bergman and Klefsj€
o (2010) but with a
slightly different wording. These and similar principles
have been suggested in many other publications, in
research papers on the quality movement, and in more
popular books as well as in ISO-9000. Sometimes these
principles are formulated as values. Here we prefer to
talk about principles for organizing—principles that
should be possible to test for when decisions and other
kinds of actions are taken in the organization. Do they
follow or are they breaking against these principles? In
the next section we will discuss these principles. Then
we will discuss the relations between principles, practices, and tools.
Principles for Organizing
In Figure 5 the principles advocated in Bergman and
Klefsj€
o (2010) are displayed but with slightly different
formulations. Below, we only give a very brief description—more detailed descriptions are found in textbooks like Bergman and Klefsj€
o (2010).
THE NORMATIVE SIDE OF A SCIENCE
OF IMPROVEMENT
A science of improvement is not a science that is
purely objective and without a purpose; on the contrary, its purpose is quite clear. The knowledge needs
that a science of improvement has to fulfill are how
improvements are achieved and sustained. In a health
care context, it should be understood that improvements are directed toward health in society: For
patients here and now and in the future and, more generally, for the health of the citizens here and now and
in the future. In the quality movement this has been
28
FIGURE 5 Principles for organizing; see Bergman and Klefsjo€
(2010).
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Take a Customer Perspective
This means that we always should try to understand
things from the patient’s point of view and also to support the empowerment23 of the patient. We judge this
to be a huge revolution in health care because better
informed patients are able to be equal partners in the
curing and caring processes, resulting in higher quality
and lower costs. Of course, not all patients have the
possibility, but the lower the resource needs for capable
patients, the more there is for those who are not as
capable and therefore in need of more support. It is
also worth considering that health care in a strict meaning
seldom create direct value to the patients—the value is
created in the life world of the person who is, has been,
or is close to the current or former patient. Health care’s
contribution is of cause very important, but it is only one
of the components in the creation of value.
As observed by Normann (2001), it seems to be a
general trend in different industries to change the way
customers are looked upon. First, customers are looked
upon as only consumers on a market, mass production
is in focus and the decisive competence is production
competence. Later on, customers are seen as a source of
information (i.e., what is quality from a customer point
of view), a customer basis becomes important and the
decisive competence is the ability to build relations
with the customers. In the final stage, the customer is
seen as a co-producer and even as a co-creator of the
value creation process. Value creation networks are in
focus and the creation of such networks becomes a
decisive competence.
Patient empowerment, co-production of care, and
even in some cases codesign of care are currently discussed.24 A case of the last kind has been performed as
an action research project on the rebuilding of neonatal
care at the Skaraborg Hospital Group (SkaS). The
method applied has been that of experienced based
co-design developed by Bate and Robert (2006) based
on concepts from the design industry. The method is
based on four key steps: capturing experiences from
patients and staff, understanding these experiences,
improving based on these experiences, and finally measuring the results. In all of these steps patients are
23
Patient empowerment has grown to a very intense area of both
practical applications and theoretical research; see, for example,
Nordgren (2008), Ekman et al. (2012), and Gustavsson (2013).
24
For illustrations of how patients have been involved in the
creation of new health care solutions, see, for example, Bate and
Robert (2007), Iedema et al. (2010), and Maher and Baxter (2009).
Improvement in Health Care
actively working first in their own group (the first and
part of the second step) and then together with the
staff. The experiences from the project at SkaS led to a
number of changes—experiences from the patients
indicated improvement possibilities of not only trivial
kinds but regarding quite complex issues such as the
cooperation between different wards, which showed
new ways to coordinate work. Interestingly, the issues
noted by the patients differed markedly from those of
the staff (for a discussion of this and similar projects
see Gustavsson 2013).
Use Intelligence for Decisions about Future
Actions
How do we know about our customers preferences
and needs and wants—how are our decisions in line
with the life situation of the patient? Understanding
the situation of the patient is an important starting
point even though in decisions regarding treatment
many other factors have to be taken into account. It is
also important to understand the value creation processes and their functioning. How do we measure and
with which results? But also intelligence concerning
the processes and results of providers that are similar
to us. From whom is it possible to learn? How do we
keep track of new research results of importance to
our activities? Earlier (Bergman and Klefsj€
o 2010) we
called this “base decisions on facts”; however, replacing this with “intelligence” is a quest for an active
process for gathering necessary information of importance for decisions. In addition, a focus on a high
degree of intelligence is pivotal for learning in the
organization.
Focus on the Value Creation Processes
The value creation processes in a system is all of those
activities that repetitively create value. These activities
may be value streams as emphasized in Lean manufacturing, more complex networks of activities, or shop solutions.25 Value is more and more created in networks of
individuals—it has to be emphasized that these networks
not only consist of the patient and other patients and
health care personnel but also of others in society. For
25
Unfortunately, the concepts here have become blurred—in many
treatises processes are thought of as linear streams of activities;
that is, value streams. Here it is interpreted as all of those
arrangements that repetitively create value.
29
further discussions on processes see e.g. Hellstr€
om
(2007), and Hellstr€
om et al. (2010).
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Improve Continuously
Improvement is, of course, at the heart of a science
of improvement but is also very demanding for an
organization that has to be able to, at the same time,
exploit the knowledge, capabilities, and capacities to
deliver value to its customers. Simultaneously, the
organization must explore new ways of working both
in small incremental improvement steps but also in
more innovative ways. Improvement in many cases
means innovation—it is important for health care to
rethink its capability and capacity for innovation; see
also Dixon-Woods, Alamberti et al. (2011). Continual
improvement is also very close to organisational learning.26 From a management point of view it is important
to understand and help the organization to build
learning mechanisms of different kinds: cognitive,
structural, and procedural learning mechanisms (see
Shani and Docherty 2003).
Support Employee Intrinsic Motivation
Earlier we discussed the importance of intrinsic
motivation. As described by Dixon-Woods et al.
(2014), most people in health care really want to help
their patients—they have a strong intrinsic motivation.
However, this is not always a principle that has been
used for organizing, which is a serious waste.
Take a Systems Perspective
Everything goes on in systems—today’s health care
is a very complex system that might sometimes make it
unpredictable. System understanding should be utilized as a basis for decision making.
Lead According to All of the Other
Principles
Good management and visionary leadership are
important assets in all organizations. The support of
26
Learning and organizational learning are a controversial area
from a theoretical point of view; what should be meant by double
loop, triple loop, or deuteron learning are discussed by Argyris
(2003), Visser (2003), and Tosey et al. (2012). A related concept is
that of transformative learning as discussed by Mezirow and Associates (2000).
30
employees’ inner motivation is an art. It is also a radical
shift from many contemporary leadership and management ideals.
It should be noted that these principles are close to
those advocated in ISO-9000 (2000) and not very far
from what Anderson et al. (1994) described as concepts
in a theory underlying the Deming management
method.
Improvement Domains
It is easy to see that the principles described above
are closely related to the domains described in the brief
review above. The reason we have chosen to describe
similar things as principles rather than domains is to
clearly show the normative side of a science of
improvement. Of course, the selection of domains is
also governed by a normative will. However, it is not
very clearly expressed.
Principles, Practices, and Tools
A large number of practices and tools are related to
the principles and are necessary for the principles to be
applied in everyday work. In many treatises on quality
management, practices are not as clearly stated as
opposed to principles and tools. One reason for this
may be that practices have to be tailored to the context
of the organization and its operation. On the high level
of principles it is easy to give these principles a nearuniversal applicability. The same goes for the precise
tools, whether quantitative or qualitative. However,
when it comes to practices—the actual work processes
applied in an organization—they have to be tailored to
the specific situation and the tasks to be performed in
the organization. To give universally valid descriptions
of work practices would be hard even though recommendations could be given.
REFLECTIONS
When reflecting upon an emerging science of
improvement in health care, a number of issues arise.
Some of these reflections are very practical with respect
to the development of health care in the future and
some more philosophical. Only a few points will be
given an initial reflection and some important questions and agendas for future research will be raised.
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Practical Reflections
Reflecting on the challenges that health care faces,
it could be asked whether it is really enough with a
science of improvement. Will the use of improvement activities alone solve the problem or perhaps
only lock health care even stronger in the current
dead-end trajectory? Probably not. But technologies
exist that could provide us with a completely different health care, but there seems to be no strong pull
for these new technologies—health care is the last
industry taking information technology solutions to
heart. With a stronger focus on improvement
issues—areas for reflection and an awareness of the
possibility to find better ways—we believe that the
ability to adopt new ways of working utilizing new
technologies will be improved. Transformative learning is needed at all levels of health care and sometimes even on the patient side. But how is
transformative learning supported in a health care
system that is fully occupied with today’s problem
(and sometimes yesterday’s omissions)?
As emphasized by Walker et al. (2007), working with
improvements makes it easier to adopt more improvements and even disruptive innovations (see also Moore
and Buchanan 2013). Today, skilful people work so
hard that they have no time to find solutions just
around the corner. “Best efforts are not enough”
(Deming 1986, p. 19).
Philosophical Questions on an
Improvement Science
Is a science of improvement really a science? What is
a science? What is its ontology? What about its epistemology? How is it related to other kinds of sciences?
What about its universality? As a combination of different knowledge areas, it could possibly be characterized by what Gibbons and collaborators (1994) called
Mode 2 research. Many of its methods have been
inspired by statistical improvement from other areas of
application, but it is also about social change. How are
these different ontologies and epistemologies combined in noncontradictory ways?
If we look upon health care as a service industry, we
should perhaps learn more from the knowledge area of
service management. It is often emphasized that value
is not created by the deliverer only but together with
the customer and in the customer’s own processes.
Improvement in Health Care
This seems to be very much related to issues in the
patient empowerment movement. How far can we go
in that direction?
What about fashions and fads? Many researchers of
the critical management school are talking about the
problems with what they call “new public management.” Is there something to take care of and be worried about from their critique? How is that done in a
constructive way? And whatever theories and sciences
we come up with, they are in a sense socially constructed—they cannot be otherwise (Hacking 1999). But the
real test is their usefulness—still a science of improvement is only in an initial phase, but it is important that
its usefulness is proven not only on local scales but on
more universal ones!
. . . and the world moves on—who wants to be left
behind?
ABOUT THE AUTHORS
Bo Bergman is partly retiring from a chair in Quality
Sciences at Chalmers University of Technology and
currently also a guest professor at Meiji University,
Tokyo. His career started with 15 years in aerospace
industry during which period he also became a PhD in
Mathematical statistics from Lund University 1978 and
was a part time professor in Reliability at Royal Institute of Technology, Stockholm. In 1983 he became a
professor in Quality Technology at Link€
oping University and 1999 the SKF professor in Quality Management at Chalmers University of Technology. He was a
co-founder of Centre for Healthcare Improvement
(CHI) at Chalmers and its first director (2004–2009).
As a professor he has supervised a large number of PhD
students many of which now are themselves professors.
He is a member of the International Statistical Institute
(ISI) and an Academician of the International Academy
for Quality (IAQ).
Andreas Hellstr€
om is a senior lecturer at the Department of Technology Management and Economics at
Chalmers University of Technology, Sweden and codirector of Centre for Healthcare Improvement (CHI)
at Chalmers – a research and education center focusing
on quality improvement, innovation, and transformation in health care. In 2007, he successfully defended
his Ph.D. thesis on the diffusion and adoption of management ideas. In 2013, he was one of four selected in
the first cohort of Vinnvard’s Improvement Science
31
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Fellowship. He has led research projects in both the private sector and the public sector.
Svante Lifvergren, MD, is a specialist in internal and
respiratory medicine. He is quality director at the Skaraborg Hospital Group (SkaS), Sweden, having also
worked as a senior physician at SkaS since 1998. Further, he holds a Ph.D. in Quality Sciences and he is
co-director for the Centre for Healthcare Improvement
(CHI) at Chalmers University of Technology. His
research mainly focuses on organizational development
in complex healthcare organizations.
Susanne M. Gustavsson, RN and Midwife, is Nursing Director at the Skaraborg Hospital Group (SkaS),
Sweden. She is also an external PhD student at the
Centre for Healthcare Improvement (CHI) at Chalmers University of Technology, Gothenburg Sweden.
Her research is focused on experience-based co-design
with patients and relatives in health care processes.
ACKNOWLEDGMENTS
First, we want to thank all those devoted health care
personnel who have participated in our research and in
our courses and have given us tremendous support. In
particular, the Skaraborg Hospital Group has been very
important as well as the V€astra G€
otaland Region.
FUNDING
We want to acknowledge the important support
from the research foundation VINNVARD, providing
resources for our health care improvement research
journey.
REFERENCES
Al€
ange, S., Steiber, A. (2011). Diffusion of organisational innovations: an
empirical test of an analytical framework. Technology Analysis &
Strategic Management, 23(8):881–897.
Anderson, J. C., Rungtusanatham, M., Schroeder, R. G. (1994). A theory
of quality management underlying the Deming management
method. The Academy of Management Review, 19(3):472–509.
Anderson, R. M., Funnell, M. M. (2005). Patient Empowerment: reflections on the challenge of fostering the adoption of a new paradigm, Patient Education and Counceling, 57: 153–157.
Andersson-G€
are, B., Neuhauser, D. (2007). The health care quality journey of Jonkoping County Council, Sweden. Quality Management
in Health Care, 16(1):2–9.
Argyris, C. (2003). Retrospective organizational learning: A theory of
action perspective. Academy of Management Executive, 17(2):
37–55.
Ashby, D. (2006). Bayesian statistics in medicine: A 25 year review. Statistics in Medicine, 25:3589–3631.
32
Aujoulat, I., d’Hoore, W., Deccache, A. (2007). Patient empowerment in
theory and practice: Polysemy or cacophony? Patient Education
and Counseling, 66:13–20.
Axelrod, R., Cohen, M. D. (1999). Harnessing Complexity Organizational
Implications of a Scientific Frontier. New York: Free Press.
Batalden, P. B., Davidoff, F. (2007). What is “quality improvement” and
how can it transform healthcare? Quality and Safety in Health
Care, 16:2–3.
Batalden, P., Davidoff, F., Marshall, M., Bibby, J., Pink, C. (2011). So
what? Now what? Exploring, understanding and using the epistemologies that inform the improvement of healthcare. BMJ Quality
and Safety, 20:i99–i105.
Batalden, P. B., Stoltz, P. (1993). A framework for the continual
improvement of health care; building and applying professional
and improvement knowledge to test changes in daily work.
The Joint Commission Journal on Quality Improvement, 19
(10):432–452.
Bate, P. (2014). Context is everything, 1–30. London: The Health
Foundation.
Bate, P., Robert, G. (2006). Experience-based design: From redesigning
the system around the patient to co-designing services with the
patient. Quality and Safety in Health Care, 15:307–310.
Bate, P., Robert, G. (2007). Experience-based design and a case study
toward more user-centric OD: Lessons from the field of experiencebased design and a case study. Journal of Applied Behavioral Science, 43(41):41–66.
Benneyan, J. C., Lloyd, R.C., Plsek, P. E. (2003). Statistical process control
as a tool for research and healthcare improvement. Quality and
Safety in Healthcare, 12(6): 458–464.
Bergman, B. (2003). On the profound knowledge of variation. In The
Best on Quality Book Series of the International Academy for
Quality, 2–13. Milwaukee: American Society for Quality, Quality
Press.
Bergman, B. (2009). Conceptualistic pragmatism: A framework for
Bayesian analysis? IIE Transactions, 41:86–93.
€
Bergman, B. (2013). Annu
Ba€ttre Vard–Vad Kan vi La€ra fran Variationen i
€
€
€
Oppna Jamforelser? [Even better care—What can we learn from
variation in the Open Comparisons?]. Stockholm, Sweden: SKL.
Bergman, B., Chakhunashvili, A. (2007). In weak statistical control? International Journal of Six Sigma and Competitive Advantage, 3(1):
91–101.
€ , B. (2010). Quality from Customer Needs to CusBergman, B., Klefsjo
tomer Satisfaction. 3rd ed. Lund, Sweden: Studentlitteratur.
Bergman, B., Neuhauser, D., Provost, L. (2011). Five main processes in
healthcare—A citizen perspective. BMJ Quality and Safety, 20:i41–
i42.
Berwick, D. M. (1989). Continuous improvement as an ideal in health
care. New England Journal of Medicine, 320(1):53–56.
Berwick, D. M. (2008). The science of improvement. JAMA, 299
(10):1182–1184.
Berwick, D. M., Blanton Godfrey, A., Roessner, J. (1990). Curing Health
Care. New Strategies for Quality Improvement. 1st ed. San Francisco: Jossey-Bass.
Best, M., Neuhauser, D. (2004). Avedis Donabedian: Father of quality
assurance and poet. Quality and Safety in Health Care, 13:472–473.
Chassin, M. R., Garvin, R. W. (1998). The urgent need to improve health
care quality. Institute of Medicine National Roundtable on Health
Care Quality. JAMA, 280: 1000–1005.
Coghlan, D., Brannick, T. (2010). Doing Action Research in Your Own
Organization. 3rd ed. Thousand Oaks, CA: Sage.
Cole, R. E. (1999). Managing Quality Fads. How American Business
Learned to Play the Quality Game. Milwaukee, WI: ASQ Quality
Press.
Cole, R. E., Scott, W. R. (2000). The Quality Movement & Organization
Theory. London: Sage.
Crossan, M. M., Lane, H. W., White, R. E. (1999). An organizational
learning framework: From intuition to institution. The Academy of
Management Review, 24(3):522–553.
B. Bergman et al.
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience. New York: Harper and Row.
Csikszentmihalyi, M., Nakamura, J. (2011). Positive psychology: Where
did it come from, where is it going? In Sheldon, K. M., Kashdan, T.
B., Steger, M. F., Eds. Designing Positive Psychology. New York:
Oxford University Press, 2–9.
Dawkins, R. (1976). The Selfish Gene. Oxford: Oxford University Press.
Dean, J. W., Bowen, D. E. (1994). Management theory and total quality:
Improving research and practice through theory development.
Academy of Management Review, 19(3):392–418.
Deci, E. L., Ryan, R. M. (2002). Handbook of Self-Determination Theory.
Rochester, NY: University of Rochester Press.
de Finetti, B. (1974). Theory of Probability. Vols. 1–2. New York: Wiley.
Deming, W. E. (1986). Out of the Crisis. Cambridge, MA: MIT Press.
Deming, W. E. (1994). The New Economics. 2nd ed. Cambridge, MA: MIT
Press.
Dennett, D. (1995). Darwin’s Dangerous Idea: Evolution and the Meanings of Life. New York: Simon & Schuster.
Dixon-Woods, M., Amalberti, R., Goodman, S., Bergman, B., Glasziou, P.
(2011). Problems and promises of innovation in healthcare: Why
healthcare needs to rethink its love/hate relationship with the new.
BMJ Quality and Safety, 20(Suppl. 1):i47–i51.
Dixon-Woods, M., Baker, R., Charles, K., et al. (2014). Culture and
behaviour in the English National Health Service: Overview of lessons from a large multimethod study. BMJ Quality and Safety,
23:106–115.
Dixon-Woods, M., Bosk, C. L., Aveling, E. L., Goeschel, C. A., Pronovost,
P. J. (2011). Explaining Michigan: Developing an ex post theory of a
quality improvement program. Milbank Quarterly, 89:167–205.
Docherty, P., Huzzard, T., de Leede, P., Totterdill, P. (2003). Home and
Away. Learning in and Learning from Organizational Networks in
Europe. London: Innoflex.
Docherty, P., Shani, A. B. (2008). Learning mechanisms as means and
ends in collaborative management research. In Shani, A. B., Mohrman, S. A., Pasmore, W. A., Stymne, B., Adler, N., Eds. Handbook
of Collaborative Management Research, 163–181. Thousand Oaks,
CA: Sage.
Donabedian, A. (1992). The Role of Outcomes in Quality Assessment and
Assurance. Quality Review Bulletin (QRB), 18: 356–360.
Donabedian, A. (2000). Evaluating Physician Competence. Bulletin of the
World Health Organization, 78(6): 857–860.
Donabedian, A. (2003). An Introduction to Quality Assurance in Health
Care. Oxford: Oxford University Press.
Ekman, I., et al. (2012). Effects of person-centred care in patients with
chronic heart failure: The PCC-HF study. European Heart Journal,
33(9):1112–1119.
Gibbons, M., Limoges, C., Nowotny, H., Schwartzman, S., Scott, P.,
Trow, M. (1994). The new production of knowledge: The dynamics
of science and research in contemporary societies. London: Sage.
Gustavsson, S. (2013). The customer in focus: Patient roles in designing
and improving care processes. Licentiate dissertation. Chalmers
University of Technology, Gothenburg.
Gustavsson, S. (2014). Improvements in neonatal care using experiencebased co-design. International Journal of Health Care Quality Assurance, 27(5):427–438.
Hacking, I. (1999). The Social Construction of What? New York: Harvard
University Press.
Hackman, J. R., Wageman, R. (1995). Total quality management. Empirical, conceptual, and practical issues. Administrative Science Quarterly, 40:309–342.
Health Foundation. (2014). Perspectives on Context. London: The Health
Foundation.
€m, A. (2007). On the diffusion and adoption of management
Hellstro
ideas. Ph.D. dissertation. Chalmers University of Technology, Gothenburg, Sweden.
€m, A., Lifvergren, S., Quist, J. (2010). Process management in
Hellstro
healthcare—Investigating why it is easier said than done. Journal of
Manufacturing Technology, 21:499–511.
Improvement in Health Care
Iedema, R., Merrick, E., Britton, K., Gray, J., Verma, R., Manning, N.
(2010). Codesigning as a discursive practice in emergency health
services: The architecture of deliberation. Journal of Applied Behavioral Science, 46:73–91.
Institute of Medicine. (1999). To Err Is Human: Building a Safer Health
System. Washington: National Academy Press.
Institute of Medicine. (2001). Crossing the Quality Chasm: A New Health
System for the 21st Century. Washington: National Academy Press.
International Organization for Standardisation (ISO). (2000). ISO
9000:2000 Quality management systems. Fundamentals and
vocabulary. Geneva: International Organization for Standardisation.
Ishikawa, K. (1982). Guide to Quality Control. Tokyo: Asian Productivity
Press.
Ishikawa, K. (1985). What Is Total Quality Control? The Japanese Way.
1st ed. Englewood Cliffs, NJ: Prentice-Hall.
Jablonka, E., Lamb, M. J. (2005). Evolution in Four Dimensions: Genetic,
Epigenetic, Behavioural, and Symbolic Variation in the History of
Life. Cambridge, MA: MIT Press.
James, J. T. (2013). A new, evidence-based estimate of patient harms
associated with hospital care. Journal of Patient Safety, 9(3):127–
142.
Juran, J. M. (1986). The quality trilogy, a universal approach to managing
for quality. Quality Progress, 19(8):18–24.
Kahneman, D. (2011). Thinking, Fast and Slow. New York: Farrar, Strauss
& Giroux.
Kuhn, T. S. (1962). The Structure of Scientific Revolutions. Chicago: University of Chicago Press.
Langley, G. J., Nolan, K. M., Norman, T. W., Provost, L. P. (1996). The
Improvement Guide. San Francisco: Jossey Bass.
Lewin, K. (1948). Action research and minority problems. In Lewin,
G. E., Ed. Resolving Social Conflicts & Field Theory in Social Science, Washington, DC: American Psychological Association,
201–216.
Lewis, C. I. (1929). Mind and the World Order. New York: Dover
Publications.
Lewis, C. I. (1934). Experience and meaning. The Philosophical Review,
43(2):125–146.
Lifvergren, S. (2013). Quality improvement in healthcare experiences from two longitudinal case studies using an action
research approach. Ph.D. dissertation. Centre for Healthcare
Improvement, Chalmers University of Technology, Gothenburg, Sweden.
€m, A. (2012). The path to
Lifvergren, S., Andin, U., Huzzard, T., Hellstro
sustainability in healthcare—Exploring the role of learning microsystems. In Mohrman, S., Shani, A. B., Eds. Organizing for Sustainable Healthcare. Vol. 2. Bingley, UK: Emerald, 169–197.
Lifvergren, S., Chakhunashvili, A., Bergman, B., Docherty, P. (2008).
Online statistical monitoring of critical patient data increases
patient safety. Proceedings of the PICMET-Conference, Capetown,
South Africa, July 27–31.
€m, A. (2010). Management by diaLifvergren, S., Docherty, P., Hellstro
logue: Joint reflection, sense making and development. Proceedings of A Time for Change: Restructuring America’s Health Care
Delivery System—Cornell University’s International Health Care
Conference, May 11–12, New York.
Lifvergren, S., Docherty, P., Shani, A. B. (2011). Toward a sustainable
healthcare system: Transformation through participation. In Mohrman, S., Shani, A. B., Eds. Organizing for Sustainable Effectiveness,
Vol. 1. Bingley, UK: Emerald, 99–125.
€m, A., Chakunashvili, A., Bergman, B.
Lifvergren, S., Gremyr, I., Hellstro
(2010). Lessons from Sweden’s first large-scale implementation of
Six Sigma in healthcare. Operations Management Research, 3:117–
128.
Maher, L., Baxter, H. (2009). Working in partnership with service users.
British Journal of Healthcare Management, 15(4):172–175.
Marshall, M. (2011). Seeing with new eyes: What can general practice
learn from the science of improvement. British Journal of General
Practice, 61(586): 352–355.
33
Downloaded by [Svante Lifvergren] at 09:50 25 December 2014
Marshall, M., Pronovost, P., Dixon Woods, M. (2013). Promotion of
improvement as a science. Lancet, 318:419–421.
Mauleon, C., Bergman, B. (2009). Exploring the epistemological origins
of Shewhart’s and Deming’s theory of quality. Influences from
Lewis conceptualistic pragmatism. International Journal of Quality
and Service Sciences, 1(2):160–171.
Mezirow & Associates (2000). Learning as Transformation, Critical Perspectives on a Theory in Progress. San Francisco, CA: Jossey-Bass.
Moore, C., Buchanan, D. A. (2013). Sweat the small stuff: A case study of
small-scale change processes and consequences in acute care.
Health Service Management, 26(1):9–17.
Neuhauser, D. (2002). Heroes and martyrs of quality and safety: Ernest
Amory Codman MD. Quality and Safety in Health Care, 11:104–
105.
Neuhauser, D., Provost, L., Bergman, B. (2011). The meaning of variation
to healthcare managers, clinical and health-services researchers,
and individual patients. BMJ Quality and Safety, 20:i36–i40.
Nordgren, L. (2008). The performativity of the service management discourse: “Value creating customers” in health care. Journal of
Health Organization and Management, 22(5):510–528.
Normann, R. (2001). Reframing Business: When the Map Changes the
Landscape. Chichester, UK: Wiley.
Parry, G. J., Mate, K., Perla, R. J., Provost, L. (2013). Promotion of
improvement as a science. Lancet, 381:1901–1902.
Perla, R. J., Provost, L. P., Parry, G. J. (2013). Seven propositions of the science of improvement: Exploring foundations. Quality Management
Health Care, 22(3):170–186.
Petersen, P. B. (1999). Total quality management and the Deming
approach to quality management. Journal of Management History,
5(8):468–488.
Pettigrew, A. (1987). Context and action in the transformation of the
firm. Journal of Management Studies, 24(6): 649–670.
Pettigrew, A. M., Whipp, R. (1991). Managing Change for Competitive
Success. Oxford: Blackwell.
Pink, D. H. (2010). Drive: The Surprising Truth about What Motivates Us.
New York: Riverhead Books.
Ramsay, F. (1926). Truth and probability. In The Foundations of Mathematics and other Logical Essays, Ch. VII, p.156–198, edited by R. B.
Braithwaite, London: Kegan, Paul, Trench, Trubner & Co., New
York: Harcourt, Brace and Company.
Reach, G. (2014). Patient autonomy in chronic caring: Solving a paradox.
Patient Preference and Adherence, 2014(8): 15–24.
Savage, L. J. (1954). The Foundations of Statistics. New York: John Wiley
and Sons.
Senge, P. M. (1990). The Fifth Discipline; the Art and Practice of the
Learning Organization. New York: Doubleday Currency.
Shani, A. B., Docherty, P. (2003). Learning by Design: Building Sustainable Organizations. Oxford: Blackwell.
Shewhart, W. A. (1931). Economic Control of Quality of Manufactured
Product. New York: Van Nostrand.
34
Shewhart, W. A. (1939). Statistical Method from the Viewpoint of Quality
Control. Washington, DC: Graduate School of the Department of
Agriculture.
Shojania, K. G., Grimshaw, J. M. (2004). Still no magic bullets: Pursuing
more rigorous research in quality improvement. American Journal
of Medicine, 116:778–780.
Sjohania, K., et al. (2007). How quickly do systematic reviews go out of
date? A survival analysis. Annals of Internal Medicine, 147:224–
233.
Stacey, R. D. (2003). Strategic Management and Organizational Dynamics. The Challenge of Complexity. 4th ed. Harlow, UK: Prentice Hall.
Swedish Association of Local Authorities and Regions (SALAR). (2008).
Swedish healthcare in international comparison (in Swedish.
http://webbutik.skl.se/bilder/artiklar/pdf/7164-363-6.pdf (accessed
November 11, 2014).
Swedish National Board of Health and Welfare. (2006). God Vard—Om
€r kvalitet och patients€akerhet I h€also- och sjukvarledningssystem fo
den [Good care—On management system for quality and safety in
healthcare]. Stockholm: Swedish National Board of Health and
Welfare.
Sztompka, P. (1991). Society in Action: The Theory of Social Becoming.
Chicago: University of Chicago Press.
Tosey, P., Visser, M., Saunders, M. N. K. (2012). The origins and conceptualizations of “triple-loop” learning: A critical review. Management Learning, 43(3):291–307.
Visser, M. (2003). Gregory Bateson on deutero-learning and double bind:
A brief conceptual history. Journal of History of the Behavioral Sciences, 39(3):269–278.
Waldrop, M. M. (1992). Complexity, The Emerging Science at the Edge of
Order and Chaos. New York: Simon and Schuster.
Walker, J. H., Armenakis, A. A., Bernerth, J. B. (2007). Factors influencing
organizational change efforts: An integrative investigation of
change content, context, process and individual differences. Journal of Organizational Change Management, 20(6):761–773.
Weick, K. E., Sutcliffe, K. M. (2001). Managing the Unexpected: Assuring
High Performance in an Age of Complexity. University of Michigan
Pressing Problem Series. San Francisco: Jossey-Bass.
Weisz, J. A., Mohrman, S. A., McCracken, A. (2013). It’s A Great Time to
Be a Physician: Building a Healthcare System That Works. Bozeman,
MT: Second River Healthcare.
Wilcox, M. (2004). Prediction and pragmatism in Shewhart’s theory of
statistical control. Journal of Management History, 42(1):152–165.
Wilmott, H. (1997). Management and organization studies as science?
Organization, 4:3309–3344.
Woodall, W. H. (2006). The use of control charts in health-care and public-health surveillance, with discussion. Journal of Quality Technology, 38(2):89–134.
Woodall, W. H., Mongomery, D. C. (2014). Some current directions in
the theory and application of statistical process monitoring. Journal
of Quality Technology, 46(1):78–94.
B. Bergman et al.