Complexity in Primary Care
Item
- Title
-
Complexity in
Primary Care - extracted text
-
If
Complexity in
Primary Care
Understanding its value
lU'iVQ
Radcliffe Publishing Ltd
18 Marcham Road
Abingdon
Oxon OX14 1AA
United Kingdom
72362 ■
www.radcliffe-oxford.com
Electronic catalogue and worldwide online ordering facility.
112
© 2006 Kieran Sweeney
Kieran Sweeney has asserted his right under the Copyright, Designs and Patents
Act 1998 to be identified as author of this work.
All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior permission
of the copyright owner.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library.
ISBN-10 1 85775 724 6
ISBN-13 978 1 85775 724 8
Typeset by Anne Joshua & Associates.
Printed and bound by TJ Internationa
Community Health Cell
Library and Information Centre
# 359, "Srinivasa Nilaya"
Jakkasandra 1st Main,
1st Block, Koramangala,
BANGALORE - 560 034.
Ph :2553 15 18/2552 5372
e-mail: chc@sochara.org
Contents
Foreword
About the author
iv
vi
Introduction
1
1 Why bother? The need to understand explanatory models
3
2 The biomedical tradition: why doctors think like doctors
15
3 Evidence-based medicine: the contemporary manifestation of the
explanatory model in medicine
27
I
I
4 The naturalistic tradition: historical overview of the epistemological
origins of qualitative research
45
5 The non-linear tradition: historical development of complexity
61
6 Developing an understanding of chaos and complexity: implications
and examples
83
7 Using complexity principles in healthcare research: examples of
data analysis using complexity principles
97
8 Complexity and medical practice: prospects for the future
117
Appendix:
Paper 1: Evidence-based practice: can this help joint working?
Paper 2: A preliminary study of the decision-making process within
general practice
Paper 3: Why general practitioners do not implement evidence:
qualitative study
Paper 4: A comparison of professionals' and patients' understanding of
asthma: evidence of emerging dualities?
125
126
140
References and further reading
147
Index
161
I
133
135
Foreword
This book describes a journey in pursuit of understanding with the author cast in
the role of everyman or, more precisely, every-general-practitioner. It is illumi
nated throughout by his delight in different forms and ways of knowing. He
begins with a single consultation and proceeds via an authoritative analysis of the
claims of contemporary evidence-based medicine to revel first in the naturalistic
intellectual tradition and then, with palpably mounting excitement, the new
insights which are generated by an understanding of chaos and complexity
theory. He shows that each way of knowing has the capacity to inform and
enrich the others and that none can lay claim to any sort of exclusive truth.
In 1998, I had the privilege of accompanying Kieran on a very small part of his
journey. By chance, we both attended an extraordinary conference in Durham
with the slightly less than seductive title of 'Advancing methodology in general
practice research'. It had been organised by Frances Griffiths on behalf of NoReN,
the Northern Primary Care Research Network, and it included a revelatory
presentation by the sociologist David Byrne outlining the rudiments of chaos
and complexity theory. Kieran and I found that this resonated so powerfully with
our intuitive and experiential knowledge of general practice that, I think, neither
of us ever felt quite the same again about the work that we do everyday. It
suddenly made sense of why guidelines are of such limited usefulness in the
reality of daily practice, why the same treatment applied to apparently similar
people carrying the same biomedical diagnosis can have such very different
outcomes and why the 'rolling out' of pilot initiatives is almost always disappoint
ing. It opened up whole new areas of understanding and Kieran went on to use
complexity theory as the intellectual basis of his future research. This book is a
large part of the result.
More than 30 years ago, the American teacher and literary critic, Lionel
Trilling, wrote about the distinction between sincerity and authenticity. Sincerity
is the notion of being true to oneself through the achievement of a consistency
between thinking, feeling and doing, whereas authenticity poses an even greater
challenge because it recognises the existence of many, and potentially conflicting,
selves which must be acknowledged and accommodated. In his introduction to
Ulysses, Declan Kibberd sees the success of James Joyce as the realisation of
authenticity. In the same way, Kieran Sweeney seeks, perhaps, to move beyond
the traditional sincerity of John Berger's Fortunate Man and to explore the
possibility of authenticity within contemporary general practice.
In Ulysses, James Joyce showed that the detail and complexity of a single day in
the life of a single individual contains within it the breadth, depth and extent of
the totality of human experience. Similarly, Kieran Sweeney finds all the
fascination and challenge of general practice within a single consultation.
Declan Kibberd writes:
■
Foreword
v
Man's littleness is seen, finally, to be the inevitable condition of his
greatness. What one man does in a single day is infinitesimal, but it is
nonetheless infinitely important that he do it.
in
niHe
he
.tic
ity
nd
lis
nn
ral
N,
ry
os
th
er
It
le
ar
nt
tse
a
el
-y
-y
?r
o
d
e
n
•f
e
i.
What happens between a patient and a general practitioner within a single
consultation is also infinitesimal but nonetheless infinitely important. The
tragedy is how poorly this is understood by those in power.
Iona Heath
April 2006
About the author
b<
Kieran Sweeney holds arts and medical degrees from Glasgow University, a
research Masters degree from Exeter University and a doctorate in medicine from
the Peninsula Medical School, Universities of Plymouth and Exeter. He has been
awarded honorary fellowships of the Royal College of General Practitioners and
the Royal Society of Arts, and the Eric Elder Medal by the Royal New Zealand
College of General Practitioners. After completing his medical undergraduate
training, he completed an extended general practice training programme in the
south-west of England, Paris and Brittany. He works as a general practitioner in
Exeter (in the sixth doctor-nurse partnership in the NHS), as an Honorary Senior
Lecturer at the Peninsula Medical School, and as Director of Research Manage
ment and Governance for the south-west peninsula. He acts as spokesperson for
the NHS Alliance on Policy and Leadership.
Between 2000 and 2004 he worked for the Commission for Health Improve
ment, as part of the development team which devised the assessment framework
for primary care organisations. He co-managed the first such assessment in
Norfolk. Between 2003 and 2004 he was seconded to the Transition Team,
which was established in order to set up the Commission for Healthcare Audit
and Inspection (now the Healthcare Commission). After its establishment, he
worked briefly as Policy Manager for Primary Care at the Healthcare Commission.
He resigned from the Commission to start up a completely new general practice in
Exeter in the autumn of 2004.
Kieran is a former Harkness Fellow and Visiting Scholar at the University of
Washington, and during the mid-1990s he was a contributing health corres
pondent for The Times in London. He is a member of the Health Complexity
Group, a collection of healthcare researchers, practitioners and philosophers who
are examining how insights from the complexity sciences might help to explain
the process of change in healthcare organisations. He has published around 100
articles, an Occasional Paper for the Royal College of General Practitioners, and
seven chapters in various textbooks. He has produced three books. The Human
Effect in Medicine: Theory, Research and Practice and A Practical Guide to PCGs and
Trusts were both produced in collaboration with GP colleague Mike Dixon.
Complexity and Healthcare: an Introduction, which he co-edited with Dr Frances
Griffiths, was published by Radcliffe Medical Press in 2002.
Kieran recently received a personal commendation from the Honourable
Annette King, MP, Minister of Health for the Government of New Zealand, for
his contributions to the annual conference of the Royal New Zealand College of
General Practitioners.
He is married to Barbara, a hospital business manager, and has four children.
Introduction
General practitioners are generalists. Most general practitioners understand that
claim in the technical sense -- they are unique among healthcare professionals in
having diagnostic and management skills that transcend the arbitrary boundaries
of specialisms. Compared with the breadth of clinical problems that general
practitioners encounter, specialists are clinical partialists, addressing the parts,
and only those parts, which their specialism can reach. The diabetic retina, the
ischaemic foot, the menorrhagic uterus and the arthritic hallux are all meat and
bone to the general practitioner's morning surgeries. Technical generalism is the
bedrock of competent general practice.
But the generalism that primary care doctors encounter and explore is much
wider than this. They have to understand the demands of contextual generalism,
where the dynamic of a consultation undergoes a sea change, usually oscillating
from the biomedical to the biographical, and sometimes - dizzyingly - back again.
I start Chapter 1 by describing a consultation where precisely this happened,
reminding us of the need to regard the biographical narrative as dignified,
legitimate and grave, with all its attendant frailties, contradictions and post-hoc
rationalisations. For the generalist, understanding context doesn't just mean
knowing where the patient lives and how long ago she was divorced. It means
accepting that the flow of a consultation may be determined by these very facts,
in a way that prejudices and confronts the arithmetical acrobatics of biomedicine.
Yet we can take generalism a step further still. Implied by this second
dimension of generalism - contextual generalism - is a fundamental shift in
the type of knowledge upon which the consultation comes to be predicated.
When we shift from biomedical to biographical perspective, we have implicitly
accepted a shift in evidentiary framework, from scientific evidence to narrative
evidence. This is the basis for a third dimension of generalism — evidentiary
generalism. This term implies that, in practice, generalists have to feel comfor
table swapping paradigms, deploying a range of world-views - the scientific, the
narrative - in order to make sense of the reckless non-sense and wilful
destruction of debilitating illness.
The purpose of this book is to reflect on the challenges concealed in this brief
apologia for generalism. The book simply asks some questions about the questions
that doctors ask. In Chapter 1, it asks why we need to bother with this exercise
anyway, exploring what an explanatory model is, and why the conventionally
accepted and hegemonic explanatory model in medicine should be scrutinised in
the first place. And, as seems only right and proper for a book on general practice,
it starts with the report of a consultation.
If Chapter 1 takes the first step of justifying why we need to ask questions about
the questions that doctors ask. Chapter 2 leads us down the first pathway, tracing
the intellectual origins of the scientific basis of clinical medicine. 'Why do doctors
think the way they do?' is the question that is asked in this chapter. Chapter 3
takes us deeper into a critique of the scientific model in medicine, by interrogating
•r
2
Complexity in primary care
its contemporary manifestation in the shape of evidence-based medicine. The aim
is not to dispute, or to render disreputable, the notion of evidence-based practice.
This would be foolish, and the chapter carefully presents the benefits to patients
of this systematic approach to clinical evidence. Rather, the aim is to help us to
think more about the way doctors think, and to introduce the notion that we
'think' in different ways. We 'know' things in different ways, too. This is what
Chapter 4 deals with, tracing as it does the origins of the naturalistic enquiry,
which has borne among its fruit the principles of qualitative research. The
argument at this stage is as follows. We know things in medicine mainly from
the perspective of medical science. The intellectual tradition of science in
medicine is glorious, its contribution is immeasurably beneficial and its progress
is magnificent. Yet it remains but one way of knowing - an explanatory model
predicated on an epistemological framework of the scientific experiment, which
in turn reflects a positivist ontology. Another way of knowing, which derives
from the naturalistic tradition, draws on another type of evidence, often (but not
exclusively) narrative evidence, and this epistemology reflects a more socially
constructed ontology.
Chapter 5 introduces us to a third intellectual tradition, a third way of knowing.
Reflecting on the fact that, in the light of advances, principally in mathematics
and biology, many of the 'hard' sciences revisited and modified their explanatory
models, this chapter traces the history of chaos and complexity. The description of
its principles continues in Chapter 6, where examples of the application of
complexity in commerce and economics as well as in clinical medicine are set
out. Chapter 7 presents some practical examples of a methodology which draws
on the principles of complexity to show how qualitative data can be re-explored,
at a second-level analysis, if you like. The final chapter speculates on what all this
can mean for the general practitioner. Is complexity the answer to life, the
universe and everything? Read on to find out.
aim
ctice.
ients
us to
it we
what
luiry.
The
from
ze in
>gress
ciodel
vhich
drives
it not
daily
wing,
natics
latory
ion of
on of
re set
draws
lored,
ill this
the
Chapter I
p-•.WaiRtwpum*
Why bother? The need to understand
explanatory models
A single consultation started the train of thought which has led to this book.
Some years ago, our practice nurse asked me to see Mrs B, an 85-year-old
widow who as I recall, at the time of consultation, had been registered as a patient
with me for about 15 years. I knew her well. Her husband, a pleasant chap who
had been a builder, had died 5 years previously. Mrs B was pretty much estranged
from her two grown-up sons, who were recurrent petty criminals, both serving
prison sentences at the time of the consultation. Box 1.1 shows the conditions
from which Mrs B suffered, and Box 1.2 shows her test results, which the nurse
wanted me to review with her.
Box 1.1 Mrs B's comorbidity
Diabetes
Hypertension
Osteoarthritis
Macular degeneration
Hallux valgus
Box 1.2 Mrs B's test results
Glycosylated haemoglobin
Blood pressure
Total cholesterol
Body mass index
9.7%
180/96 mmHg
8.0 mmol/1
29 kg/m2
Mrs B is not unusual, and my guess is that many people reading this book will
know patients like her. When we met, at the practice nurse's request, I rehearsed
the abundant evidence supporting interventions to lower her blood pressure, to
improve the control of her diabetes and to reduce her lipid levels. I remember
even thinking where the reference for this all lay (with a resume in Clinical
Evidence). I confess to feeling just a shade confident as I explained the abnorm
alities and how we could 'help' to reduce her risk. After a few moments I stopped
- resting my case, as a barrister might say.
Mrs B remained silent for a moment or two. Then she said, 'Well, Jack's dead
and the boys have gone.'
r
4
Complexity in primary care
This has remained one of the most privileged communications I have ever
received. As she delivered her words I sensed that she was saying something very
profound, although its full implications eluded me for a number of years.
Certainly on the day, the consultation changed tack and, looking back, we
muddled through with a compromise strategy, and agreed to review the situation
later. As Mrs B left, I sensed that the balance of influence in the consultation had
rested firmly with her.
I
I
r
t
t
i
t
I"
Analysing the consultation: ‘Jack’s dead, and the
boys have gone’
s
This is really the pivotal sentence, out of which many of the concerns explored in
this book arose. At the simplest level, one can say that the consultation, at the
point when Mrs B made this contribution, moved from being doctor centred to
being patient centred. It moved, one could say, from the biomedical domain to
the biographical domain, or from clinical, evidence-based medicine to a consulta
tion predicated on narrative-based evidence. But the shift was profound. When
the consultation moved from its biomedical phase, it shed its parameters of
F-values, absolute risk and numbers needed to treat. These were replaced by the
parameters of the biographical phase of the consultation - led by Mrs B. Here
despair, hopelessness, regret, guilt perhaps, and defeat were the parameters.
Physical parameters had been replaced by metaphysical ones - two intellectual
worlds seemed to have collided.
It is clear that, when Mrs B offered her contribution, the consultation took off
in another direction. Up until that point, a fairly straightforward consultation was
proceeding, drawing on scientific evidence gleaned from good clinical trials, many
of them randomised and controlled, in the great tradition of scientific medicine.
The remainder of the consultation, led by Mrs B, had nothing to do with that way
of thinking, and arose from her lived experience. Yet in that context Mrs B's
narrative evidence had more impact on the outcome of the interaction between
Mrs B and myself than the clinical evidence-based observations with which I led
the consultation. There were, one could argue, two ways of explaining things
which were competing for influence - two explanatory models which at first sight
did not seem to overlap much. At a deeper level, there were two types of
knowledge jostling for influence. Two different ways of viewing and making
sense of the world were at stake. But what constituted these three levels of
understanding? This is what this book tries to explore.
F
t
t
i
s
t
s
I;
r
F
a
o
t
if
e
c
I
u
v>
n
(]
Explanatory models, types of knowledge and world-views
Why bother? Why interrogate the explanatory model in medicine? An explana
tory model provides a framework from which one can explore the receptive
context within which professionals and patients conduct their conversations
during consultations. Its propositions create boundaries within which these
conversations can take place and also, in so doing, create constraints. For
example, the postulates of homeopathy do not conventionally feature in these
conversations, because they are not supported by the paradigm within which the
current explanatory model in medicine is located. However, to describe
P
tl
ii
P
/
is
p
r
The need to understand explanatory models
5
medicine's explanatory model adequately, one needs to consider the world-view
upon which it is based, and the type of knowledge constructed to populate and
make sense of that world-view. Let me try to clarify succinctly the proposition
that I want to explore.
The nature of an explanatory model, I argue, betrays a predilection for a certain
type of knowledge - the collection of 'facts' which populate one's explanatory
model. Medicine's conventional explanatory model is based on the scientific
tradition. It populates that model with 'facts' arising from that tradition, in the
shape of the results of scientific experiments, among which, for clinical medicine,
the randomised controlled trial stands at the pinnacle. That preference, for one
type of model over another, expresses the world-view that underpins the ex
planatory model. For medicine, I propose that the basis of the world-view
underpinning that model is scientific positivism.
Before going any further, let me clarify some terminology and some initial
standpoints.
I am using the term 'explanatory model' in a pretty straightforward, dic
tionary-based way. Thus I am taking 'model' to mean a set of postulates which
serves to represent an entity that cannot be observed, and 'explanation' (simi
larly derived) to indicate the process of arriving at a mutual understanding or
reconciliation (Kirkpatrick, 1994). In the context of this book, then, an ex
planatory model in medicine consists of a series of sense-making postulates that
are located within the contemporary medical paradigm (using that term in its
original Kuhnsian sense) (Kuhn, 1970). I am arguing from the viewpoint that
the contemporary explanatory model in medicine is dominated by science - that
is, science occupies a hegemonic role in that model. This is what I want to
explore and reconsider.
In this book I will use the term 'science' as Chalmers (1982) does, in what he
calls the 'widely held common-sense view of science.' Chalmers states:
Scientific knowledge is proven knowledge. Scientific theories are derived in some
rigorous way from the facts of experience acquired by observation and
experiment. Science is based upon what we can hear, see and touch. Science
is objective. Scientific knowledge is reliable knowledge because it is objectively
proven knowledge.
This is similar to the way in which another philosopher of science, James Brown,
uses the term 'normal science.' 'Most science,' Brown says, 'is normal science. It is
what is done by all scientists who agree on the basics, that is what the world is
made of, how things interact with it: normal science is a puzzle-solving activity'
(Brown, 2001). Scientists, Brown goes on to say, work within a particular
paradigm, or an accepted set of beliefs within a background of unquestioned
theory. The scientific paradigm (Kuhn, 1970) involves some associative practices,
including a basic agreement about ontology (what the world is made of) and
epistemology (the nature of knowledge, its possible scope and general basis).
I can now put the central proposition of this book in slightly more formal,
philosophical language. An explanatory model, one can say, is an expression of a
particular epistemological standpoint, and this in turn reflects the ontological
view held by the person who is supporting the model itself. An explanatory model
is the product, if you like, of the interaction between the ontological view and the
epistemological framework.
»
6
I'
Complexity in primary care
Science, in the sense that I have defined it above, accepts an objective singular
rational reality that can be measured and verified. This associates science with
positivism, first described by Comte (Honderich, 1995), by virtue of its acceptance
of empiricism (all knowledge is based on sensory experience) and verificationism
(if a statement is to be meaningful it must be empirically testable) (Brown, 2001).
It also shares with positivism a sense of inexorability - of an inevitable progress
inherent within scientific pursuit. And it shares with positivism a notion of
hierarchy of knowledge, with knowledge derived from physics having the highest
value. Scientific positivism is central to the ontological view held, usually more
tacitly than explicitly, by those who support medicine's conventional explanatory
model.
Thus defined, two characteristics are fundamental to science, namely linearity
and reductionism. Linearity assumes a regular, proportionate and stable relation
ship between effect and antecedent cause. Reductionism refers to an approach to
understanding phenomena by reducing the whole to its constituent parts, and
assuming that the whole is the sum of its constituent parts. It is really important to
hold these attributes of the scientific approach in mind, as I shall challenge
medicine's explanatory model precisely at this level, by exploring another ex
planatory model (based on complexity) which emphasises the importance of
interaction between parts, rather than reduction to parts.
In addition to the strict definitional issues, which it is important to get clear at the
outset, it is also important to grasp the implications of this description of medicine's
current explanatory model. By implication, 'science' as understood in this way is
regarded as a purer form of understanding, where fact is considered more import
ant than value and where explanation is confined to the expression of measurable
and verifiable correlations between phenomena (Ruse, 1995). This is justified, I
argue, because of the way in which the conception of science is embodied by the
latest expression of the explanatory model in medicine, namely evidence-based
medicine. Here evidence amounts to knowledge distilled from observation and
experiment. These observations sit in a well-recognised hierarchy, with rando
mised controlled trials at the top, followed by partially or uncontrolled trials, with
expert opinion very much at the lowest position in the league. The whole approach
to evidence-based medicine, with its five steps - from defining the patient's
problem to auditing one's performance in solving it - represents the puzzle-solving
approach of normal science. And it makes clear assumptions, albeit tacitly, about
the world and the practitioners who make sense of it, using the epistemological
principles laid down by its most revered exponent, David Sackett.
In Clinical Epidemiology: a Basic Science for Clinical Medicine, Sackett and colleagues
state 'the assumption is that medicine is rational and so are you' (Sackett et al.,
1985). Evidence-based medicine depicts a world that is rational and objective, and
which can be measured empirically. In addition, it depicts a world in which
experiments can be performed in closed systems where it is assumed that the
researcher can stand outside the system, apart from it, manipulating one or two
key variables in order to measure the outcome precisely. If in the course of this
book this model is criticised, it is not to diminish the extent to which develop
ments in science, which lies at the model's heart, have benefited mankind. Rather
it is to place the role of science in the contemporary explanatory model in a wider
context - a context in which, I shall argue, several different explanatory models
are deployed, exchanged and accepted.
The need to understand explanatory models
7
Now this use of the term 'science' is open to criticism. As it stands, it is strongly
associated with the principle of induction, the process of generalising from a series
of particulars. That argument is not logically justifiable (Chalmers, 1982; Honderich, 1995; Brown, 2001), nor are the issues of authority of scientific knowledge or
Popper's falsificationism (Popper, 1963) accounted for. Nor, finally, is the issue of
all observations being theory driven given proper accommodation (Polanyi,
1958). But the use of the term 'science' in this way is justified pragmatically. It
accords with Chalmer's common-sense view of science (Chalmers, 1982) and
Brown's definition of normal science (Brown, 2001), which are in themselves
compatible with the way in which science is conceptualised in the current
explanatory model in medicine (Sackett et al., 1985).
Facing up to the evidence: chinks in the armour of the
gold standard
Although 'O'-level Latin was still a prerequisite for entry to my medical school in
Glasgow in the early 1970s, the training there was resolutely scientific. Bio
medical science was riding the crest of an intellectual wave (from which, as we
shall see, it was to fall), doctors were held in indisputably high regard, and
advances in technology, especially in the field of molecular sciences, held out the
promise of dramatic new interventions which would conquer the most common
fatal diseases affecting Western societies. It would be nearly two decades before
the breathtaking manifesto of the Evidence-Based Medicine Working Group
(1992) exhorted us to concentrate on assessing precisely what kind of (scientific)
evidence we had available to address our clinical questions. To have questioned
the medical model at that point would have appeared heretical (and stupid).
However, some tried. Ivan Illich introduced the notion of iatrogenesis, the
process by which medical care itself could cause illness (Illich, 1975). Cartwright
and Anderson (1981) published their seminal study of general practice patients
not long afterwards, giving us the first hint that perhaps they were not so
unthinking and passive as doctors assumed them to be. However, there was no
real attempt to disarticulate the medical model, and the hegemony of science
remained virtually impregnable up until the turn of the century.
I confess, along (I imagine) with many other healthcare professionals, to
holding a fairly simplistic view of evidence during that time. The whole thing
was rather a mystery really, but one took comfort from the fact that a randomised
controlled trial was the best one could get, and one could pretty well bet one's life
on the truth of their outcomes. The fact that many patients did, only to
experience an unsatisfactory outcome, was one of the trends that led some
practitioners to reappraise the nature of such trials, and to examine just how
they were set up. The practice to which I belonged in the last two decades of the
twentieth century did precisely that in relation to what appeared, to the
untutored eye, to be a powerful body of evidence supporting the anticoagulation
of patients who were suffering from atrial fibrillation. Around the time when this
evidence was published, I had responsibility in my general practice partnership
for clinical policies in this area. The publication of this evidence, authoritative
reviews of it and policy documents based upon it had not been incorporated into
our clinical practice. Although this could be seen as a simple oversight, excused by
8
Complexity in primary care
the bustle of daily general practice, my initial reading of the evidence led me first
to articulate a more serious reluctance to incorporate the policy wholesale, and
secondly to carry out a more in-depth review of the primary evidence. For me,
the conclusion of this critique was the first step on the road to rethinking the
whole model (Sweeney et al.r 1995). For this reason, the key points of the review
are set out in the next section.
Warfarin and atrial fibrillation: a commentary from
general practice
-I”
In 1992 the NHS Management Executive published a document which for the
first time linked advice about commissioning services at the level of regional
health authorities to contemporary clinical evidence (NHS Management Execu
tive, 1992). I found one piece of guidance striking. In this particular section, the
document advised those commissioning health services to link the provision of
services for people with coronary heart disease to new evidence about the
treatment of atrial fibrillation with the anticoagulant warfarin. My sense of
failure upon reading this document was due to an instant recognition that this
was not a policy that I was implementing in my own daily practice, nor was it a
clinical policy in our practice as a whole. The sense of failure was magnified by
our status as a training practice, where junior doctors passed through from time to
time, relying on us - their trainers - to advise them about new developments.
More importantly, we acted as an example of contemporary and exemplary
general practice for them. Yet here was a policy that had already moved from the
position of research reporting, through editorial comment in medical journals to
the position of health policy in an executive document. All four partners in the
practice, as it was then, held academic appointments. How could we have missed
this new body of knowledge and consequently failed to act upon it? The solution
seemed to be simple. Ascertain the evidence, build a clinical policy for the practice
around that, audit our actions using the well-developed computer system that
was already in use, and prepare to demonstrate the rapid change to our
incumbent registrar.
Reviewing the evidence from the randomised controlled trials
In the early 1990s, six randomised controlled trials in Europe and North America
produced results that supported the use of warfarin in both primary and
secondary prevention of stroke in patients with non-rheumatic atrial fibrillation
(Petersen et al., 1989; Boston Area Anticoagulation Trial for Atrial Fibrillation
Investigators, 1990; Connolly et al., 1991; Stroke Prevention in Atrial Fibrillation
Investigators, 1991; Ezekowitz et al., 1992; European Atrial Fibrillation Study
Group, 1993). Three trials also reported on the beneficial effect of aspirin
compared with placebo. The design and results of these studies are summarised
in Table 1.1 at the end of this chapter. The authors of the six primary prevention
trials reported the findings of a collaborative meta-analysis of their results. The
estimates of the reduction in relative risk of stroke with warfarin are shown for
each trial separately in the right-hand column. Overall, these trials suggested that
warfarin decreased the relative risk of stroke by 68%. The meta-analysis resolved
other questions which had not been clearly answered by the individual trials.
The need to understand explanatory models
9
Warfarin reduced the risk of both major and minor stroke, and was shown to be
equally effective in men and women. The overall effect of aspirin was statistically
significant but smaller. When data from both studies were combined, aspirin
decreased the risk of stroke by 36%. Table 1.2 (at the end of this chapter) presents
the results of these trials.
Within all of these trials the rate of serious complications from warfarin was
remarkably low. In the meta-analysis the annual rate of cerebral haemorrhage
was 0.3% in patients who had been treated with warfarin, and 0.1% in the
control group. Taking these studies together, 40 patients with atrial fibrillation
would have to be given an anticoagulant treatment for one year in order to
prevent one stroke. Out of 1000 patients treated for one year, between 15 and 50
episodes of ischaemic stroke or systemic embolism would be avoided at a cost of
between four and six measured episodes of bleeding over the same period.
Reviewing this evidence for general practice
On the face of it, the evidence from these trials suggested that a substantial benefit
would be gained from using warfarin. Despite this, I together with my partners
had been laggardly in applying this evidence. Early reading of one or two of the
original studies provoked a series of questions about how reproducible these
results might be in routine general practice in the UK. From this reading emerged
the structure of a wider critique of the design and execution of this sextet of
studies. In this critique, the salient questions about these studies concerned the
participants in the trial, and the feasibility of reproducing the results in routine
clinical practice.
• Were the characteristics of the populations that were studied comparable with
the general population who may be offered this form of anticoagulation in
primary care in the UK?
• Is the type of follow-up that was employed in these studies to ensure
compliance feasible in day-to-day general practice?
• Is it possible to stratify risk and thus to individualise therapy in general
practice?
The study populations in the atrial fibrillation trials
I
If the evidence from these trials was to be part of routine care for patients with
this condition, we thought it was important to know the entry and exclusion
criteria. In general these trials involved older patients, of mean age 69 years,
about half of whom had hypertension and about a quarter of whom had angina.
Around 20% had a history of heart failure and 14% were known to have
diabetes. However, there was no standardisation between the trials with regard
to the exclusion criteria. The rate at which patients were excluded from some of
the trials surprised us, and made us reconsider the potential generalisability of the
evidence. For example, in the SPINAF trial (Ezekowitz et al., 1992) 93% of the
eligible patients were excluded, of whom one-third had 'chronic alcoholism or a
psychiatric or social condition rendering the patient unsuitable for anticoagula
tion.' A further 1600 patients were deemed ineligible for inclusion in the study,
according to 'undefined administrative criteria.'
In the SPAF study (Stroke Prevention in Atrial Fibrillation Investigators, 1991)
I0
Complexity in primary care
only 3% of a potential 18 000 eligible patients were entered into the warfarin arm
of the study. Nearly 1000 patients in this study were excluded because the
investigators could not be sure that they were followed up, and about 1700
patients refused to enter the trial once invited to do so. A separate list of exclusion
criteria was applied to over 700 patients in the SPAF study, who were entered into
the programme but not assigned to anticoagulant therapy. Of this group, onethird refused anticoagulant therapy and 6% were excluded because of 'repeated
falls or unstable gait predisposing to head trauma.'
It began to look to us as if the patients who were entered into these trials
represented a population with atrial fibrillation that was, to begin with, at low risk
of bleeding on warfarin. They also seemed to constitute the population most
likely to comply with the treatment and be amenable to follow-up. We checked
the drop-out rates.
Despite the cautious entry criteria, quite large percentages of patients in all of
the trials were withdrawn from warfarin therapy after entering the programmes.
These included 38% of patients in the AFASAK study (Petersen et al., 1989), 10%
in the BAATAF study (Boston Area Anticoagulation Trial for Atrial Fibrillation
Investigators, 1990), 26% in the CAFA study (Connolly et al., 1991), 11% in the
SPAF study (Stroke Prevention in Atrial Fibrillation Investigators, 1991), 31% in
the SPINAF study (Ezekowitz et al., 1992) and 21% in the EAFT study (European
Atrial Fibrillation Study Group, 1993). The largest number of withdrawals was in
the AFASAK trial, which was the closest to being a community study comparable
with a UK primary care population.
Compliance and monitoring
Could the standard of care that these patients received be reproduced in routine
general practice and achieve similar outcomes?
During the studies, patients were vigorously monitored in a hospital outpatient
setting and underwent repeated physical examinations for the side-effects of
warfarin treatment. The rigour of this follow-up programme did at least introduce
the possibility that the perceived safety of the treatment, and in particular its low
complication rates, was associated with the close medical monitoring of the
patients who were receiving treatment. The concern among general practitioners
in the UK was that such rigorous monitoring of patients was unlikely to be
reproduced in routine clinical practice.
The clearest description of follow-up came from the Copenhagen AFASAK
study, in which each patient had clinical check-ups twice in the first 6 months
and every 6 months thereafter. Complete physical examination was undertaken,
and echocardiography was performed to assess left atrial size. During the second
year, echocardiography was repeated and the researchers obtained confirmatory
evidence of continuing atrial fibrillation. Many of the other studies deployed
similar rigorous clinical monitoring. The relevance of this to the problem of
implementation was the difficulty of reproducing such rigorous monitoring,
which was deemed sensible in the light of the potentially catastrophic side-effects
of warfarin, the most serious of which is intra-cerebral haemorrhage.
Because of the potential side-effects of bleeding on warfarin, blood tests were
performed frequently to ensure that the dose of warfarin was appropriate. In the
trials that constituted the much cited body of evidence to support the use of
The need to understand explanatory models
I I
warfarin in atrial fibrillation, blood tests were performed at monthly intervals,
which probably fitted with the recognised schedule for monitoring in UK general
practice. However, anticoagulant control and subsequent determination of the
appropriate dose of warfarin were difficult, even in the hands of the experts who
were conducting these trials. Table 1.2 at the end of this chapter shows the
percentage of study days on which anticoagulant control fell outside the accept
able range, along with the annual rate of major bleeding episodes and the
percentage of patients who reported minor episodes of bleeding on treatment.
To our surprise, many of the trial patients were under-anticoagulated for a
considerable proportion of time during the studies - nearly half of all the days
on treatment in the CAFA study. Could this, we asked ourselves, be related to the
low levels of serious bleeding that were observed in these trials?
Implications of this evidence for practice
There were conflicting views about the implications of this evidence for routine
clinical care. Academic opinion, in the shape of editorials in well-respected
journals, advised full and rapid implementation of the evidence (Laupacis,
1993; Lowe, 1993). However, routine clinical care seemed to be slow to catch
up (Rassam, 1993).
On the basis of these studies, editorials in peer-reviewed journals encouraged
doctors to consider giving anticoagulant therapy to patients with atrial fibrillation
if there were no contraindications (Laupacis, 1993; Lowe, 1993). Some commen
taries on the studies called for lifelong treatment, despite the fact that the mean
duration of treatment in these trials was about 18 months (Laupacis, 1993). The
results of these trials were brought to the attention of the NHS Management
Executive, whose focus-group research identified anticoagulant treatment for
patients with atrial fibrillation as a key element in purchasing negotiations for the
then regional health authority corporate contract (NHS Management Executive,
1992). This was a crucial development. Here fresh research evidence was being
linked to commissioning strategy at management level, despite a lack of vigorous
reflection on the generalisability of these trials to routine clinical practice.
What did this review say about the nature of evidence?
Two important implications emerged from the publication of this review. This was
the first time an extended appraisal of the components of a set of randomised
controlled trials had been published.1 It raised questions about the generalisability
of scientific evidence, about the selection and exclusion criteria for trials and,
emerging from these two points, concerns about how widely or safely this
evidence could be incorporated into routine clinical care. In this respect it
appeared to raise some questions about the model of evidence-based medicine,
which was becoming an important influence at that time (Evidence-Based
Medicine Working Group, 1992).
For me, it raised deeper questions. What was the basis of the randomised
controlled trial, which had caused it to occupy pride of place in the hierarchy of
scientific knowledge in medical practice? How and why had medicine come to
1 The review paper was cited around 30 times in the next 5 years.
I2
Complexity in primary care
rely on the scientific paradigm, within which the randomised trial was one of the
most powerful tools? These were questions about epistemology, about scientific
positivism, and about the philosophical basis of research methods in general. A
simple clinical query about the widely reported value of anticoagulation for this
common cardiac dysrhythmia was leading us to ask basic questions about the
scientific method.
To answer these questions, it seemed to be necessary to research the historical
development of these ideas within clinical medicine. As we shall see, that
intellectual journey simply led to more questions - about the origins of qualitative
data, as opposed to the quantitative data of randomised controlled trials, and
about the ways in which the other natural sciences were facing up to the massive
paradigm challenges resulting from developments in mathematics and thermo
dynamics. My conclusion was bleak, namely that other sciences had assertively
revisited their explanatory models, and some, including physics, had made
sweeping changes to their understanding of the world. Medicine was yet to
wake up to these challenges.
Table 1.1 Summary of randomised trials of warfarin and aspirin in patients with non-rheumatic atrial fibrillation
Study
(year, country)
Design
Comparison
Setting
Duration
(years)
Target
Person
Annual event rate
years of
follow up Placebo
Warfarin
of
% RRR
AFASAK (1989)
Denmark (Petersen et al.)
Randomised
Double blind
Aspirin/Placebo
Warfarin,
aspirin and
placebo
OPDa
2.0
2.8-4.2
INR
398
4.8
1.4
71
BAATAF
(1990)
USA
(Boston Area Trial)
Randomised
Controlled
Unblinded
Warfarin
versus
aspirin
OPD
2.3
1.2-1.5
PTR
435
2.9
0.4
86
CAFA
(1991)
Canada (Connolly et al.)
Randomised
Double blind
Warfarin
versus
placebo
OPDb
2.5
2.0-3.0
INR
241
3.7
2.1
43
Warfarin
versus
placebo aspirin
versus
placebo
OPD
1.3
1.3-1.8
PTR
245
7.4
2.3
67
SPAF
Randomised
(1991)
Aspirin/Placebo
USA
Double blind
(Stroke prevention in atrial
fibrillation)
SPINAF
(1992)
USA
(Ezekowitz et al.)
EAFT
(1993)
Netherlands
(European Atrial
Fibrillation Study Group)
Randomised
Warfarin
Placebo controlled versus
Double blind
placebo
-
warfarin
I
Q_
c
OPD
1.7
Q.
1.2-1.5
PTR
483
4.3
0.9
79
; o
7
5
x
"O
Randomised
Secondary
prevention trial
Warfarin
versus
aspirin versus
placebo
OPD
2.3
2.5-4.0
INR
517
OPD = outpatient department; INR = iinternational normalised ratio; PTR = prothrombin ratio; RRR=relative risk reduction.
a Echocardiography laboratory; b University centres.
17.0
8.0
53
I' ST
I
g
I B
‘
I
Q.
n>
w
I
Table 1.2 Percentage of study days where anticoagulant control fell outside stated range, annual rate
of major bleeding episodes and percentage of patients with minor episodes
Study
AFASAK
BAATAF
CAFA
SPAF
SPINAF
EAFT
% of days where INR/PTR
Bleeding episodes
Below
lower limit
Above
higher limit
Annual rate
of major (%)
% of patients
with minor
26
9
0.6
8
17
0.5
15
9
1.2
0.4
2.5
1.5
1.3
2.8
A
17.9
16.0
A
24.6
20.9
40
23
29
32
INR = international normalised ratio; PTR = prothrombin ratio; A = Not reported.
h
r: 2
I*
£
i
Chapter 2
The biomedical tradition:
why doctors think like doctors
Introduction
So why do doctors think like doctors? What kind of intellectual tradition has
spawned the progeny of the biomedical model? To find out the answers to these
questions, we need to go back to the origins of this approach in ancient Greece
and trace the development of that type of thinking up to the present day. This
ciapter presents a historical overview of the main influences that shaped and
informed the current accepted medical model. It then looks at the current
manifestation of that model, enshrined in the principles of evidence-based
medicine (EBM) and, by presenting a brief critique of EBM, begins to unpick
the features of a different, complementary way of thinking, whose intellectual
pathway has yielded the principles of qualitative research. The next chapter
explores that pathway in greater detail, defining its intellectual origins, and
reflecting on what that means for the way in which doctors think in consultations.
Although this seems to be a helpful way of exploring how the contemporary
medical model evolved, I accept that a pervasive weakness of historical accounts
is that they are exposed to assumptions about the status of past knowledge. Such
assumptions necessarily entail a degree of bias, through the interpretive prism of
the historian. Thus they run the risk of conferring on past knowledge a degree of
significance that it may not initially have had. So although historical accounts of
medicine may not provide a perfect mirror of past events, they are useful for
making explicit the assumptions that underpin the theory upon which the
current accepted model rests.
The historical overview presented here is necessarily brief and eclectic, focusing
on the principal contributors to our current understanding of the medical model.
This historical overview will show that the twentieth century can be divided into
two periods, the first one ending around the mid-1970s. Up until then, rational
scientific progress in medicine seemed unstoppable, exciting discoveries peppered
the medical landscape, and the status of the medical professions (at least in
hospitals) seemed unassailable. However, by the last quarter of that century the
supremacy of interventionist medicine was being called into question, and
developments in mathematics, biology and computing were introducing the
possibility of another paradigm applicable to the biological sciences, this time
predicated on non-linear modelling rather than on the linear rationality of
scientific positivism. I shall explore this paradigm in Chapter 5.
3
I6
Complexity in primary care
The origins of contemporary medicine: ancient Greece
and the Dark Ages
III-"
Ancient Greece is generally regarded as the home of medicine, whose point of
origin is commonly identified in the writings of the Hippocratic collection (Singer
and Underwood, 1962; Porter, 1987; Greaves, 1996). Plato proposed a distinction
between the healthy soul (in which reason occupied a superior position to
passion) and organic social order, in which rational guardians possessed true
authority (Porter, 1987). Dubos (1960) points out that Hippocrates' writings have
had a biblical influence on the thinking that has underpinned medical practice
throughout recorded history. According to Dubos, Hippocrates stands for rational
concepts based on an objective knowledge of science in general, and of medicine
in particular, liberating it from mystic and demonic influences (Dubos, 1960).
There is an important identification here of the superiority of rationality and
reason as conceived within positivism,2 to the detriment of all that was seen to be
irrational or non-rational.
Greek medicine had three fundamental characteristics that have contributed to
its fundamental status in contemporary medicine, ft had a unified theory of
medicine (referred to as naturalism), it held an ontological view of diseases as
specific and real entities awaiting discovery and classification, and it advocated an
outlook based on empirical observation as a way of progressing knowledge
(Greaves, 1996).
During the Roman period that followed, the second great medical name of
ancient times, Galen, produced his writings. Their main contribution was to
organise and restate systematically what had gone before, rather than to produce
anything really new (Phillips, 1973). The prolonged period after Galen, from ad
200 to ad 1500, is still regarded as a regressive period for mankind in general,
and medicine in particular - a view that is symbolised by the use of the term
'Dark Ages' to describe that era. Singer and Underwood (1962), for example,
present the still widely held view of the Dark Ages of medicine as a 'period of
progressive deterioration of the intellect', criticising medicine up to around ad
1500 mainly on the grounds of its lack of precise observation in anatomy and
pathology.
The origins of contemporary medicine:
Renaissance and Enlightenment
I;
In contrast, the next three centuries, from 1500 to 1800, are projected in an
altogether different way - the Renaissance and Enlightenment - in which medical
progress was reawakened and medical theory and practice advanced. In the
context of the history of medicine, the previous period of the Dark Ages was to be
set aside, as an interruption to the inexorable progress of medical understanding
which could be ignored (Cunningham, 1989). However, the same author
(Cunningham, 1989) suggests that it was Hermann Boerhaave (1668-1738), a
2
I use Brown's definition of positivism as predicated on empiricism (all knowledge is based
on sensory experience) and verificationism (for a statement to be meaningful, it must be
empirically testable) (Brown, 2001).
The biomedical tradition
I7
physician and teacher who lived in Leiden, who had a seminal influence on
portraying (retrospectively) this progressive scientific view of medical history,
which has continued to be widely held up to the present day:
In the early decades of the eighteenth century Boerhaave created a history
whose peaks were Hippocrates, Bacon, Sydenham and Newton. Hippocrates
first practised proper medicine, Bacon pointed out the means for its restoration,
Sydenham effected its restoration into practice, and Newton provided the
means to understand properly the working of the body.
King (1982) has summarised the contributions of Francis Bacon (1561-1626) (to
which we shall return in Chapter 4 on qualitative research) in his demand for
precision in observation, and for repeated experimentation leading to cautious
generalisation. The chief elements of the modern scientific method. King argues,
are found clearly expressed in Bacon's writings. 'He recognised the need for
controls'. King writes, 'pointing out the dangers of hastily drawing conclusions,
the need for verification and the return to particulars once the generalisation had
been made.' Singer and Underwood (1962) argue that this huge influence
ascribed to Bacon must be seen in the light of the achievements of the great
Renaissance pioneer Andreas Vesalius (1514-1564). 'The masterpiece of Vesa
lius , they write, 'is not only the foundation of modern medicine as a science, but
the first great positive achievement of science itself in modern times.'
The historical and intellectual link between Hippocrates, Vesalius and Bacon
and Isaac Newton (1642-1727) now seems obvious to contemporary practi
tioners. Newton developed a physical system of science that claimed to be unified,
absolute and objective. Medicine enthusiastically embraced the Newtonian ideal
of this single schema, and each of its newly produced theories claimed to support
and develop that ideal. In addition, Newton influenced modern medical thinking
by developing the principles of mechanics, heralding a mechanistic model of
science, populating the medical vocabulary with mechanical metaphors, and
inexorably imbuing the medical model with linear thinking. The essence of the
linear method was the acceptance of a proportional, steady, regular and pre
dictable association between variables. Newton's approach was also reductionist,
implying that phenomena could be described and hence understood by reducing
the whole to its constituent parts and, more importantly, by assuming that the
whole was the sum of its constituent parts. These parts were thought to be
regulated by a small number of core laws, and were assumed to change in a
smooth and predictable manner. With the adoption of these mechanical meta
phors, a new discipline of iatrophysics (Greaves, 1996) emerged, which promoted
the study of the body as a machine. Perhaps the most elegant application of
Newtonian mechanics to clinical practice was the description of circulation by
William Harvey (1578-1657) in 1628. Singer and Underwood (1962) strongly
emphasise this important mechanical approach to medical practice and under
standing exemplified by Harvey's description of circulation. 'The knowledge of
the circulation of the blood has been the basis of the whole of modern
physiology', they wrote, 'and with it the whole of modern rational medicine.'
Greaves (1996) sees Harvey as rivalling Vesalius as the founder of modern
scientific medicine. However, both can in part be accused of maintaining a
rather solipsistic view of the rise of medicine in the West, by ignoring the
staggering contribution to this field that was made by the Syrian physician
1
1
I8
Complexity in primary care
Ibn al-Nafis (1200-88), who described how oxygenation of the blood took place
around 300 years earlier than anyone in Europe (Brown, 2001).
The last member of Boerhaave's heroic quartet was Thomas Sydenham (16241689), who practised in London between 1656 and 1689. His contribution to
improving medical practice resulted from systematic observation of patients
independently of any medical theory, and his focus on the typical manifestations
of disease, rather than on a particular individual's unique experience of illness.
Relying heavily on induction, his contribution was to derive general accounts
from a multiplicity of individual case histories. Greaves (1996) argues that
Sydenham had an ontological conception of disease, viewing diseases as real,
distinct and natural entities that were awaiting discovery, ready to be allocated
their pre-ordained position in a natural taxonomy. The conventional view of
history sees Bacon, Harvey, Newton and Sydenham as key figures laying down
the basis for the modem understanding of medicine and medical science.
IK '
The rise of science in medicine: the nineteenth century
The dawning of the nineteenth century marked an important watershed, not only
in world politics (less than a quarter of a century had elapsed since the
Declaration of Independence by the United States of America), but also for the
rise of science in medicine. And it is perhaps fitting that scientific progress in
medicine had its origins in Paris, when not just France but Europe in general was
reeling from the consequences of the French Revolution of 1789-95. One of the
consequences of the Revolution had been the removal of hospitals from the
hands of the Church into those of the nation state. Medical politics, policies and
institutions were vigorously reformed, and new programmes of medical enquiry
and research practice were introduced by an enthusiastic and ambitious commu
nity of physicians based around Paris, where their salaried appointments gave
them access to 20 000 beds in the city alone, outnumbering England's entire
inpatient population at that time (Porter, 1997). The most significant change in
clinical practice was the central role of the autopsy to corroborate bedside
diagnoses. The tumours and infections that killed people became the focus of
attention, and doctors were continually searching for correlates between clinical
presentation and pathological lesion. Porter (1997) points out that early signs of
this shift to a disease focus in medical practice had not exclusively occurred in
Paris. The previous century had seen parallel traditions developing in Scotland
and Germany in particular, and also in England - for example, through the
contribution of Thomas Sydenham, whose work has already been mentioned.
Bichat (1771-1802) wras probably the key influence in the Parisian movement
at this time (Greaves, 1996). He emphasised the importance of anatomical
dissection. 'Start cutting bodies open,' Bichat said, 'and hey presto, this obscurity
will soon disappear' (quoted in Porter, 1997). It was ignorance of anatomy that
had led physicians to neglect internal diseases. Before the rise in influence of
physicians such as Bichat and his contemporary Corvisart (1755-1821), diagnosis
had relied on the patient's history as well as clinical observation. However,
although anatomy and the study of pathological lesions became dominant,
clinical observation was not completely ignored. Rene Laennec (1781-1826)
wrote an important treatise on the stethoscope (published in 1819), which
encouraged physicians to bypass patients' accounts, thereby rendering diagnosis
The biomedical tradition
I9
more 'objective.' Corvisart, Laennec and Bichat revelled in the amount of clinical
material at their disposal in Parisian hospitals. Clinical medicine, Greaves (1996)
argues, 'aimed to be a science, hinging on clinical detachment where empirical
data were acquired through relentless examination of pathological lesions.' What
this analysis implies is the increasing acceptance of reductionism in the medical
explanatory model, expressed in the enthusiasm for anatomy to reveal the
structure of the body, confident that, armed with this mechanical, Newtonian
approach, an understanding of its function would follow.
Working alongside this cohort of clinicians, Pierre Louis (1787-1872) con
tributed significantly to the evolution of clinical science by advocating the use of
numerical methods, using simple arithmetic to test the relative merits of com
peting therapies (Bynum and Porter, 1993). He encouraged clinicians to group
together large batches of patients undergoing different treatments for the same
condition, pointing out how differences in mortality would indicate the appro
priateness of the chosen therapy. In doing this, it is clear that he paved the way
for the clinical trial, whose randomised controlled form would emerge around 60
years later as the gold standard for evidence (Bynum and Porter, 1993).
Under the influence of the French, medical education across Europe and in
America became more scientific and systematic. Foucault (1963) observes that
this was 'the great break in the history of western medicine, dating precisely from
the moment clinical experience became the anatomo-clinical gaze.' In Vienna,
pathology dominated all other emerging medical specialties through the work of
Rokitansky (Bynum and Porter, 1993). In England, Hunter occupied a similar
privileged position to Bichat in Paris. The central position of anatomy in
Edinburgh and London around the mid-point of the nineteenth century was
highlighted by the gruesome activities of Burke and Hare, who first had the idea
of selling dead bodies to anatomists (for £7 each), bypassing the grave (Porter,
1997). Their enthusiasm for the fees that such anatomical specimens could
command led them to suffocate victims in preparation for their sale - murdering
to dissect, as Wordsworth was later to observe.
This anatomical gaze, and the constant search for correlation with clinical
abnormalities, was given a huge boost by refinements in optics, particularly the
advances in microscopy that were introduced by Lister (1786-1869), the father of
the more famous surgeon. This in turn had benefited from the brilliant lens
making capabilities of Carl Zeiss (1816-1888), among others (Bynum and Porter,
1993). It was in Germany that the new sciences of biology and histopathology
were developed. The term 'histology' was coined in 1819, the year in which
Laennec's treatise on the stethoscope was published.
Coming together (more by chance than by design) in Germany around the
middle of the nineteenth century was the powerful triad of clinical ambition,
technological advance, and financial support through the educational reforms of
German rulers who invested heavily in academic science (Greaves, 1996).
Understandings in physical chemistry helped physicians to interpret previously
unfathomable physiological findings. A key figure in this domain was William
Prout (1785-1850), who worked in England and was generally regarded as the
father of biochemistry. Curiously, Prout's approach to clinical chemistry revealed
one of the emerging tensions in the epistemological basis of scientific medicine.
Prout was a vitalise who believed that the chemistry of biological systems was
20
mu
k:.
Complexity in primary care
something quite different from that of any physical system, being supported by
what he called a 'vital force' (Brock, 1985, 1993).
Later in the nineteenth century, the attempt by Karl Ludwig (1816-1895) to
debunk this mystic vitalism in physiological science did more than anything to
root advances in this domain in quantitative positivistic and materialistic science.
'Every illness', Ludwig wrote in his Textbook of Human Physiology (cited in Porter,
1997), 'is a physiological experiment and each physiological experiment is an
artificially produced illness.' By the late nineteenth century, technological
advances had enabled scientists and clinicians to make the cell the focus of
their research. These developments led to the crowning achievement of nine
teenth-century medicine, namely the theory of contagionism and the germ
theory of disease. This infection model. Ten Have (1990) asserts, 'turned out to
be the most powerful paradigm of modern scientific medicine', singularly
responsible for securing the decisive ascendancy of scientific positivism in
medicine. The work of Pasteur and Koch is most closely identified with this
domain, and although Koch's postulates were first published in 1891, they had in
fact been described theoretically at least half a century earlier (Koch, 1891). Koch
proposed that there were three essential elements in the ideal disease model,
namely the causal agent, the pathological lesion and the clinical syndrome. It
was, according to Greaves (1996), the most unambiguous expression of the
dominance of scientific positivism in medicine, and it carried a series of important
consequences and assumptions. Medical knowledge could henceforth only be
determined by doctors, and progress would now only be linked to developments
in science. Health was conceived as the absence of disease and 'illness' an
imperfect account of it. Diseases existed as discrete real entities, with a universal
nosology that was potentially completely discoverable (Wright and Treacher,
1982). Despite some swift modifications to the doctrine of specific causation
(which was expanded to allow for more than one causal element), the rational
and scientific basis of Koch's postulates has had an enduring influence on medical
knowledge, practice and status ever since (Wright and Treacher, 1982).
Almost alone among contemporary commentators, Greaves (1996) identifies a
weakness in the portrayal of Koch's postulates - in the way that they were
portrayed as the triumph of positivism in medicine. Greaves asserts that the
postulates essentially emphasise the primacy of the causal organism, but refer to it
as if it were separate from the disease, in which case the other elements of the
triad could take precedence, or occupy equal status with that of the causal
organism. Virchow, who made a substantial contribution to the medical implica
tions of cell biology by painting a picture of 'the republic of the cells', in which
diseases were firmly situated in 'cellular abnormalities, multiplied through
sequential divisions' (Virchow, 1858, cited in Porter, 1997), identified this
confusion early on (Byron and Boyd, 1991). He described how 'the hopeless
and never-ending confusion, in which ideas of being and causation have been
arbitrarily thrown together, began when micro-organisms were finally discov
ered' (Virchow, 1895). What happened in practice, Greaves (1996) argues, was
that the causal element was given primacy, but in so doing undermined the
notion of objectivity, as a degree of judgement had been introduced into the
concept of disease. In addition, the exercise of this judgement meant that the
postulates in this 'ideal' model of disease were actually being used to construct the
notion of disease, rather than to reveal a pre-existing (or even pre-ordained)
The biomedical tradition
21
notion of disease, which had been the original intention of the model. This
attacked the very notion of value-free scientific knowledge as the source of
medical understanding, but it did not diminish the importance of Koch's
postulates in the progression of medical science in the twentieth century.
However, its continued centrality to the medical model resulted from a complex
interplay of social and political factors. The general optimism about human
progress which characterised the nineteenth century and was captured, for
example, in the writings of Comte and Mill (whom I shall discuss in greater
detail in Chapter 4 on the origins of qualitative research), the desire to
professionalise medicine (for example, through the Medical Act of 1858) and
the statutory provision of healthcare, particularly in the area of public health and
infectious disease (although the latter had been predicated on a quite different
notion of infection called miasmatism; Singer and Underwood, 1962), all
combined to sustain the centrality of the germ theory of disease in medicine
and, more importantly, of the epistemological and ontological propositions on
which that theory rested.
Porter (1997) summarises the progress of clinical medicine during the nine
teenth century as follows: The pathological gaze penetrating the diseased body
and the eye of microscopy formed part of the wider attempts to apply the methods
of science to the whole medical enterprise, including the regular business of
clinical medicine.' Advances in microscopy and optics led directly to the devel
opment of haematology as a medical specialty. Physiology was also supported by
technical advances, and physiologists had achieved unprecedented influence,
arguing that medical science had to understand the normal no less than the
abnormal. The inexorable trend was towards objectifying findings accumulated at
the bedside, and the monitoring of pulse, monitoring of temperature and serial
measurements of chemical functions all became part of what was later to be called
the 'work-up' of the patient. Clinical science itself became a byword for the
advances of medical investigation that characterised the twentieth century. Huge
advances in technology assisted developments in endocrinology and cardiology,
and later assisted the introduction of two new specialties, namely genetics and
immunology (Reiser, 1991; Devor, 1993).
This overview shows that science already occupied a hegemonic position in
medicine by the end of the nineteenth century. A medical model was emerging
which was firmly rooted in a positivist ontology - that is, a world-view that insists
that anything which is measurable is real. Within this medical model, the body
was regarded as a machine, the dominant metaphor was mechanical, and the
relationship of antecedent cause to subsequent effect was regular, proportional
and predictable.
However, the inexorable trend, first described by Boerhaave as an unassailable
progression, does seem to have come together as a result of quite separate
activities. Advances in strict biomedical understanding co-evolved with develop
ments in optics, and both were helped by visionary political investments, such as
the re-integration of hospitals into state control in France, and the huge
investment in educational programmes in Germany in the early decades of the
nineteenth century. And the triumph of rational scientific positivism in medicine,
in the shape of Koch's postulates, is no longer seen to be as objective or
intellectually celibate as was once thought.
y
C7
22
Complexity in primary care
Medicine in the twentieth century: an eclectic overview
Although it is impossible to describe each of the revolutionary innovations in
scientific medicine that occurred during the course of the twentieth century.
Table 2.1 shows a selection of the key achievements.
The application of scientific medicine to many diseases has been so successful
that it is difficult to imagine what life must have been like during the epidemics of
polio, diphtheria and whooping cough that ravaged society during the first half of
the twentieth century.
Table 2.1 Selected milestones in medicine (adapted from Le Fanu, 1999)
H n>
Year
Achievement
1935
1941
1944
1946
1948
1950
1954
1957
1960
1967
1969
1973
1978
1979
1987
1996
1998
2003
Introduction of sulphonamides
Introduction of penicillin
Development of kidney dialysis
Developments in general anaesthesia (curare)
First intra-ocular implant for cataract
Streptomycin cure for tuberculosis
Zeiss operating microscope
Discovery of Factor VIII for haemophilia
Development of the oral contraceptive pill
First heart transplant
Prenatal diagnosis of Down syndrome
Introduction of the CAT scanner
First test-tube baby
Routine coronary angioplasty
Thrombolysis for myocardial infarction
Triple therapy for AIDS
Sildenafil for the treatment of impotence
First birth of a baby screened for genetic and therapeutic compatibility with sibling
Life expectancy has increased in developed countries, and medical interventions
do appear to have contributed substantially to this (Bunker, 1995). There has
been major progress in the field of antibiotic therapy, and in surgery, where
advances in anaesthesia (particularly the introduction of the first heart-lung
bypass machine in 1952 by John Gibbon) supported progress in heart surgery,
leading to the first heart transplant at Groote Schuur Hospital by Christian
Barnard (1922-2002) in 1967. Although it is beyond the scope of this book to
present a systematic account of all the developments in scientific medicine that
have taken place in the twentieth and twenty-first centuries, it would be a major
omission not to mention the publication of the human genome map in 2002,
hailed by the then American President Bill Clinton as the 'greatest discovery of
mankind, something which will revolutionise the diagnosis of all diseases' (Times,
2002). The publication of this map represents the epitome of the reductionist
approach, in which an understanding of function can be elicited from a detailed
scrutiny of structure.
However, there is another side to the story. Despite the huge advances in
scientific medicine, the number of acute serious illnesses in comparable groups of
adults in the USA nearly tripled between the 1920s and the 1980s (Gillon and
L.
■
The biomedical tradition
23
Wesley, 1998). In 1974, the year before Ivan Illich published his seminal Medical
Nemesis (Illich, 1975), a Senate investigation reported that approximately 2.4
million unnecessary operations were being performed per year in the USA at a
cost of $3.9 billion and, more importantly, with a consequence of 11 900 deaths more than the annual number of military deaths in Vietnam (Porter, 1997)
The analysis by Illich (1975) further fuelled this debate, arguing that conven
tional medicine had claimed a monopoly on the interpretation and management
of health, well-being, suffering, disability, disease and death, ultimately to the
detriment of health itself. Illich took the view that health broadly encompassed
the processes of growing up, ageing, disease and death, using the coping
mechanisms embedded in the culture and traditions of communities. Physicians,
Illich argued, were now in the unenviable position of feeling that they had an
obligation to make available any intervention which was available. There seemed
to be an imperative among physicians to adopt a 'can do, will do' technological
approach. Patients themselves began to register their dissatisfaction with what
Skrabanek and McCormick (1989) later dubbed 'coercive healthism' by consult
ing alternative therapists, whose total number in the UK in the last quarter of the
twentieth century was around 30 000, nearly equal to the total number of general
practitioners (Le Fanu, 1999).
The dominance of the medical model predicated on science was called into
question in the last quarter of the twentieth century by Engel's critique entitled
The need for a new medical model: a challenge for biomedicine (Engel, 1977). Engel
recognised that the science-based biomedical model was 'now the dominant
model of disease in the western world' (Engel, 1977), and criticised it on three
grounds, namely that it was reductionist, dualistic (disconnecting the body from
the mind) and perfused the descriptions of bodily functions with mechanical
metaphors. Although this model had a fairly wide folk appeal (Morris, 1998), it
has suffered from confusion as to how the model addresses the notion of
positivism and the role of science in medicine. In essence, in Engel's model the
molecular phenomena remain firmly as the basis upon which the biological and
social levels of reality are built. His analysis purports to give these levels equal
weight, but he does not address the prerequisite elements at the core of medicine.
Indeed he does not actually advocate that the psychosocial issues be placed at the
centre of medicine, but only that they need to be taken into account (Puustinen,
2000). On the one hand Engel's position can be seen as subjecting psychological
and social factors to scientific positivism, in which case any and all areas of life are
vulnerable to medicalisation. On the other hand, the biopsychosocial model can
be interpreted as a mechanism for excluding some areas from the domain of
scientific positivism - but then one is left with the contentious decisions about
which areas are, or are not, included within those realms. This confusion has
never really been addressed, with the result that, although the model has never
een actively opposed, its influence has been indirect at best and marginal at
worst (Wulff, 1990; Morris, 1998).
Around the same time as Engel was advocating his new model, patients
themselves developed a less meek and accepting role in the medical encounter,
oregoing the assumption that doctors always know best and that patients do not
really want to know what is wrong with them, because it might cause them
anxiety. One English GP who was interviewed in the 1980s admitted that T find
t ie older working classes and generally the lower middle classes of all ages easier
24
Complexity in primary care
•• -i.'
to deal with than my own sort. My own sort ask complicated questions and are
often dissatisfied with the answers' (Cartwright and Anderson, 1981).
A closer examination of the chronology of scientific discoveries in medicine
during the twentieth century suggests that by the late 1970s many of the more
spectacular discoveries had been made, and that from then onward there was a
relative decline in major advances (Le Fanu, 1999). Even self-confidence within
the profession itself has been damaged - for example, by the inability to contain
the AIDS epidemic, and the confusion over the relationship between bovine
spongiform encephalopathy (BSE) and Creutzfeldt-Jakob disease (CJD), which
sparked further fears with regard to other potential pathogens introduced into the
food chain as a result of ignorance.
The propriety and respect generously afforded to the medical profession
throughout most of the twentieth century have been seriously damaged by the
egregious and well-publicised failures of some prominent clinicians, most notably
the cardiac surgeons in Bristol Royal Infirmary and the mass murderer Harold
Shipman. Offering his predictions for the twenty-first century at the end of his
vast tome. Porter concludes:
We have invested disproportionately in a form of medicine whose benefits often
come late, which buy little time, and which are easily nullified by external,
countervailing factors. Punitive interventionist medicine has played a modest
part in shaping wider morbidity and mortality patterns within the community,
and in terms of its professed aims - the greatest health of the greatest number the Olympian verdict must be that much medicine has been off target. (Porter,
1997)
Sir David Weatherall, Regius Professor of Medicine at Oxford (quoted in Porter,
1997) pointed to the root cause. 'The trouble is'. Sir David observed, 'although we
have learned more and more about the minutiae of how these diseases make
patients sick, we have made little headway in determining why they arise in the
first place.'
During the last quarter of the twentieth century, developments in fields that
had not previously been directly related to medicine, such as computing,
mathematics and ecology, began to inform a paradigm debate that was emerging
within the medical profession (Kernick and Sweeney, 2001). This debate focused
on observations provided by complexity theorists on a wide range of biological
systems, which appeared to challenge the reductionist approach and linearity of
the contemporary medical model. I shall explore the contribution of this field and
its relevance to the medical model in Chapter 7.
Linearity in the National Health Service
The success of the scientific method extended well beyond the fields in which it
originated. Drawing on the scientific paradigm, Adam Smith and David Ricardo
advanced the laws of economic interaction (Fukuyama, 1993), and sociology and
politics also tried to become sciences. This linear paradigm reached its zenith
towards the middle of the twentieth century, when it was applied to modernisa
tion theories of Third World development, international relations and public
policy (Geyer, 2001).
It also influenced organisational development, not least in the thinking that
1
The biomedical tradition
25
underpinned the development of the National Health Service (NHS) in the UK.
The machine metaphor for Newton's universe was translated into a desire to
understand the structure and parts of all systems, and to build organisations on
that basis. As early as 1911, Taylor published his theories of scientific manage
ment, with the underlying model of the organisation as a machine. The import
ance of mechanical metaphors in this explanatory model cannot be
underestimated. Even up to the 1990s, these deeply embedded beliefs were
expressed in the management predilection for re-engineering - a movement
whose weaknesses were attributed to its failure to acknowledge the role of
ordinary frail people in organisational life (Hammer, 1995).
In the UK, the NHS enthusiastically embraced this Taylorist approach, but only
after an initial period, spanning its first 25 years, of management by diplomacy
(Harrison, 1988). The command-and-control approach in the NHS, recalling the
wartime circumstances in which hospitals were first brought under central
control, is epitomised by the far-reaching Hospital Plan for England, which was
devised in the 1960s but was not implemented for another decade (Department of
Health, 1962). In the early 1970s, healthcare policy strategists agreed that the
administration of the NHS was in need of an overhaul, but their solution, in the
form of the Grey Book (Department of Health, 1972), testifies further to the
confidence in linear, algorithmic solutions at that time. The Grey Book adopted a
rigid, centralised approach to resolving the administrative confusion in the NHS.
However, its insensitivity to local context exposed its inherent weaknesses almost
as soon as its recommendations were implemented (Kember and MacPherson,
1994), and the structural reorganisation that it underpinned is generally viewed
as a failure (Klein, 1989). The Thatcher reforms of the NHS, which were
instigated after the review conducted at the then Prime Minister's request by
Sir Roy Griffiths, demonstrate the extent to which strategy in healthcare was
influenced by commercial business philosophy. In his much quoted comment in
which he lamented the lack of managerial accountability, Griffiths wrote that 'If
Florence Nightingale were carrying her lamp through the corridors of the NHS
today, she would almost certainly be searching for the people in charge'
(Department of Health and Social Security, 1983). Over a decade later, when
Robinson and LeGrand (1994) evaluated another set of NHS reforms, they were
unable to say whether the reforms had had a positive or negative impact. Hinting
that the problems which were encountered as a result of the contemporary
approach to healthcare strategy were more complex than had been envisaged,
Robinson and LeGrand (1994) wrote that 'There are rarely simple answers to
simple questions, usually because the questions are not actually simple.'
Summary
The general picture painted in this overview follows a widely accepted view of the
history of medicine which characterises medical developments as involving a
continuous, inexorable rise in the understanding of science, predicated on a
rationalist positivist ontology. The evidence to support this picture is substantial,
and while open to the considerations discussed above, it has helped to shape the
view that the dominant biomedical explanatory model is predicated fundamen
tally on a reductionist conception of disease, in which the body acts as a machine,
with disease events proceeding in a linear way through a sequence of measurable.
<
2
26
Complexity in primary care
regular and predictable causes and effects (Morris, 1998). The way to understand
a diseased state, according to this model, consists of a detailed scrutiny of the
structure of that state, achieved by reducing the whole to its constituent parts,
and assuming that the 'whole' of that diseased state was the sum of those parts.
Moreover, drawing on Newtonian thinking, it is assumed that those parts interact
in a regular and predictable way, from which generalities of the diseased state
may be constructed.
Not all commentators agree with this view. Almost a lone voice, Greaves (1996)
proposes that this view, whose central pillar is the notion of a predetermined
thread leading to the current understanding, is in fact a post-hoc rationalisation,
and that it is weakened at its very source, namely the germ theory proposed by
Koch. Theorists explored the contribution of social and psychological factors
(Engel, 1977) but, despite some intriguing evidence - presented in the medical
profession's leading journals - that cultural factors seemed to play a part in
determining physicians' and surgeons' own practice, the impact of Engel's
expanded model remained modest (Brook et al., 1988; Garratini and Garratini,
1993). Others, notably Le Fanu (1999), argue that the spectacular advances in
medicine occurred across a rather narrow spectrum and were more prevalent in
the first three-quarters of the twentieth century than in the last 20 to 30 years.
The scientific paradigm influenced fields well beyond those in which it originally
developed. Throughout its history the NHS has experienced a series of reforms, all
of which demonstrate a linear, mechanical approach to organisational develop
ment, testifying to the influence of that paradigm.
Had the conventional biomedical model, by the late twentieth century,
exhausted its repertoire of problems to address, as Pisek (2002) argues? How
might the observations from the complexity sciences, which were becoming
increasingly widely reported in the medical literature, affect the model? I shall
return to these questions in Chapter 5. To take this argument further, I shall now
look at the contemporary manifestation of the biomedical explanatory model, in
the form of evidence-based medicine. This model currently constitutes the
bedrock of clinical medicine, and its integration into mainstream clinical thinking
has embedded a hierarchy of evidence into current clinical practice. The nature of
this evidence, refined as it is in various forms, the purest of which is considered to
be the randomised controlled trial, remains resolutely scientific.
But think again of Mrs B. 'Jack's dead', she exclaimed, 'and the boys have
gone.' Her intervention was pivotal in that seminal consultation. What explana
tory model was she deploying? Chapter 4, which traces the origins of the
principles of what we now call qualitative research, addresses this question.
First, however, I shall explore the current manifestation of the scientific method
in clinical practice, in the form of evidence-based medicine.
li
ilb.
Chapter 3
• .... -
Evidence-based medicine:
the contemporary manifestation of
the explanatory model in medicine
Introduction
The overview of the historical rise of science in medicine that was presented in
Chapter 2 supports the proposition that, in the nineteenth and twentieth
centuries, science came to occupy a hegemonic position in the explanatory
model in medicine.
There are certain important epistemological implications for that model,
namely that it is predicated on rational positivism, with an ontological conception
of disease as real and distinct entities awaiting discovery and classification. In this
chapter, I shall dig deeper into the explanatory model by describing and critiquing
its contemporary manifestation in the form of evidence-based medicine (EBM).
The purpose here is to develop the argument that is unfolding in this book, from a
justification of the proposition that science dominates the model to an exploration
of the implications of that model for everyday general practice. The aim is to
arrive at a position in the argument where it can appropriately be held that,
although science is a necessary (and indeed vital) component of the explanatory
model, it is insufficient in itself to constitute that model.
Let us recall what is meant by the term 'evidence-based medicine.' The method
is described as having five steps (Evidence-Based Medicine Working Group, 1992;
Oxman et al., 1993; Davidoff et al., 1995; Rosenberg and Donald, 1995; Dawes,
1996; Sackett et al., 1996).
• A clear clinical question is formulated from the patient's presenting problem.
• A literature search is conducted in order to identify relevant published articles
that consider the problem.
• The evidence is accessed and evaluated (or critically appraised).
• Valid and useful findings are implemented in clinical practice.
• An audit of performance is conducted.
The 'evidence' in EBM refers to a hierarchy in which the findings of fully
randomised controlled trials are at the top, followed by less well-controlled
trials, cohort studies, expert consensus views and authoritative opinion.
This chapter explores EBM in four stages. First, the principles of EBM are
explained and placed in context (Dawes, 1996). Next, 1 reflect on a description
of the method (Rosenberg and Donald, 1995) and an early example of its
application in an acute medical unit of a teaching hospital (Ellis et al., 1995),
1
■
28
Complexity in primary care
both of which were quite heavily criticised in the correspondence columns of
the peer-reviewed journals in which they appeared. In the third section, I briefly
explore how EBM has been integrated into the current framework of general
practice. Finally, I pose the following question. How could patients in general
practice benefit from EBM?
The principles of EBM
||l! I
i
Right at the outset, the instigators of EBM claimed that it 'aimed to de-emphasise
unsystematic clinical experience and intuition as sufficient grounds for taking
decisions in clinical practice' (Evidence-Based Medicine Working Group, 1992).
Rather, the appraisal of evidence from clinical research was to be seen as the
evident and proper basis for clinical practice. The process through which EBM
was to be applied seemed to be straightforward. Patients present with problems
which are turned into clinical questions. The answers to these questions are then
based on the best available evidence, which is obtained by searching for and
critically appraising the relevant studies.
After this initial description was published, a series of articles in the Journal of
the American Medical Association entitled the 'Users' guides to the medical literature'
illustrated the principles of EBM in greater detail (Guyatt and Rennie, 1993;
Guyatt et al., 1994), and demonstrated the importance of four basic outcome
statistics, namely risk ratios, relative risk reductions (the complement of the risk
ratio expressed as a percentage), the calculation of absolute risk reduction, and
finally the numbers needed to treat (the inverse of the absolute risk). These
proved to be extremely helpful in clarifying evidence at the population level.
What, then, were the advantages to which the advocates of EBM wished to
draw attention? First, it offered a better way of deciding which treatments or
clinical approaches could be usefully incorporated into practice, and which
should be discarded. In turn, this could inform better decisions with regard to
commissioning and providing services. It also provided a common language for
critical appraisal of new evidence, which in itself might encourage the design of
better trials and provide a sounder basis for undergraduate and continuing
postgraduate education.
EBM wasn't rocket science. Its principles could be shared by healthcare
professionals from different backgrounds (including non-clinicians), and by lay
people, at any stage of their careers (Rosenberg and Donald, 1995).
However, early on in the adoption of EBM, practical problems emerged. It was
conservatively estimated that it would take two hours to focus on a clinical
question and then find, appraise and act on the evidence (Rosenberg and Donald,
1995). To do this, one would require access both to the appropriate computer
software and to the Internet. Although it wasn't difficult, it took time to acquire
the skills necessary to practise EBM properly. A plethora of protocols, guidelines
and clinical synopses was produced to address these practical teething problems
(Dawes, 1996).
Applications
Within three years of the publication of the principles of EBM, two seminal
examples of its application in clinical practice in the NHS were published, one in
=
Evidence-based medicine
29
secondary care and the other in primary care (Ellis et al., 1995; Gill et al., 1996). In
the former example, over 100 acute medical admissions were studied in order to
ascertain how many of the treatments that had been offered to the patients were
evidence based. Around 82% of the patients were judged to have received
evidence-based interventions; 53% had received interventions supported by
one or more randomised controlled trials, and 29% were offered interventions
that had been unanimously judged to be supported by convincing non-experimental evidence. Similar estimates were obtained in a study of two days of
consecutive consultations in general practice (Gill et al., 1996). Around 81% of
the patients in this study received interventions that were either supported by
one or more randomised controlled trials or were supported by convincing nonexperimental evidence.
Thus EBM seemed to represent the pinnacle of the scientific tradition under
pinning clinical medicine. It embedded a hierarchy of evidence, encouraged
critical appraisal to distinguish between good and poor evidence and, with its
clear explication of the concepts of absolute and relative risk, it made a major
contribution to increasing our understanding of population-based evidence. How,
then, could it attract the substantial body of criticism that it did? Many of the
critiques focused on a central issue. Doctors don't practise medicine in popula
tions - they consult face to face with individuals. Let us now briefly consider the
gist of these objections.
!
a
Reservations
Iggo (1995) and Charlton (1995a) both challenged EBM on the basis of its central
strength, namely the robustness of the evidence that forms its basis: 'a major fault
with evidence-based medicine is its emphasis on randomised controlled trials'
(Iggo). The results of such trials cannot be easily extrapolated to a particular target
group, they argued, without taking into account relevant contextual knowledge.
Some critics, such as Fowler (1995), were clearly angered by some of its claims.
'Evidence-based medicine', argued Fowler, 'is a neologism for informed decision
making . . . the presumption is made that the practice of medicine was previously
based on a direct communication with God or by tossing a coin.' Other critics
(Bradley and Field, 1995) of the paper by Gill et al. (1996) argued that assignment
of diagnoses in the study achieved a reduction in complexity of management at
the cost of 'drifting from the reality of many patients presenting with more than
one problem.' And, they asked, 'what about people who did not have a
diagnosis?'
One of the most enduring criticisms of the model accused it of measuring only
what is measurable: 'There are no suggestions on how practice relating to un
measurable aspects is to be guided' (Smith, 1995). And Bradley and Field (1995)
stated that: 'Not all that is measured is of value, and not all that is of value can be
measured.'
One of the key claims repeatedly made by proponents of EBM was that it
constituted a new paradigm (Evidence-Based Medicine Working Group, 1992;
Sackett et al., 1996). It is important to look at this claim in some detail as, if
substantiated, it would have serious implications for the unfolding of a new
chapter in the epistemological evolution of biomedicine. Was the claim justified?
30
Complexity in pnmary care
The claim that EBM constituted a new paradigm
on
I:
Kuhn (1970) uses the term 'paradigm' to describe a set of shared assumptions or
received beliefs in a scientific community. His central thesis is that groups of
scientists imagine, think and research within a clearly defined set of assumptions
which define the possibilities and also the limits of their endeavours. They use a
technical language that encompasses a core group of theories, and then occupy
themselves largely by solving only those problems that this technical language
can express. Only when the need to answer questions that cannot be dealt with
within the assumptions of a current paradigm becomes pressing can the assump
tions be seriously challenged. Kuhn (1970) suggests that it is at (rare) times like
these that new paradigms appear. He describes their appearance as 'a reconstruc
tion of the field from new fundamentals, a reconstruction that changes some of
the most elementary theoretical generalisations.'
Judged against these two criteria, one cannot confidently support the proposi
tion that EBM constituted a new paradigm. First, its conceptual origins go much
further back than the key synthesis of EBM thinking. Clinical Epidemiology: a Basic
Science for Clinical Medicine (Sackett et al., 1985). In the last chapter I referred to the
work of Pierre Louis in Paris in the 1830s, who developed the first firm
epidemiological approach to clinical thinking by comparing the outcomes in
two populations of patients who were offered different treatments (Lilienfield and
Lilienfield, 1979). Clinical epidemiologists had been working with evidence
based approaches for much of the twentieth century (Armenian and Lilienfield,
1994), and in the 25 years or so preceding the publication of the EBM manifesto
(Evidence-Based Medicine Working Group, 1992), the number of published
randomised controlled trials - mainly in secondary care, but increasingly in
primary care - increased exponentially (Silagy and Jewell, 1994). Case-control
methods and statistical techniques were developed throughout the first half of the
century as sociologists began to tackle health problems as social issues (Lilienfield
and Stolley, 1994). Doll and Hill (1950) were clearly publishing the best available
evidence when in 1950 they reported one of the first studies linking smoking and
lung cancer. In a seminal overview of medical research, the pivotal importance of
randomised controlled trials was championed by Cochrane (1971), who described
them as a 'beautiful technique of wide applicability.'
Secondly, before the EBM protagonists laid out their position, others had laid
claim to introducing a paradigm shift. In theoretical terms, one could argue that
McWhinney (1983) had a stronger case. As we saw in the previous chapter, he
was one among a range of commentators who appeared to be reconsidering the
explanatory model in medicine. In the clinical paradigm that was presented in the
historical overview, disease was viewed 'objectively', independently from the
person who was suffering from it. Mind and body were considered separately,
each with their own diseases. In the early 1980s, McWhinney outlined what he
claimed constituted a new paradigm which defined a view of medicine that
'would emphasise the patient, the doctor-patient relationship, and the language
of illness' (McWhinney, 1983). Indeed, on a more muted level, the Royal College
of General Practitioners (1972) had been supporting this model as one of the
principles underpinning its vocational training.
Evidence-based medicine
3I
The notion of evidence-based diagnosis
However, it is within the notion of diagnosis that the epistemological preferences
of the EBM model are most interestingly revealed. In its first analytical step, EBM
encourages us to define a clear clinical question. Two implications follow from
this. First, if there is no clear clinical question, the model does not apply. Those
who support the model are clearly in agreement with this, and it is not intended
to apply to those circumstances, thus excluding the substantial number of general
practice consultations which proceed without a firm diagnostic label. Secondly,
where there is a clinical question, the framing of that question betrays a
preference for the type of knowledge that biomedical science can bring to bear
on the issue. Take, as an example, one of the EBM diagnoses that was identified
and explored in the key paper by Gill et al. (1996) from general practice. A patient
presents with worry, insomnia, palpitations and poor appetite. If a 'diagnosis' of
anxiety is made, the literature can be searched for randomised controlled trials of
benzodiazepines, and an appropriate evidence-based treatment applied. What has
happened is that those aspects of the issue that can be usefully collated to form a
'clinical' question, and only those aspects, are abstracted. In doing this one has
tacitly expressed a preference for selecting out those elements of the issue that can
possibly be illuminated by the type of knowledge which is constructed from the
epistemological standpoint of the model, in this case randomised controlled trials.
However, if in this patient's presentation one identifies a set of related contextual
issues as important - for example, a background of grinding poverty, a poor
marriage and estranged children - one could construct a different but equally
relevant question. In doing this one would again tacitly be expressing a prefer
ence for the epistemological standpoint in which narrative knowledge was the
accepted currency. One's gaze on the 'case' would have moved from the
biomedical to the biographical, just as occurred during Mrs B's consultation
described in Chapter 1. And, just as in her case, one could see the 'case' in
metaphysical terms, exploring the issues of hopelessness, despair, guilt, shame
and humiliation. Thus we can see that the EBM model sits as one way of gazing
upon issues - an extremely well-developed and robust way, but not a unique
analytical method. The analogy can be made with scientists exploring the nature
of light. They can 'look' at light as it 'appears' in terms of photons and packets of
energy, and they will learn a lot, but they will be completely oblivious to the
wave properties of the same phenomenon.
Thus, in general terms, the concept of 'diagnosis' in general practice is broader
than the strictly biomechanical diagnosis that is used in randomised controlled
trials, the cornerstone of the edifice on which the claims of EBM are based.
Patients most often present to their GPs not with diseases, but with illness stories
consisting of a collection of unorganised symptoms and signs. The process of
defining a disease from such an illness story is a selective and narrowing
abstraction (Kleinman, 1988).
Making a diagnosis is often an evolutionary process. Definitive diagnoses (in
this broad sense) may be made retrospectively or indeed not at all. Formulating a
diagnosis in general practice can take weeks or months, and the exact moment
when a diagnosis becomes 'clear' may be difficult to determine. Here the evidence
suggested that the general practitioner uses time appropriately as a powerful
diagnostic instrument (Balint, 1957). Often the general practitioner formulates a
2
d
3
32
Complexity in primary care
series of diagnostic 'hunches' early in the consultation, using clues from their
background knowledge of the patient and perhaps of the patient's family (Elstein
et al., 1972). Often the practitioner is forced to bypass confident diagnosis and
adopt a decision pathway that approximates to symptomatic treatment (Howie,
1972). Diagnoses are sometimes made retrospectively, when the treatment
confirms the putative diagnosis. However, if antibiotics are given for a sore
throat, the diagnosis of bacterial tonsillitis may be supported, but it is hardly
confirmed.
Until recently, the way in which medical students were taught diagnostic skills
reflected this predilection for the scientific method. Traditionally diagnosis is
taught to medical students (at least initially) as though it were an inductive
process. First the facts are gathered without the prejudice of preformed bias, and
only then, when all of the relevant pieces of the jigsaw have been assembled, does
the full pattern reveal itself. However, most clinical problem solving is better
taught as hypothetico-deduction (Royal College of General Practitioners, 1972)
or, putting the same idea more simply, 'guessing and testing' (Marinker, 1981).
For general practitioners brought up in this tradition, the early years in practice
involve a lot of 'unlearning' of approaches that have been acquired during
medical school education (Marinker, 1970; Williamson et al., 1979).
The evidence in EBM
EBM encourages the critical appraisal of the best available evidence in order to
solve clinical problems (Sackett and Rosenberg, 1995). For the most part, the
evidence referred to is found in the results of randomised controlled trials or is
collated in systematic reviews or meta-analyses. In setting out the principles of
EBM, its proponents argued very persuasively that the breadth and explanatory
power of a randomised controlled trial are the cornerstone not only for EBM, but
also for evidence-based public health, evidence-based hospital administration,
evidence-based purchasing and evidence-based consumerism (Sackett and Cook,
1994).
Let us remind ourselves what, precisely, a randomised controlled trial involves.
In a randomised controlled trial, a group of individuals - the study cohort - is
identified by virtue of its members sharing a common characteristic - for
example, a particular disease or risk factor. The researchers randomly allocate
each member of the cohort to one or more study groups, each of which receives
different interventions, preferably in a way that is not known to either the
researcher or the subject - so-called 'blinded intervention.' After a period of time
the outcome for each group is ascertained and the findings are compared. Thus
randomised controlled trials aggregate the benefits and disadvantages of alter
native interventions in comparable populations within the same experimental
environment. The idea is to balance the populations, the intervention and the
assessment process, so that the only difference distinguishing the two, at the end
of the observation period, will be attributable to the intervention. The key issue
for the clinician is to consider how that population-based outcome may be applied
to the individual patient.
The problem for all clinicians when using the results of EBM is simply this.
Faced with a patient's particular clinical problem, the doctor needs to know
whether the outcome statistic from a relevant trial applies to this patient, in this
Evidence-based medicine
r
1
1
t
7
>
I
33
consultation, in this particular environment. However, these vital questions
cannot be answered. Outcome statistics reflect the average experience of each
group, and extrapolation with precision from that group average to a particular
patient's chances is impossible. An extrapolation can be made, matching the
consulting patient's characteristics with those in any relevant study, but the
accuracy of this extrapolation is a matter of degree. What can be judged is the
average order of probability in a large group of people whose personal and disease
characteristics bear a resemblance to those of the patient.
In a clinical trial, a group of patients who share the same diagnosis may exhibit
quite different patterns of illness behaviour. They may exhibit different symptoms
or experience differences in the severity or rate of progress of the subject
condition. In randomised controlled trials, these individual experiences are
grouped together for the sake of simplicity and clarity, and the results are
summated results from the combined experiences of a heterogeneous population
of individuals. Although the intention is to distribute potential biases equally
between the two groups, this does not eliminate the biases - it just disperses them,
albeit in a particular manner. In this sense, comparisons between the groups will
be unbiased but they run the risk of conflating several causal processes. A second
difficulty concerns the relevance of the data from the population in the trial to the
situation of the individual patient in the consulting room. To find out the extent
to which the results of a trial should apply to a particular patient, the doctor needs
to know what kind of patients were in the trial.
This critique is not intended to diminish the contribution of the randomised
controlled trial to the clinical care of patients. Some of the simplest interven
tions have been subjected to this kind of trial. A classic example is the use of
aspirin in secondary prevention of ischaemic heart disease (Second Inter
national Study of Infarct Survival Collaborative Group, 1988), trials of which
have resulted in subsequent worldwide benefit for individuals with this
condition (Collins et al., 1996). However, the issues of context and extrapolation
remain difficult, particularly for individuals in primary care. In this domain,
some authors take a fairly robust stance. For example, Barbara Starfield (2001)
reported that: 'Evidence-based medicine is surely a desirable approach to
ensuring the quality of practice; however, existing evidence is not for the
most part appropriate for primary care.' Starfield identifies three major flaws in
the design of the trials that contribute to the evidence base of primary care. In
general, they are seriously underpowered to detect any but the commonest
adverse events. This means that when we extrapolate from small or even
modest trial populations to large national or even continental populations, we
do not know quite what degree of harm we might inflict on those populations.
Secondly, they fail to take into account the nature of the primary healthcare
that the study subjects receive while they are participating in trials. Starfield
herself has shown how the absence or presence of a relationship with a source
of primary care can itself be expected to influence the outcome of medical
interventions. However, the greatest flaw in the evidence base is the absence of
evidence with regard to comorbidity. A defining feature of the randomised
controlled trial is that it excludes people with coexisting medical conditions, as
this would confound the relationship between intervention and outcome. Yet
we know that a quarter of people over the age of 65 years will have three or
more comorbid conditions. What we know less about is how the parallel
I
34
i
1
Complexity in primary care
medications for such comorbid conditions might interact over the decades of
treatment that such patients will endure.
The issue of concern, then, is the epistemological framework that is deployed to
address the genuine uncertainty about treatment benefits in important clinical
conditions. However, the type of knowledge that is produced in randomised
clinical trials can only provide clarification at the level of populations. We can say
what the general direction of advantage is and what the general proportion of risk
might be, but even when the evidence is clear (according to the conventions of
that model), practitioners are left uneasy about the numbers needed to treat.
Consider the evidence with regard to anticoagulation. If we treat 1000 patients,
we will prevent between 15 and 40 ischaemic strokes or systemic embolisms, at a
cost of inducing between four and six serious haemorrhagic events. Smeeth et al.
(1999) have drawn attention to the subtle difficulty involved in calculating the
numbers needed to treat (NNT) from meta-analyses. Describing them as 'some
times informative, usually misleading', those authors remind us that the numbers
needed to treat are sensitive to factors that change the baseline risk, such as
outcomes considered, patient characteristics, secular trends in incidence and case
fatality, and clinical setting. Pooling the numbers needed to treat derived from
meta-analyses may be misleading because the baseline risk often varies between
trials. In some ways this is like a social contract with the population al risk people voluntarily agree to expose themselves to a degree of personal risk in order
to benefit both themselves and, by extrapolation, the social whole.
Although clinical decisions in this context are reasonably clear (to clinicians at
any rate) in terms of a risk-benefit ratio, this type of approach becomes much
more problematic when the clinical intervention itself is set to trigger not
following a particular clinical event (such as the confirmation of atrial fibrilla
tion), but at a biochemically determined level of risk. Consider how this
population approach now plays out in the field of cardiovascular prevention. If
you take just one European guideline for cardiovascular disease, and extract just
two risk factors (blood pressure and cholesterol level), three-quarters of the entire
European adult population will be identified as being at risk (90% of those over
50 years of age), all requiring external monitoring, and many requiring medica
tion to modify these risk factors. Although nearly all of the risk scores for
cardiovascular disease are based on the Framingham risk equation, Fahey and
Schroder (2004) have pointed out that this equation does not provide a truly
accurate assessment of an individual's cardiovascular risk. Their review suggests
that the Framingham figures overestimate both fatal and non-fatal coronary heart
disease by about 60%. There is also a documented variation in the way in which
these figures are applied, with overestimation occurring in areas where the
mortality rate from heart disease is lowest - in England, where the average
overestimation is 70%. The overestimation is lowest in areas where the mortality
rate from heart disease is highest - in Scotland, where the average overestimate is
around 30%.
One discerns in this trend towards clinical intervention at a biochemically
determined level of risk a consolidation of the hegemony of the contemporary
explanatory model in medicine. There is an authority percolating the provision of
clinical advice based on the types of algorithms critiqued by Fahey and Schroder
(2004) - reflected, for example, in the precision of the risk calculation figures. The
concern here is that patients, on receiving information about their level of risk,
■
Evidence-based medicine
35
might feel bullied into accepting the need for continuous vigilance with regard to
their health, medical supervision, and often the consumption of medicines for
decades. The philosopher Lionel Trilling cautions us about transactions of this
nature: 'Any proposition delivered without a hint of doubt about its validity is a
form of bullying' (Delblanco, 2001).
At its most simplistic, it may be unwise to assume that what is statistically
significant to the researcher and clinically significant to the doctor will also be
personally significant to the patient. Consider the rationale for treating hyperten
sion in a middle-aged man. One could explain to such a patient that 170 middleaged men would have to have their blood pressure treated for five years in order
to prevent one stroke. The patient may well elect to take his chances with the
other 169 rather than accept the sick role of a patient who requires daily
medication for the rest of his life.
We know that people's attitudes to health are not necessarily or exclusively
logical (Johnson, 1995). A patient's attitude to health, and the actions that are
taken on that basis, are determined by how that person perceives a particular
threat to health, the strength of their belief in the advantages of changing their
behaviour to accommodate that threat, and how difficult they believe that
behavioural change will be. The beliefs that form attitudes to health are
influenced by personal, family, social and demographic factors (Becker, 1974).
An individual's actions based on these personal health beliefs are not always
apparently rational (Johnson, 1995). For example, research has demonstrated a
correlation between the frequency of prescribing psychotropic drugs for mothers
and the frequency of prescribing antibiotics for their children in order to treat
respiratory infections (Howie and Bigg, 1980). People's actions are influenced by
what they think others might expect them to do and by how much importance
they attach to those expectations (Fishbein and Azjen, 1975).
This analysis begins to call into question the broader nature of 'significance'. Let
us explore this a little further by considering one further criticism of EBM.
>
r
3
-i
Measuring the measurable
The descriptions of EBM and the illustrations of its application, to which I have
referred above, were criticised by Ellis et al. (1995), among others, for posing only
questions the answers to which could be measured, and therefore for implying
that if something cannot be measured it is of little value. Rudebeck (1992)
expresses it bluntly: 'the requirements of medical research are limited by insisting
that an answer should be numeric, otherwise it is not a real answer.' Patients or
their representatives pose different types of questions, some of which have
numerical answers whereas others are more reflective, philosophical or existen
tial (Dixon and Sweeney, 2000). The problem of the nature of research questions
is touched upon in a trenchant critique of academic research in clinical medicine
(Rudebeck, 1992). It is worthwhile considering the basis of these criticisms, as
they shed light on the epistemologies that compete for focus and meaning in
consultations.
Consider the following example, which is taken from one of the articles in the
users' guide series published in the Journal of the American Medical Association by
the EBM Working Group (Laupacis et al., 1994).
36
Complexity in primary care
Mrs J is a 76-year-old retired schoolteacher who consults because she thinks she
is becoming forgetful. The doctor's assessment includes a formal test of her
mental state and a series of biochemical tests, the results of which are all
normal; this confirms that the patient is suffering from dementia. Her son asks
to see the doctor about the problem.
HUI
i'
£
The authors of this article suggest the type of questions which the son might ask
the doctor in this situation - questions which can be appropriately dealt with by
EBM. They explore two questions: 'What is my mother's prognosis?' and 'Will my
mother be alive in five years' time?' They then go on to show how a rigorous
appraisal of the relevant literature can produce quantifiable answers to these two
questions. The selection of these questions is instructive because they can be
given answers that are expressed numerically.
However, consider the example of the 76-year-old woman with dementia. The
patient's son may ask the general practitioner the following questions: 'Can my
mother be cured?', 'Will she die?' and 'Can her suffering be relieved?' The
answers to these questions will certainly involve harnessing skills that are set out
usefully by the EBM Working Group. However, the son may also ask 'Why has
this happened to her?' and 'What will happen to us now?' He may well reflect on
the impact that his mother s illness will have on his own family: 'Should my
mother come and stay with us? and 'What will be the effect on my marriage if
she does?'
So it is not always possible to formulate from the patient's presenting story a
question whose answer can be obtained from the findings of biomedical research.
This should not be interpreted as an inherent weakness of EBM, but it does define
its epistemological boundaries.
From the viewpoint of general practice, the concern is that an overemphasis on
the questions that can be answered by EBM may run the risk of devaluing those
questions that cannot be so expressed - and in turn deflecting attention towards
the type of knowledge that populates those questions. The consultation in general
practice is not simply a place where a patient seeks scientific answers to questions.
Toon (1994) states that: 'The consultation is the patient's forum for coming to
understand her illness, not merely a rational understanding, but an under
standing which involves the emotions and which contributes to the growth of
the individual. In a patient-centred consultation, the doctor recognises this and
tries to see the illness through the patient's eyes (Peppiatt, 1992). Patients'
questions are answered partly by reference to science, but also by reference to
what Heath (1995) calls the search for meaning.' There is something in all of
these descriptions that calls us to consider the relationship of the frameworks
from which quantitative and qualitative knowledge arises. They also demand that
we revisit the notion of 'significance', which in the conventional biomedical
model is presented on two levels - statistical and clinical. I shall pursue this point
later in this chapter. First, in order to develop a clearer idea of what a
complementary epistemological framework might look like, let us reflect briefly
on the broad nature of general practice.
Evidence-based medicine
37
The nature of general practice
Any experience that an individual perceives as threatening or problematic and
introduces to the consultation becomes the legitimate concern of the general
practitioner. Being competent to deal with the range of potential problems that
are encountered in general practice requires a portfolio of intellectual, profes
sional and personal skills. Some aspects of the general practitioner's competence
have been described as doctor centred (Guilbert, 1987). They include curative and
rehabilitative care, promotion of health, and organisation of preventive activities.
Personal skills are equally important components of competence. They are the
more patient-centred qualities of competence, and include moral and personal
attributes.
In the USA, the American Board of Internal Medicine (1985) stated that: 'A
major responsibility of training residents in internal medicine is to stress the
importance of the humanistic qualities in the relationship between the patient
and physician.' In the UK, these competencies now form a central plank of the
recertification processes in the NHS, and it is recommended that 'the recertifica
tion of all clinicians should include an assessment of honesty, self-awareness in
the professional context, empathy, respect for patient autonomy and confidenti
ality' (Southgate and Jolly, 1994). The recent acceptance of the new general
medical services contract by the medical profession makes the consultation
process with patients mandatory, and rewards practices who can show that
they have acted upon such consultation (British Medical Association, 2003).
What these initiatives emphasise is the importance of forming a human
relationship between the patient and the doctor. The formation of such a
relationship is the basis of good consulting, and it requires the ability to identify
imaginatively with what a patient experiences subjectively during an illness, and
to recognise the validity and importance of that experience for the patient. The
term 'human relationship' is used here with a specific meaning to distinguish it
from the term 'personal relationship.'
For doctors this distinction is important. In a personal relationship, two human
beings decide spontaneously to exchange benign sentiments based on mutual
care, concern and affection. As the relationship develops, the mutuality and
reciprocity become more profound on both sides, and the way in which a
personal relationship develops depends on the equal participation of both parties.
In a human relationship, one person, responding to the human condition of
another, seeks to initiate an exchange of benign sentiments (Sweeney, 1992).
Human relationships require a genuine concern from the instigator and a positive
response from the participant. Such relationships are based on self-knowledge,
tolerance, self-confidence, patience and the ability to listen and communicate.
In human relationships the depth of feelings and personal involvement should
never be as profound as in a personal relationship, and doctors have a duty to
maintain these relationships at an appropriate level. It is this element of control maintaining the relationship al the appropriate level - which epitomises the
difference between the two types of relationship. To illustrate how the tasks of
applying best evidence and forming human relationships can be combined,
consider the following example.
w
J
.•f
J
38
Complexity in primary care
Bert, a former professional soldier, is 63 years old. He consults one day saying
that he feels low. The doctor carries out a standard depression inventory to
confirm the diagnosis based on Bert's symptoms, and he identifies Bert as a
'case' of depression. Calling on the evidence from clinical trials, he prescribes a
tricyclic antidepressant.
noii
hi:::
h!:::
i
A
Here the appropriate use of EBM helps the doctor to arrive at the diagnosis of
'depression' by using a validated instrument, and helps the patient to benefit from
the best treatment suggested by the relevant clinical trials. After talking further
with Bert, the doctor establishes that the depression was initiated by the failure of
Bert's son to get into the army. Bert had been a soldier himself, loved the army,
and had always wanted his son to join the same regiment. His son also wanted
this, but when he failed to do so he became antisocial, drinking too much and
experimenting with soft drugs. Bert had found the experience devastating.
This example illustrates the three levels at which a doctor can understand and
empathise with a patient. At the first level, the doctor categorises the patient on
the basis of established, recognisable patterns of disease - Bert is a 'case' of
depression. At the second level, the doctor makes use of the patient's life history
to try to understand what has been happening to him - Bert's aspirations for his
son are based on his own experiences as a soldier. At the third level, the doctor
recognises the uniqueness of the patient's human condition and the significance
and gravity of the illness from the patient's perspective. Accordingly, in this
consultation the use of EBM and the formation of the human relationship are
both necessary ingredients of a productive consultation. Neither the diagnostic
technique nor the relationship-building skills are in themselves sufficient to
constitute an effective consultation. Heath (1995) identifies two fundamental
roles for the general practitioner, namely acting as an interpreter and guardian at
the interface between illness and disease, and acting as a witness to the patient's
experience of both.
Consider the patient with end-stage chronic obstructive airways disease, for
whom no more can be done by the respiratory specialists. Such patients may
linger on in a terminal state for months or years, deriving gradually less benefit
from steroids, antibiotics and bronchodilators. Their burden of illness and the
burden on their carers need to be witnessed and supported despite the failure of
biomedicine to help them.
The distinction which appears to be fundamental for the general practitioner is
that between illness and disease. Illness may exist without disease, which is a
narrower reconfiguration of the patient's story based on biomedical science
(Kleinman, 1988). Doctors have a duty to explore the meaning of illness,
recognising its potential origin in unhappiness, unfavourable socio-economic
circumstances or chronically unsatisfactory personal relationships (Morris,
1998). They need to recognise that the severity of an illness can be a function
of an individual's health beliefs and particular life circumstances (Becker, 1974).
There is a further duty to protect patients from overenthusiastic interpretation of
illness as disease. Kleinman (1988) cautions doctors against the 'over-literal
interpretation of accounts best understood metaphorically.' Here the doctor
works with the patient to make sense of the patient's experience in the context
of the rest of the patient's life. This has been described elsewhere as the
hermeneutic (Toon, 1994) or biographical (Marinker, 1994) tasks of the doctor.
Evidence-based medicine
•f
r
f
1
1
i
i
f
39
What this analysis emphasises is the gravitas and relevance of knowledge that
arises out of context, out of personal narrative and, crucially, out of the
interaction between doctor and patient in consultations. This interaction - the
undisputed basis of any relationship - is a mutually changing process, as it
gradually and subtly changes both participants, their own relationship, and the
way in which they interact with others.
The EBM Working Group argued that many patients are being denied the
benefits of evidence from well-conducted clinical trials (Sackett and Rosenberg,
1995). The evidence supports this view (Haines and Jones, 1994). Accordingly,
doctors have a responsibility to acquire new skills in accessing, evaluating and
applying the results of the vast number of clinical studies that are now being
published worldwide. In pursuit of this end, the integration of EBM into the
clinical work of general practice is both necessary and timely. The danger lies in
regarding EBM as the only and last word in the pursuit of clinical quality and the
search for relevant evidence. For example, the conventional approach to the task
of modifying the behaviour of doctors has been to offer educational inputs or to
formulate clinical guidelines. To some extent these work (Grimshaw and Russell,
1993). However, the decision to change a clinical habit can be the result of a series
of influences, which may not be dealt with by a conventional approach where the
researcher has to guess in advance which intervention is most likely to succeed.
Yet patients, or their representatives, also pose different types of questions, some
of which have numerical answers, while others have answers that are more
reflective, philosophical or existential (Dixon and Sweeney, 2000). The problem
of the nature of research questions is touched upon in a trenchant critique of
academic research in clinical medicine, which argues that the requirements of
medical research are limited by insisting 'that an answer should be numeric,
otherwise it is not a real answer' (Rudebeck, 1992).
However, this is not a dichotomous taxonomy of questions, numerical or
otherwise. Mary Midgley (1992) states that: 'The mere presence of an emotional
factor in any kind of decision does not take it out of the realm of rational thought.
All our thinking involves emotional factors as well as rational ones, just as every
physical object has size as well as shape. These are not alternatives. The presence
of one does not mean the absence of another.' The challenge lies in integrating
one with the other. Midgley (1992) expresses it thus: 'We need ways of thinking
which are unifying enough to give us guiding patterns, but not so strongly
reductive as to leave out something important.'
A seminal study of doctors' reasons for changing their prescribing behaviour
illustrates this point (Armstrong et al., 1996). The researchers encouraged the
clinicians to explain and understand their own reasons for altering their pre
scribing habits. The researchers identified three models of change. Some doctors
altered their prescribing policy as a result of the sheer pressure to move in a
certain direction, evidence for which came in various forms, such as articles,
letters from consultants, and talks that they had attended. A second model of
change was more abrupt, and occurred in response to near 'clinical disasters' or to
changes in prescribing initiated by a second doctor, sometimes a locum. A third
group of doctors seemed to embrace change in general more enthusiastically.
They were more prepared to be influenced by a variety of sources, some of which
would be regarded as conventionally authoritative, such as journal articles, while
others would not, such as magazine advertisements and television programmes.
i
j
i
J
40
MIHII
S'"
Complexity in primary care
This important article is saying something fundamental about personal epis
temologies - that is, the type of knowledge that an individual regards as 'true' or
valuable. One can have what psychologists call a 'naive' epistemology where, at
risk of simplifying a complex literature, knowledge is regarded as something
certain, carefully teased out of issues that are seen as solvable puzzles. Con
ventionally, people who hold such naive epistemologies defer to experts and
authorities as the source of valuable knowledge. At the other end of a knowledge
continuum, one can hold a more complex view of the nature of knowledge,
seeing it as always contingent, uncertain and arising from a range of domains,
including personal narrative and context as well as more objective 'factual'
knowledge. Here individuals tend to be more anti-condescensionist, regarding
authority as always potentially fallible. At its extreme, this stance may be
described as postmodern, sceptical and constructionist. It can, for example, be
seen in the lack of authority of secularism and science in contemporary society in
the USA (Pederson, 2005). One is driven, through this analysis, to consider how
we know things, to entertain the possibility that there is more than one way of
knowing and that, by implication, the current explanatory model in medicine,
predicated as it is on biomedical positivism, is one way of knowing, which reflects
a particular world-view and a preference for a particular type of knowledge. At
the spearhead of medicine's explanatory model we find the notion of significance
- the convention whereby we judge whether knowledge produced through this
framework's processes is worthy of interest or action. If we now consider that
there may be more than one way of knowing, then we should revisit the notion
of significance, take a view of its limitations, and theorise about how it might
usefully be extended. This brings us to the notion of personal significance
(Sweeney et al., 1998).
Personal significance
At stake in this part of the argument are the nature, importance and relevance
of subjectivity, a notion of little relevance within the conventional medical
model. In that context, subjective evidence is anathema. In this context, EBM is
almost always doctor centred. It focuses on the doctor's objective interpretation
of the evidence, and it either jeopardises or diminishes the importance of
human relationships and the role of the other partner in the consultation the patient.
Earlier in the chapter I described the current position, whereby the importance
of research evidence is weighted with mathematical models which, within each
study, describe its importance according to two levels of significance - statistical
and clinical. There is no mystery attached to statistical significance - it is simply
the mathematical likelihood that the result did not occur by chance. Clinical
significance, we agreed, describes what the results would mean if applied to a
population similar to that studied. But therein lay its limitation. Although clinical
significance attempts to clarify the potential impact of the research, it only applies
to populations or groups of patients. As clinicians, particularly in primary care, we
still face the difficulty of extrapolating such population-derived information to the
individual patient, who may not enter the consulting room with a discrete one
dimensional problem that can readily be turned into an answerable question
(Charlton, 1995b; Lancet, 1995; Greenhalgh, 1996).
■
Evidence-based medicine
4I
In a traditional model, new research findings are passed to the professional
community via the conventional pathways of peer-reviewed journals and clinical
meetings. The next step involves the receipt of that evidence by the wider health
community - the move from Rogers' (1983) earlier adopters to the majority and
ultimately the so-called 'laggards.' Evidence from medical research reaches
doctors and patients in different ways, but it is self-evident that doctors do not
always hear of new research findings first. Increasingly, such findings are rapidly
disseminated through the media, or are learned about through family or friends.
And patients are not passive recipients of information - they have their own ways
of interpreting and responding to new findings. The crucial step in dissemination
of clinical information occurs during the interaction between doctor and patient,
when an individual practitioner interprets and explains the clinical information
for and with an individual patient during a one-to-one consultation. This is not a
simple transaction. The term 'personal significance' helps us to understand the
components of this part of the consultation from the perspective both of the
clinician and of the person (patient) who may benefit from the information.
Although it is acknowledged that factors associated with the doctor can affect the
way in which a message is transmitted and interpreted, what really matters is
what the message means to the patient (Balint, 1957).
Once again, in this critique of statistical and clinical significance I do not intend
to diminish their vital contribution to clinical thinking. The crucial advantage that
these mathematical models constantly seek is the reproducibility of the data
produced by investigation. My aim, rather, is to debate their limitations, and in so
doing to call into question their intellectual impregnability.
Consider first the derivation of statistical significance. The accepted value for
statistical significance, P < 0.05, is based on the understanding that in any
Gaussian distribution of a continuous variable, 95% of the data are included in
a zone that is covered by 1.96 standard deviations of the mean. The 5% of data
that do not fall within this span should be regarded as inconsistent with the main
distribution. Statistical significance is the term used to indicate the probability of
clinical data falling within this span (Feinstein, 1992). Is this valuable? Indeed it
is, but let us think of the notion in the context of clinical care.
The primary reasoning behind this calculation of P does not fit easily with the
distribution of many kinds of data that are found in clinical medicine. In
addition, this type of frequentist theory, which has dominated medical research
for much of the last century, creates a dichotomy in which results are regarded
as either significant or not. Arguing in favour of a change of model towards
Bayesian statistics, Tilford and Braunholtz (1996) proposed that dichotomous
results ('either it is or it isn't') do not take into account relevant evidence
obtained outside the index experiment, which may be important in formulating
clinical policies. Out of such dichotomous reasoning emerges the paradox that
very small differences will eventually 'become' significant if repeated often
enough in a very large series. Thirdly, human factors can also affect statistical
interpretation. As we discussed earlier in the chapter, a single summary out
come statistic cannot capture satisfactorily the heterogeneity of the combined
experience of the individuals in a clinical trial. Such calculations fail to address
the issue of auxometry (the rate of progression of illness). Patients who have
the same disease at the point of entry in a trial may have different patterns of
the illness, with different rates of progression and clinical features, which can
I
42
•iniin
JJ.'.n..
II"...
Complexity in primary care
affect both treatment and prognosis (Feinstein, 1992). Thus the importance of
statistical significance lies in the early stages of the interpretation of research
findings.
Although the results of a clinical trial may be statistically significant, the
clinician is helped to interpret such evidence with the calculations that collec
tively constitute clinical significance. In general there are four key calculations,
namely absolute risk reduction, risk ratio, relative risk reduction (the comple
ment of risk ratio expressed as a percentage) and the concept of the number
needed to treat (the inverse of the absolute risk reduction) (Sackett et al., 1985).
These concepts, particularly the number needed to treat, have been of especial
value to clinicians in helping them to interpret the results of trials. However, what
remains unresolved is the precise boundary at which a distinction between two
means or two rates can be regarded as quantitatively significant - to the same
extent that the stochastic component of statistical significance lies at 0.05. In their
review, Burnand et al. (1990) noted the boundaries that were being used by
general medical journals to trigger such quantitative decisions in research that
contrasts two means, two rates or two correlation coefficients. They found that
this boundary was reached when the ratio of the smaller to the larger mean was
greater than 1.2, where the odds ratio was greater than 2.2, and where the
r-value for the correlation coefficient was greater than 0.32. However, these
conclusions are based only on the authors' reporting of the interpretative
comments of the component studies in the review - they do not represent a
consensus view.
Clinical significance is thus an important additional factor that can aid the
interpretation of research findings for regional, district or practice populations.
However, the whole aim of producing such evidence and calculating statistics is to
clarify the clinical dilemma for the patient. Thus the most important application
of such findings lies in the context of the individual consultation, during which
such information is tailored to the individual personal context of the patient.
Personal significance adds a further dimension, and is the key to the transfer of
an idea to, and the evaluation and interpretation of that idea by, the doctor and
the patient together. Personal significance is thus a dialectic consisting of a
contribution from the practitioner, who outlines the concept as he or she
understands it, and the person who receives and evaluates the new idea. At
the heart of this definition is the interaction between the two participants in a
consultation - and it is this that drives the definition towards a greater apprecia
tion of subjectivity (and indeed inter-subjectivity, the process of their inter
action). In the context of the argument that is unfolding in this book, we are close
to considering - if we accept the processes involved in personal significance - a
new way of knowing.
The contribution of the doctor is threefold, namely evaluating the research
evidence, exploring the patient's philosophy of health, and delivering an opinion
that is based on a synthesis of the two. Doctors evaluate evidence in different
ways. Usually they will have an opportunity to evaluate information independ
ently, before engaging in a consultation. But doctors are people, too - they are not
immune from fears, prejudices and attitudes to health by virtue of having
received a medical education. The evaluation process is not simply an intellectual
procedure, but includes both cognitive and intuitive components. This process
has been best described by Neighbour (1987) as an 'inner consultation' in which a
Evidence-based medicine
43
dialogue takes place between what he terms the organiser (the logical part of the
process) and the responder (the more intuitive component).
Doctors conduct an inner consultation with biomedical evidence before
deciding how to apply it. Although the doctor's organiser responds in an
analytical, logical way, the evaluation of new evidence is also influenced by
the doctor's background, experience and other individual factors. Thus the
doctor's responder will act in a more intuitive manner, using pattern recognition
and the association of ideas in a Gestalt manner. The responder is sensitive to
internal messages determined by the doctor's feelings and emotion, and this
affects the interpretation of information in a way that recognises context,
experience, apprehensions, failures and successes. This elevates and dignifies
the doctor's subjectivity.
Furthermore, the consultation occurs at one point in what may be a long
standing doctor-patient relationship. Such relationships are dynamic and ever
changing, so that the interaction between organiser and responder, and between
the product of that dialogue and the influence of the patient, evolves over time.
For the experienced doctor the logical (organising) processes become less import
ant, and are replaced by historical pattern analysis or script recognition from
exposure to previous similar problems - a responder function. This is what
distinguishes the thinking of novices from that of experts (Van der Vleuten and
Newbie, 1995), and focuses the analysis on the interaction between doctor and
patient.
The second part of the doctor's role in personal significance involves exploring
the patient's health philosophy. The clinician and patient may have different
priorities, so the traditional healthcare philosophy of the medical profession may
not be shared by an individual person. For the most part, medicine assumes that
disease-free longevity is desirable, even at the expense of matters that patients
consider to be of more immediate and substantial concern (Landau and Gustaffson, 1984). Let us recall the case of Mrs B. She was conveying a despair and
hopelessness about her predicament, to the effect that, put bluntly, she no longer
wanted to live. Consequently, she began to take decisions about the care of her
diabetes that were incompatible with expert professional advice. Mrs B is telling
us something extremely important. If patients' priorities differ from those of the
clinician, the quality of medical evidence matters little, as the clinician's advice
based upon it will be ignored.
Finally, the style and method of communication may also affect the message,
and contribute to the patient's personal significance. A doctor's own experience,
either privately or professionally, will influence the words and nuances that he or
she uses in discussion. For example, the doctor who wrongly diagnosed a benign
breast cyst that turned out to be a breast carcinoma will alter his future behaviour
and management of women who present with fibroadenosis in a way that is
coloured by this experience. It is almost impossible for doctors to be clinically
dispassionate or completely neutral about a topic - their view is a product of both
cognitive and experiential evidence. This aspect of the definition of personal
significance encourages us even further to focus on the interaction between the
two participants in the consultation. The crucial point here is that this interaction
is unpredictable - its outcome is emergent, in the sense that it cannot be known in
advance. And it is serendipitous, constantly at the whim of the frailty of the
human predicament. We shall revisit these features of personal significance in
1
44
uioii'ii
' '"t:;
itc...
:i
1
Complexity in primary care
Chapter 5, when I discuss the complexity theorist Ralph Stacey's definition of
complex responsive processes (Stacey, 2000).
However, it is the patient's contribution that is more important in creating
personal significance. In a consultation the patient adds to previous intellectual
and emotional understanding of an illness experience (Heath, 1995; Cromarty,
1996). Patients are not passive recipients waiting for doctors to make decisions
about their health. The evidence suggests that the more actively patients
participate in consultations, the better controlled are their chronic diseases
(Kaplan et al., 1989). Attitudes to health are not exclusively logical. People's
attitudes to health, and the decisions that they take, are determined by how they
perceive a particular threat to health, their belief in the advantages to be gained
from a change in behaviour to accommodate that threat, and how difficult they
believe that behaviour change to be. The beliefs that form attitudes to health are
influenced by personal and family factors, and also by social and demographic
factors (Becker, 1974). Actions that are based upon these beliefs are not always
rational - they can be emotional or habitual (Johnson, 1995). The actions that
patients take are influenced by what they think others might expect them to do,
and by how much importance they attach to those expectations (Fishbein and
Azjen, 1975). Of course this analysis applies equally to the doctor as a person, not
just as a professional. As Kant said, 'we see things not as they are, but as we are'
(Russell, 1961). This firmly roots the definition of personal significance in the
patient's personal history, personal narrative and personal epistemology. The
subjectivity of the interaction has become the major feature of the definition.
Summary
The contemporary explanatory model in medicine boasts a long and illustrious
pedigree, which was outlined in the previous chapter. From that tradition has
emerged the model of EBM, which has contributed directly and extensively to
improvements in patient care, and which has anchored critical appraisal of
evidence firmly in the clinical arena. In this chapter we have reflected generically
on the criticisms of the model. By considering the individual criticisms, we have
been encouraged to interrogate the nature of diagnosis in clinical practice, the
properties of a randomised controlled trial, the type of knowledge produced by
that process, and the conventions (statistical and clinical significance) whereby a
value can be attached to such knowledge. I argue that this process has helped to
define the limitations of these two conventional levels of significance, and
encouraged, by way of a response, the introduction of a third level of significance.
At stake in the definition of that third level - personal significance - is the
centrality of subjectivity and interaction, and of emergence and serendipity, in
the clinical encounter. We are forced to consider that this represents a different
'way of knowing.' It is now appropriate to explore the intellectual pedigree of this
'way of knowing', by tracing the historical antecedents of a constructivist,
naturalistic way of understanding the world. Although these terms might be
unfamiliar to the jobbing clinician (and indeed to your humble author), they
might be more easily recognised as the building blocks for the principles of
qualitative research. This subject is explored in the next chapter.
.f
g
1
s
s
s
s
/
1
/
Chapter 4
The naturalistic tradition: historical
overview of the epistemological
origins of qualitative research
Introduction
t
1
t
This chapter describes the evolution of a second intellectual tradition, described as
naturalistic. This tradition is characterised by a search for truth predicated on
scepticism, a constant comparison of the natural and the social world, and the
emergence of a socially constructed ontology. The evolution of the naturalistic
tradition has led to the development of sociology as a legitimate domain of
enquiry, and of qualitative research methods as appropriate tools for exploring
the meaning of human actions. The history of thinking from the seventeenth
century to the present day illustrates how philosophers' struggle with the notion
of truth' has influenced the development of these qualitative research methods.
The purpose of this chapter is to encourage us to compare this naturalistic
tradition with the development of the scientific tradition set out in the previous
chapter, and to prepare us to consider a third intellectual tradition, predicated on
non-linearity, which we shall explore in the next chapter. The key idea to bear in
mind is the relationship between the way we explain things, the type of
knowledge we consequently prefer to create in order to furnish those explana ions, and a fundamental world-view which underpins that way of thinking
Putting those thoughts in more philosophical language, I argue that these three
traditions (scientific, naturalistic and non-linear) are predicated on different but
complementary ontologies. Their comparison reveals how each has created its
own epistemological framework with which to populate that world-view in the
iorm of an explanatory model.
Historical development of qualitative research methods
It IS convenient, following the proposals of Auguste Comte (1798-1857) in his
Course of Positive Philosophy (Comte, 1875), to consider the history of a naturalistic
tradition in three periods.
The first era covers roughly the seventeenth to the nineteenth centuries.
During this period the natural and social sciences were indistinguishable, and
every form of scientific enquiry was regarded as philosophical (Dunn et al.,
2). This period includes the work of Montaigne and Bacon, sets out the
basis of Cartesian duality and doubt, and includes the contributions of
Berkeley and Hume.
i
j
j
j
46
Complexity in primary care
• The second period includes the writing of Kant, Comte, Weber and Mill.
During this period positivism was developed, and the hierarchy of the sciences
was described, placing the physical sciences at the pinnacle of knowledge.
• The third period covers the twentieth century. Tensions between qualitative
and quantitative approaches arose and then matured from a position of
opposition to one of greater collaboration (Rossman, 1985).
The first period: the seventeenth century - a point of departure
lllIWU'HI
"I!:.;..
„i.....
!i: ■
The seventeenth century, regarded by some as the beginning of the modern
world, can be taken as a convenient starting point for this historical overview of
qualitative research (Grbich, 1999). At this time, the dominant intellectual
authority lay with the Roman Catholic Church. Although the Church approved
scholarship, it only afforded seniority and influence to scholars who did not
challenge its authority. As Hawthorn (1976) points out, it resisted any rise in a
rational form of human enquiry - this was really a political dislike of challenge to
its static epistemological position of an unchanged and unchanging Creation. The
church's assertion of the privileged position accorded to divine final cause would
have been fatally weakened by the rise of a science that revealed 'divine' purpose
in what we would now describe as a naturalistic way. The notion of absolute truth
dominated (Crosby, 1997), and was strategically linked to the concept of power.
However, the Roman Catholic Church could not stifle the frustration caused by
its inflexible position, nor could it eliminate the inexorable rise in scepticism that
characterised subsequent Renaissance science. And it is with Renaissance science
that a more detailed description of the history of qualitative research can begin
(Murphy et al., 1998). The names of the main contributors during this period, and
their principal contributions in the context of this chapter, are presented in Table
4.1. Although this may be an egregious conflation of their contribution to human
knowledge, its purpose is to clarify their relevance to the propositions explored in
this book.
Table 4.1 Key figures during the first period
Philosopher
Dates
Principal contribution in the context of this chapter
Montaigne
Francis Bacon
1533-1592
1561-1626
Thomas Hobbes
1588-1679
Rene Descartes
1596-1650
Development of sceptical approach to knowledge
The notion of induction, permitting generalities to be
accumulated from particular observations. Embryo of the
scientific method
Hyperbolic doubt. How do we know how accurate our
observations are? Embryo of social construction
Cartesian duality and Cartesian doubt. Application of
mechanical and mathematical principles to natural
phenomena
Immaterialism and further origins of socially constructed
ideas of knowledge
The nature of uncertain knowledge, our knowledge of the
future, and of causality
The integration of the natural and social sciences (or
natural and moral philosophy). Combination of research
methods in the Wealth of Nations
George Berkeley 1685-1735
I
David Hume
1711-1776
Adam Smith
1723-1790
The naturalistic tradition
47
The Renaissance and the Reformation
An increase in scepticism characterised intellectual progress during the Renais
sance and Reformation, and is epitomised by the work of Montaigne (1533-1592)
and Bacon (1561-1626). Bacon exhorted scientists to undertake a completely
'fresh' examination of particulars in an orderly and considered manner. 'There
remains simple experience which, if taken as it comes, is called accident', he
wrote. 'But if sought for experiment, the true method commences with experi
ence truly ordered and digested' (Bacon, 1858). Montaigne, on the other hand,
was more explicitly sceptical. 'Que sais-je?' was his personal motto (Burke, 1981).
The notion of experimentation was much broader at this time, and could refer
to any occasion during which one tested out the validity of a proposition - this
could include what we would regard as natural experiments in contemporary
qualitative research (Murphy et al., 1998). Bacon advocated induction. He
suggested that one could draw up particularities of observation from experience
and experiments, and combine them to create lesser, then greater generalities
(Tuck, 1993). Although Bacon illustrated his methods mostly by reference to
physical phenomena, he advocated that a sceptical approach should also be taken
to the study of history, literature and philosophy. The historian Lester King
(1982) asserts that it is in Bacon that we find the principal elements of the
modern scientific method, as he recognised the need for controls, verification and
reduction of bias (Greaves, 1996).
Hyperbolic doubt, Cartesian doubt and duality
The philosophers who followed Bacon, notably Hobbes (1588-1679), who was in
fact a younger contemporary, developed scepticism in a much more fundamental
way, focusing on the very act of observation itself. The radical nature of their
scepticism came to be known as hyperbolic doubt (Tuck, 1993).
The central core of hyperbolic doubt concerns the nature of observation. How
can we know that our observations are accurate reflections of the external
world? These questions arose in the seventeenth century mainly as a result of
developments in optics, which at that time strayed into areas which we would
probably now regard as the domain of psychology, and Hobbes himself was
much inspired by Galileo (Russell, 1961). Seventeenth-century advances in
optics encouraged scientists to define more precisely and at the same time to
doubt the veracity of perceptions. Indeed the use of mirrors and lenses was
sometimes associated with intrigue and falsehood (Hurwitz B, personal com
munication). Although these advances might allow individuals to make observ
ations that would be impossible with the naked eye, that same eye could easily
be tricked by their misuse. So how could accurate and critical observations be
distinguished from unreliable ones?
Philosophers in the seventeenth century regarded this issue as a practical
problem rather than an ontological one. They proposed, as a way of closing the
logical gap between sensory observation and interpretation, that observations
should be regarded as signs of the world. In doing so, they still relied on divine
inspiration, arguing that God would not play tricks on his creatures, and that at
the very least a thinking body could be certain of its own existence and the
correctness of immediate sensations. In the end we could trust our perceptions,
they argued.
48
iniiii
>’• I
C
Complexity in primary care
This is encapsulated in the well-known dictum, cogito ergo sum, of Descartes
(1596-1650). Descartes epitomised the scepticism which flourished at that time in
his explication of Cartesian doubt (Russell, 1961). The irreducibility of the T led
Descartes to assert that the soul (mind) was different from the body - this
Cartesian duality would influence medical thinking for the next 200 years.
Descartes' influence on the subsequent development of epistemologies is central.
His notion of truth was predicated on a view of the world as linked to but separate
from subjective interpretations, emotions, reflection and consciousness (Des
cartes, 1912). The reality constructed from such truths was measurable and
controllable - by using logic and mathematics one could make accurate predic
tions of future events (Crosby, 1997). In this lies the kernel of the scientific
method and subsequent positivist epistemology.
On the other hand, the importance of hyperbolic doubt was that it created a
rather different notion of how scientific enquiry should proceed. Rather than
seeing and understanding the world by inferring the rules that may govern it,
science began to regard the world as the creation of our observations, rather than
regarding our observations as being a copy of it. Our own observations and
interpretations, then, could be a proper focus for enquiry - and for some, the
origins of social construction.
The approach developed by Hobbes and others arose from their concern about
Bacon's reliance on induction (Murphy et al., 1998). Induction is the process of
proceeding from a list of singular statements to a universal statement. It is defined
by Chalmers (1982) as follows: 'If a large number of A's have been observed
under a variety of conditions, and if all those observed A's without exception
possessed the property B, then all A's have the property B.' Oilman (1973)
describes induction as occurring 'where we reason from a piece of information,
however complex or elaborate this may be, to a conclusion which is logically
independent of it.' By contrast, he argues, deduction occurs when 'the relation
between premise and conclusion, by virtue of which I am justified in inferring the
latter from the former, is internal and can be gathered from the premise and
conclusion alone. What the conclusion states is already contained in the premise'
(Dilman, 1973).
Collingwood (1946) points out that Bacon asserted that one of his contributions
was to develop induction as a break from the dominance of deductive reasoning
which had characterised most historical writing up until then. It was such
deductive reasoning, for example, that allowed medieval history to be viewed
as illustrations of the Divine Plan. Such revelations contained useful (divine)
knowledge, because its premises determined its conclusions (Collingwood, 1946).
However, the sceptics were arguing about the assumptions which underpinned
the very particular observations that constituted the cornerstone of the inductive
method.
Immaterialism: the origins of social constructionism?
The approach to philosophical enquiry predicated upon profound scepticism was
subsequently developed by Bishop George Berkeley (1685-1735), and it is to him
that Murphy et al. (1998) and Bloor (1976) attribute much of the foundations of
contemporary qualitative research. Berkeley is important in philosophy because
of his denial of the existence of matter - that is, immaterialism. His principles,
which were all set out when he was relatively young, are best seen in his Dialogues
F
The naturalistic tradition
49
Of Hylas and Philonous. which was written in 1713 (Berkeley, 1967). Hylas, who
represents the scientifically educated common-sense person, debates with Philo
nous (effectively Berkeley himself) about the nature of material substance 'Can
there be anything so repugnant to common sense than not to believe in matter?’,
asks Hylas. Philonous does not deny the existence of sensible things, but asserts
that one does not see the causes of colours, or hear the causes of sounds. The reality
of sensible things consists, he says, of being perceived (Russell 1961) His
philosophy was irreverently captured in the following limerick, and reply by
Monsignor Ronald Knox:
There was a young man who said God
Must think it exceedingly odd
If he finds that this tree
Continues to be
When there's no one about in the quad.
Reply:
<
Dear Sir:
Your astonishment's odd: I am always about in the quad.
And that’s why the tree
Will continue to be,
Since observed by,
Yours faithfully,
God.
(Quoted in Russell, 1961)
It was unnecessary, Berkeley argued, to propose the idea of a separate world
composed of physical matter, because what was much more important was the
way in which human beings classified and organised the world, and acted
according to those classifications. While insisting that he started from a
common-sense view of the world, Berkeley recognised a number of objections
to his position. First, the most obvious objection was the idea that objects would
cease to exist if we did not attend to them - the main message in the limerick
quoted above. Secondly, if material objects were irrelevant, then the world as we
know it would have the same attributes or lack of attributes as dreams, illusions
and fantasies. Finally, there is a seemingly obvious distinction between a real
blow and an imagined blow, between the imagining of being struck by a blow and
the actual experience of it. Berkeley (1967) responded to these arguments by
stating that ideas can be distinguished from each other by their having a different
nature or order, such that the ideas of reality are stronger and exhibit some
regularity. The quality of ideas also allows us to establish procedures for counting
or deciding what is real (Russell, 1961).
Within the context of this chapter, Berkeley's influence on qualitative research
*os in this rather arcane notion of immaterialism. From Berkeley's perspective,
what we know is the way in which we construct the world - there is nothing
knowable other than minds and their contents (Berkeley, 1967). This does seem
recognisable in the notion of social constructionism upon which much qualitative
research rests. Berkeley starts from a common-sense view of the world, resisting
re introduction of theoretical assumptions or technical notions. According to
50
ii-
ri
t:
Complexity in primary care
Collingwood (1946), Berkeley was at pains not to demean science by this route.
His main point was not to make the mistake of assuming that science offered
absolute truths. New scientific hypotheses that were more predictive than their
antecedent hypotheses were indeed more powerful, but despite their predictive
capacity, they were always vulnerable to alternative descriptions and more
accurate results. Many contemporary qualitative researchers, drawing on
Berkeley's influence, continue to adhere to the view that it is entirely appropriate
to study the world from the perspective of social construction, including a social
construction of the world created by science (Chalmers, 1982).
The application of Berkeley's approach in recent research in medicine is
illustrated by a paper by Bloor (1976) on decision making among ENT surgeons.
Bloor observed 11 ENT surgeons in various clinic settings, gradually drawing
together a picture of the various processes and procedures that the surgeons used
to decide whether a patient who had been referred should be listed for
adenotonsillectomy. He analysed the data inductively, identifying the way in
which each surgeon took their clinical histories and made decisions. Both of these
activities, history taking and decision making, differed among the 11 specialists.
The data suggested that the surgeons attributed different weights to the same
physical findings. From Berkeley's perspective, the surgeons were socially
constructing their decisions in this context in a way that reflected their experi
ence of the domain.
Causality
Although the main focus of Berkeley's philosophy was on immaterialism, much
less prominence was given to the notion of causality, which Berkeley formulated
merely as a linkage between two repeatedly observed phenomena. It was David
Hume (1711-1776) who refined the notion of causality during the eighteenth
century (Russell, 1961). In his Treatise on Human Nature, Hume (1739) identified
two kinds of statement. Drawing on algebra and geometry, the first statement,
which he termed impressions, concerned statements that can be shown to be
demonstrably true because the conclusions are inherent in the ideas themselves.
An example is the proposition from Euclidean geometry that the square of the
hypotenuse is equal to the sum of the squares of the opposite two sides.
Contained in the section of his treatise entitled Of Knowledge and Probability, the
second kind of statement from Hume's work related to all knowledge deriving
from empirical data that cannot be directly demonstrated. Thus it includes all our
knowledge about the future (Hume, 1975). His discussion here has real relevance
for current research in medicine and for qualitative and social science research, as
it explores the notion of predictability. For Hume, the notion that the sun would
not rise tomorrow was as logically intelligible as the statement 'The sun will rise
tomorrow.' However, we disbelieve one and firmly believe the other. Hume's
description of causation had three components, namely contiguity in time and
space, priority in time (the cause preceding the effect), and a necessary con
nection between the two (Hume, 1739). Essentially, Hume was arguing that we
observe regular conjunctions, and it is the regularity of these conjunctions that
creates the notion of necessity about any connection between a regularly
conjoined antecedent and subsequent event. This is falling short of the notion
of absolute causation. Hume's answer was to say that although we cannot be sure
L
The naturalistic tradition
5I
of the laws of physics, nature continues in such a uniform way that
we can treat
conjunctions as causal for practical purposes (Hume 1739)
In the context of contemporary medicine, Hume might ask how we can predict
that the effect of a treatment which has worked well in many thousands o
prevtous pattents wtll work well in a particular patient with the same condition
The regularity of the connection between the treatment and subsequent bene-’
^“e ir^^theT t " T
abS°IUte
^nefit in a
particular case in the future. For contemporary qualitative research, Hume's
legacy also lies in his integration of the natural and social sciences. The first social
scientists - Smith and Ferguson in Scotland, and Montesquieu and Vico in France
S™ to TXoV We7 bea/^ influen-d
fume's support for induction
(Schneider, 1967) One could induce generalisation or regularities about society
Hume argued, only if they were based on extensive observations
putotoTn1 1 tT
inZ
°fthe Wealth
first
published in 1776, uses a wide range of methods to present its conclusions
Th ocu"”n's' m"'“
“d
=
ations (Smith, 1976, 1993). These two notions, namely Berkeley's immaterialism
and Hume s constant conjunction, heavily influenced the way in which qual-
,a“d M“
s
“ “d
»>'
I
The second period: positivism and the hierarchy of the sciences
™e second period covered in this analysts includes the work of Kant Comte
Weber and Mtll. During this period positivism was developed and the hierarchy of
he sctences was described, placing the physical sciences at the Znade of
in tlw comew ofnth?sechf
h1"” C°ntributOrS and their Pr"’^Pal contributions
m me context of this chapter are presented in Table 4.2.
Table 4.2 Key figures in the second period
Philosopher
Dates
Immanuel Kant
1724-1804 Observations were theory driven. Exploration of the
balance between experience and systems of thought
(phenomenon and noumenon), reflected in the balance
between data and theory in contemporary research
1798-1857 The introduction of positivism. Presenting the sciences in a
hierarchy, with the physical sciences at the pinnacle
1806-1873 Exploring the notion of positivism introduced by the
French philosopher Auguste Comte. Embedding the
process of induction in the social sciences
1864-1920 Rejection of the notion that quantification and
measurement were the only proper tools for a scientific
enquiry (the word 'science' meaning rigorous and precise)
Auguste Comte
John Stuart Mill
Max Weber
Principal contribution in the context of this chapter
Kant and the emergence of social science
onTr8 thf 'S,SUeS that PhlIosoPhers were grappling with at this time was the
quest,on of what
titutes reliable knowIedge _ as important a notion nQw for
emporary medical research as it was then for Immanuel Kant (1724-1804)
A
!7
52
IE'"'
K- •'
Complexity in primary care
and John Stuart Mill (1806-1873), two of the philosophers who were most
directly influenced by Hume and Berkeley. Mill in particular explored the theme
of how people could grasp reliable knowledge. He developed both the notion of
induction and the notion of causal laws. In the context of this chapter, the
importance of his contribution lies in applying both of these to the fields of
psychology and economics, as well as sociology and history. Perhaps more
importantly, the rather difficult writings of Kant developed thinking from
where Hume and Berkeley left off. It is important to spend some time considering
Kant's main points because of their influence on German and American qual
itative social science in the nineteenth and twentieth centuries.
Kant rejected the view that our knowledge of the world is founded simply and
exclusively on our experience of it. Although experience provided the contents of
knowledge, Kant argued in Critique of Pure Reason, reason was also necessary to
provide a structure or order. While the outer world causes only the matter of
sensation, our own mental faculties order this matter in space and time, and
supply the concepts by which we understand experience (Russell, 1961). He
distinguished between analytic and synthetic propositions, and between empiri
cal and a priori propositions. An analytic proposition is one in which the predicate
is part of the subject — for example, a fat man is a man. A synthetic proposition is
one that is not analytic, and so includes everything we can know through
experience. An empirical proposition is one that we can know only by sense
perception, either directly or via reliable testimony — for example, the facts of
history and geography. An a priori proposition is one that, although elicited by
experience, is seen to have a basis other than experience - all the laws of
mathematics are in this sense a priori (Russell, 1961).
In this sense, Kant's thinking influenced contemporary commentators on the
role of science, fuelling the debate about theory-driven observations, elegantly
explored by Chalmers (1982). Kant argued that we cannot usefully and produc
tively focus our gaze on the world unless we have a prior theory which directs our
observations. The form of knowledge that derives from such an activity is
objective in the sense of being independent of the observer. Kant applied the
term 'transcendent idealism' to this idea to suggest its ability to overcome the
notion that knowledge is derived purely from (sensory) ideas of the world.
Although we draw knowledge from our sensations, Kant argued, we also have
knowledge of theories and concepts which organise the way in which we
experience these sensations. Scruton (1982), in a commentary on this work,
writes: 'A mind without concepts would have no capacity to think; equally, a
mind armed with concepts but with no sensory data to which they could be
applied would have nothing to think about' (his emphasis).
In Kant's work, theory and data are given equal status. Kant ascribed
distinguishing terms to these two notions. A phenomenon is what appears to us
in perception, and it consists of two parts, namely the sensation, and the part due
to our subjective apparatus, which allows us to order or classify the perception.
Noumenon, in Kant's taxonomy, refers more to systems of thought, a mechanism
by which ideas can be classified and not something that can be directly
experienced (Russell, 1961). Murphy et al. (1998) argue that this taxonomy
had a major influence on subsequent social theory. By the 1870s, this kind of
epistemological debate was influencing the methods and practice of the social
scientists (Adorno et al., 1976). In contemporary terms, this debate is expressed in
The naturalistic tradition
53
the arguments about whether the social sciences are informed by abstract or
empirical knowledge.
Addressing this issue, the German economist Menger identified two types of
knowledge, m a classification broadly similar to that of Kant, namely a knowledge
of concrete instances and a knowledge of forms or types. Types can be experi
enced in individual form, which allows us to characterise them. However it is our
understanding of typicality that makes prediction possible. Menger, who was an
economist was able to distinguish between theoretical economics, an analysis of
type, and historical economics, concentrating on individual cases.
At the turn of the nineteenth century, debate about this distinction was aided
by considering the idea that disciplines could be categorised to reflect the
distinction between type and case - between what came to be called nomothetic
disciplines (concerned with laws and generalisations) and ideographic disciplines
(concerned with specific instances) (Freund, 1968). For example, a nomothetic
study of society could be part of a natural science concerned with looking at the
regu anties m institutions and individuals. An ideographic study of the same
society could generate a cultural science, concerning itself with questions about
what leads us to think about that society in a particular way. This is an important
distinction to bear in mind when reflecting on contemporary research. The kind
of evidence that is produced by the randomised trial is in the form of generalisations - more nomothetic than ideographic. This leaves the clinician with the
i eographic challenge of applying such nomothetic evidence in individual
ratZ thaOnS’ r’u ValUe °f "0"sidering ,he taxonomy as complementary,
rather than as dichotomous, will be explored later.
The new idea of a cultural, ideographic science challenged scholars to think
about just how precise, unique context-specific evaluations could be produced It
was at this point that 'reflexivity' - so central to contemporary qualitative
research was introduced. Dilthey described the notion of 'verstehen' which for
he first time set out the contribution of the subjectivity of the analyst, which
reflected their own experience and led them to a particular perspective or
description of the experience being studied (Hughes, 1959). The connection
between this debate and current research was secured through the work of
ead and the Chicago school, which gave rise to symbolic interactionism - for
r5SKearchfrS the bedrock for many qualitative approaches (Holloway,
). he debate also had practical implications for how the various disciplines
that were emerging under the broad umbrella of social sciences organised and
P01i?icsC d ■' e,mselves and their approach to study. Sociology, psychology and
politics divided into two camps - positivist and normative. The positivist school
saw themselves as scientists, conducting value-free enquiries, whereas the
normative school offered descriptions that raised questions about how things
snould work.
&
Weber and the development of sociology
Max Weber (1864-1920) is perhaps the most prominent example of the implica
tions of such separation. 'The object of study for sociology,' he argued 'is the
scienttfic mvestigation of the general cultural significance of the socio-economic
structure of a human community' (Weber, 1963). Although values of culture
could also be a proper route of enquiry, they had a non-rational foundation and
consequently a scientific study could not address their validity However
54
4
Complexity in primary care
sociology had a key contribution to make by systematically describing discrepan
cies between professed values and observed actions. According to Murphy et al.
(1998), Weber had a profound and lasting influence on the qualitative methods
currently in use, and his arguments have been very influential in legitimising
qualitative research. One important contribution that he made was to reject the
notion that quantification and measurement were the only tools available to a
proper scientific enquiry (Freund, 1968; Aron, 1970). Weber himself used
quantification extensively, but argued equally that coherent and logically oiganised systems of concepts could produce a firm enough account of a domain to
form the basis for clear interpretation and effective action. Echoes of this insight
are seen today. For example, they can be found in an editorial by Sackett et al.
(1996) allocating quantitative and qualitative research their appropriate places in
evidence-based medicine. Even in case studies, Weber used systematic descrip
tions of unique events in order to elicit some regularities from those events, if not
precise laws in the more strict scientific sense. Interpretative understanding,
Weber argued, was one means of developing a general science of society (Freund,
1968).
S'
Mill, positivism and the laws of induction
Two years after Kant's death saw the birth of John Stuart Mill (1806-1873),
whose philosophy was to develop further some of the arguments laid out by
Hume, Berkeley and indeed Kant himself. It is in the work of Mill that we see the
first application of the idea of positivism that was introduced by the French
philosopher Auguste Comte (1798-1857) in his Positive Philosophy. Comte (who
also coined the term 'sociology') emphasised the belief that natural science was
the most important paradigm and the source of all possible valid knowledge. He
expressed the view, typical of that held in the nineteenth century, that the
sciences could be seen as a hierarchy, with the basic natural sciences, physics and
chemistry at the pinnacle (and in that order), and with those sciences more
closely related to the behaviour of societies, institutions and individuals at the
bottom, each tier being constrained by the laws of the ones above. Comte's work
is considered to be the classical expression of the positivist view, namely that the
empirical sciences are the only valid source of knowledge (Comte, 1993). It is
important to remember that Comte was developing and publishing these ideas
just at the time when Parisian medicine was revisiting many of its institutions and
policies, focusing on the importance of precise systematic observations, correlat
ing anatomical findings with clinical presentations and, to all intents and
purposes, as a discipline practising the positivist philosophy that he espoused.
For the purposes of this chapter, it is the relationship of the sciences in the
hierarchy which is important, as it clearly distinguishes in terms of value between
those at the top (the natural sciences) and those at the bottom (studies of human
relationships and societies).
Murphy et al. (1998) argued that positivism, initially described by Comte,
developed two distinct strands when used by philosophers on the one hand and
by social scientists on the other. For the former it emphasised phenomenalism,
which recognised experience as the basis of valid knowledge. For the social
sciences it had three implications, all of which retained the notion of the natural
sciences as the dominant paradigm (Giddens, 1974). First, there was the implica
tion that the processes and methods of natural science could be adapted to the
(
V
0
s
s
c
1
c
t
(
I
(
(
The naturalistic tradition
/.
s
g
e
a
i
)
t
t
55
social sciences. Secondly, the application of positivism in the social sciences was
was
value-free activity. Thirdly, as a result of the first two implications, the outcomes
of a social enquiry would have the same characteristics as those of a natural
scientific enquiry. Throughout the twentieth century, positivism has dominated
social and scientific research. This was a system of enquiry that recognised only
observable phenomena, objective relationships, and the laws governing them
The belief that the method of logic upon which the physical sciences were
constructed could be applied to the social sciences became one of the central
tenets of positivism (Grbich, 1999).
However, according to both Giddens (1974) and Murphy et al. (1998), the
application of positivism to the social sciences has been threatened by accumulat
ing evidence that the social world is different from the natural world, and the
emerging argument that human action is better understood as a creative act of
rule orienting, rather than the more deterministic notion of rule following.
Central to the notion of positivism was the idea of objectivity, and the relation
ship between cause and effect. Mill was one of the first to describe the reaction of
effects upon causes: 'the circumstances in which mankind are placed, operating
according to their own laws and to the laws of human nature, form the characters
of the human beings; but the human beings in their turn mould and shape the
circumstances for themselves and for those who come after them' (Mill, 1974)
Mill's great contribution was to embed the notion of induction firmly in the social
sciences. His sequence - observe, induce, formulate, deduce, hypothesise, test and
observe - is his consistent legacy (Fletcher, 1971). Although this has clear
parallels with the natural sciences, exactness in the social sciences was always
going to be more difficult to achieve, simply because of the inherent complexity of
the phenomena involved.
5
c:
The third period: qualitative research in the twentieth century
.1
The historical origins of contemporary qualitative methods are to be found in the
developments in social anthropology that occurred in the years between the two
World Wars. Before Malinowski, who really introduced the notion of participant
observation in his Argonauts of the Western Pacific (Malinowski, 1922), anthro
pological accounts, mainly from travellers, had lacked context (Urry, 1993). At
the turn of the century Haddon introduced and refined fieldwork, taking up the
challenge of going directly to traditional societies in person and questioning key
informants individually via an interpreter. Fieldwork was further developed by
Haddon's colleague Rivers in anthropological studies in Australia and elsewhere,
and by Rivers' student Radcliffe Brown (Stalking, 1995).
Although it was Malinowski who did most to develop participant observation
in qualitative research, in doing so he still espoused the importance of 'scientific
values': 'The results of scientific research in any branch of learning ought to be
presented in a manner absolutely candid and above board' (Malinowski, 1922).
Some of these emphases are still important today. He distinguished between data
obtained from direct observation, data received indirectly through an interpreter,
and inferences drawn by a researcher in a summary report (Malinowski, 1922)'
He also emphasised context, advocating that descriptions of typical events should
be accompanied by accounts of the ways of thinking and feeling about those
events among the participants. For Malinowski, the goal was 'to grasp the native's
5^
56
IMIX"'
Ji"
I
Complexity in primary care
point of view, his relation to life, to realise his vision of his world' (Malinowski,
1922) (his emphasis).
In the 1930s, Malinowski endorsed what became known as mass observation
(Mass Observation, 1938). This was a movement which sought to re-democratise
politics by asking participants, who were spread nationwide throughout the
British Isles, to keep diaries in which they recorded their own daily experiences.
It was Malinowski's contribution to the mass observation report that addressed
the still delicate tension between subjectivity and objectivity in qualitative
research. On the one hand, there is the unique and personal experience of
acting in a human society. Social scientists, on the other hand, attempt to observe
and record human actions as a clue to these inner processes, document their
findings and produce data - in much the same way as science did in natural
experiments. In this way, Malinowski said, the subjective behaviour of human
beings could become the objective data of the social scientist.
Harrison extended this debate further in The future of sociology (Harrison, 1947).
He imagined a spectrum of observation, with at one end a philosophical approach
which, in his own words, produced laws without observation, and at the other an
absorption with quantitative methods which simply satisfied mathematical
criteria (Harrison, 1947). Sociology was a potential mediator between these two
extreme positions, and could become the 'anthropology of civilised societies'
(Harrison, 1947). Harrison took the notion of precise methodologies forward by
listing a variety of different types of method, ranging from interviews to
observation and what he called penetration, or observation in private settings.
The Mass Observation movement, which had been the exemplar of qualitative
research in Britain, became increasingly influenced by Government contracts,
producing the first qualitative medical sociology report - Meet Yourself at the
Doctor's - in 1949 (Mass Observation, 1949). Further developments in qualitative
methods and social science were delayed until the 1960s, when the incumbent
Labour Government demanded more and better social research to inform policy
making (Murphy et al., 1998).
Qualitative-quantitative tensions in the twentieth century
At the beginning of the twentieth century, the debates about what constituted
'truth', which the historical description in this section has addressed, focused on
two basic approaches to research - qualitative and quantitative. Quantitative
researchers assume a singular ontology that is objective, independent and
measurable (Brown, 2001). Their epistemological approach, deriving from this,
proposes that measurable influences (termed independent variables) affect out
comes (dependent variables) proportionately, as cause and effect. Precise rela
tionships between phenomena can be described by distilling raw numerical data
using the conventions of statistics.
Qualitative researchers do not, on the whole, assume a singular ontology. For
most of them the notion of 'truth' is not absolute, but rather it resides in gaining
an understanding of an individual's frame of reference, by acquiring a detailed
knowledge of their views, attitudes and beliefs. The frame of reference thus
described is recognised as being socially and historically constructed, and
influenced by passing through the researcher's interpretive prism (which is
itself socially and historically located) (Denzin and Lincoln, 1994). However,
there is considerable variation among qualitative researchers with regard to how
The naturalistic tradition
i.
n
e
e
i
d
e
>f
e
r
il
1
1
1
7
)
57
one can gain access to 'truth', and it is accepted within qualitative research that
one can both create and test hypotheses, and use statistics, to describe the
relationship between variables (Grbich, 1999).
Smith and Heshusius (1986) identify three stages in the debate between these
two approaches to describing truth, which they term conflict, detente and co
operation. They are each identified with a particular view about the approach to
research, which can be described as purist, situationalist or pragmatist (Rossman,
1985). Conflict, Smith and Heshusius argue, was the position for virtually threequarters of the twentieth century. Characterised by a purist approach to research,
qualitative and quantitative techniques were considered to derive from quite
different theoretical positions, divided fundamentally by their respective notions
of objectivity and subjectivity. However, the last quarter of the twentieth century
saw a gradual thawing of this position, with an acceptance that the methodologies
could act in parallel. Philosophical differences became submerged in the detente
of comparability. This then evolved into a more active co-operation between the
two approaches, in which some of the methods of quantitative research could
enhance the rigour of qualitative techniques. In the pragmatism of compatibility,
the epistemological differences seemed to have become obscured (Grbich, 1999).
But does this desire for pragmatic collaboration between the two approaches
work? Several studies in the 1990s have attempted to combine both of these
methodologies, either to seek convergence of data, or to produce a fuller explan
ation of a phenomenon. Although a study by Pradilla (1992) of students'
perceptions of their academic supervisors produced broadly consistent results in
its quantitative and qualitative arms (and was thus declared to have achieved
convergent triangulation), a study by Prein (1992) in the same year did not. That
study investigated the links between women's professional careers and their
private family biographies. The qualitative approach, using mainly interviews,
identified family as the most important factor influencing decisions. However, the
quantitative arm, using mainly cluster analysis, identified the particular profes
sion as the dominant influence on decisions. The researchers declared the results
completely contradictory (Prein, 1992).
In medicine, the last decade of the twentieth century did see a gradually
increasing acceptance of qualitative research processes in a field that had
previously been dominated exclusively by quantitative research. Some medical
institutions, including grant-awarding bodies, tried to draw up guidelines for
qualitative research, in a move that could be interpreted both as legitimising and,
at the same time, as constraining the boundaries of such research. Although this
has been a welcome development at an institutional level, qualitative researchers
in the field still report serious problems with having their research methods
understood let alone accepted by senior medical professionals (Sweeney G, 2002,
personal communication).
Within the last decade, this debate has been further complicated by an
acceptance of the possible limitations of the explanatory model in medicine
(Kernick and Sweeney, 2001), and calls for an enquiry into the potential
advantages of an explanatory model, predicated on non-linear change, known
as complexity, which will be described in detail in the next chapter (Pisek, 2000;
Sweeney and Kernick, 2002). The need to consider a revision of medicine's
explanatory model, which these papers called for, was heavily influenced by the
implications of Heisenberg's uncertainty principle for the notion of accuracy and
5:
58
Complexity in primary care
precision (Sweeney and Griffiths, 2002). At the risk of oversimplifying some
complicated mathematics, Heisenberg asserted that the very act of measurement
influenced the system that was being measured (Cohen and Stewart, 1994). His
quantum state - that is, the combination of velocity and position - could only be
approximated. The mechanism of observation determined the observability of a
phenomenon, collapsing its other potentialities. This is the basis of the Copenha
gen agreement (Cohen and Stewart, 1994). On the face of it, this understanding
confronts scientific determinism as applied to social and some physical phenom
ena, and its implications are still being worked out (Feinstein, 2002; Sweeney and
Kernick, 2002). In the second half of the twentieth century, advances in the
power of computers led researchers, who were modelling biological systems with
increasing accuracy, to recognise the extreme sensitivity of systems to their initial
conditions. An understanding began to emerge, which we shall explore shortly,
that it was important to understand not just what the structure of a system was,
but also how those structural elements related to and interacted with each other
(Sweeney and Griffiths, 2002). A reductionist approach to understanding how
any system worked was necessary but in itself insufficient (Evans and Sweeney,
1998). While these debates continue, what they seem to imply is that the
assumption that any research can capture absolute truth has been seriously
undermined.
Summary
Within the naturalistic tradition one discerns the view that, initially, an under
standing of the nature of human action could be determined by adopting the
same approach as that deployed to acquire scientific knowledge. The origins of
this trend can be found in Bacon's writings, and were developed through
Descartes' duality and Kant's description of causality. They reached a pinnacle
in the hierarchy of knowledge, described by Comte, in which scientific know
ledge was placed at the top. Running parallel with, but as a counterpoint to, this
hierarchy was the view that the inherent nature of social action is more complex
and less predictable than scientific knowledge. The theoretical origins of this
viewpoint are found in Berkeley's immaterialism and its relationship to a socially
constructed theory of knowledge. Mill's ideas about the impact of effect on
antecedent causes represent an important milestone in the development of this
trend, which was expressed at the beginning of the twentieth century by Weber's
acceptance of the non-rational nature of enquiry into human communities, and
his abandonment of quantification as the sole mechanism for proper enquiry. The
product of the tension between these two trends was the distinction, much fretted
over in the second half of the twentieth century, between quantitative and
qualitative approaches. For some - for example, Grbich (1999) - this tension was
dissipated in detente, and then methodological co-operation, towards the end of
the last century.
Let us reflect on the two traditions discussed so far, namely the scientific
tradition from which the contemporary explanatory model in medicine arose,
and the naturalistic tradition, providing a different model for observing, recording
and interpreting the human predicament. In the previous chapter, exploration of
the explanatory model revealed its basis in a particular ontology (a singular
reality) and a related epistemology, based upon empiricism and verificationism.
1
c
c
(
I
A
I
(
(
(
1
1
1
The naturalistic tradition
59
The historical lineage of a naturalistic tradition, described in this chanter offers a
different, complementary view of how the nature of reality might be established
considers another possible ontology, based upon a socially constructed reality
and offers a related epistemological framework, framed in the precents of
qualitative research. If the biomedical tradition expresses a predominantly
reductionist world-view, the naturalistic tradition adopts
adopts a more relational
view, arguing for an understanding not just of the structure of a system but
also of the relationship between the structural components. In the third quarter of
the twentieth century, a third intellectual tradition, predicated on non-linear
change, emerged from a diverse range of research fields - mainly biology
computing and mathematics. It is termed complexity, and I argue that this
could constitute a third tradition, and an extension of the two traditions that
have already been described. The next chapter explores this tradition by
following, as best as can be done, its historical origins (conscious always of the
post-hoc rationality inherent in that kind of analysis), considering the implica
tions of that analysis for clinical medicine, and speculating on a methodology
which might allow us to deploy the principles of complexity in contemporary
neaimcare.
r
1
I
7'
5-
Chapter 5
The non-linear tradition: historical
development of complexity
Introduction
This chapter presents the key ideas in the development of complexity, describes
the nature of complex adaptive systems, and reflects on how those principles can
be applied in healthcare. The term System' follows the definition given by Pisek
(2000), namely the coming together of parts, their interaction and sense of
purpose. Complexity is the term used to describe one of four generic types of
dynamic behaviour that a system can exhibit. The first two system behaviours are
stasis and order. Stasis denotes the absence of dynamic behaviour, and order
denotes a behaviour that is predictable, linear and stereotypical. Chaos refers to a
system that appears random, but within which there is determinism and hidden
order. Complexity is the dynamic state between order and chaos. Battram (1998)
gives the analogy of the breaking surf-wave. The tube in such a large curling wave
can be regarded as the complex phase of the wave's behaviour, the phase in the
wave's development before it crashes into chaos on the beach. Complexity exists
at the edge of chaos.
Three points will be conveyed in this chapter. These are the pervasive nature of
non-linear systems, the importance of the interaction between the components of
such systems (termed the system's organising relations), and emergence. Emer
gence denotes the ability of such systems, through the iterative patterning of their
interactive relations, to create fresh behaviours and properties, whose nature
could not have been predicted simply by understanding the system's components
alone. This chapter presents the principal developments of complexity in a
chronological sequence in order to build up a picture of the paradigm developing
in a wide range of increasingly related disciplines - biology, mathematics, ecology
and computing.
The origins of the non-linear paradigm: the debate
about structure and pattern
The detail of this historical overview focuses on the twentieth century, when the
principles of complexity were first defined, and their potential application in a
wide range of disciplines was recognised. The debate, which led to the current
detailed understanding of complex systems, began early in the twentieth century
among biologists who debated the nature of cell differentiation. However, this
debate about the relative importance of structure and pattern in systems was not
new. It can be found at the dawn of Western thought, when an enquiry into the
62
Complexity in primary care
relationship between structure (or matter) and pattern (or form) was first
recorded by Thales, Parmenides and Pythagoras. Aristotle also recognised the
distinction between matter and form. Matter contained the essence of all things,
but only as a potentiality, Aristotle argued, and form or pattern was what gave
this essence actuality (Sweeney and Griffiths, 2002). The Greeks also struggled
with determinism. Was the universe governed by deterministic laws? Can we
predict precisely what will happen to systems, and if so, how?
Epicurus set out what became the conventional position at that time, asserting
that the world was made up of atoms and a void. The Greeks believed that the
atoms fell through the void at the same speed and on parallel paths. This model
immediately posed the problem of human freedom. In what could the meaning of
human freedom consist if the world was thus deterministically composed of
atoms? Epicurus proposed a solution which he termed 'clinamen.' Lucretius
described Epicurus' solution as follows:
H”
i: :
While the first bodies are being carried downwards by their own weight in
straight lines through the void, at times quite uncertain and at uncertain
places, they deviate slightly from their course, just enough to have been defined
as having changed direction. (Bailey, 1947)
Heraclitus, contributing to this debate, argued that novelty need not be intro
duced if the nature of becoming was emphasised. He argued that 'truth lies in
having grasped the essential becoming of nature, that is having represented it as
implicitly infinite, as a process in itself' (cited in Popper, 1963). Later, in the
Sophist, Plato concluded that man needs to incorporate both being and becoming
into any explanatory framework, a duality which has tested Western philosophy
ever since (Plato, 1979).
To gain a more detailed understanding of how complexity emerged as a
valuable way of making sense of the world, it is worth considering the tension
between the two schools of biology - vitalism and organicism - at the beginning
of the twentieth century.
Early twentieth-century biology: the problem of cell
differentiation
Early in the twentieth century, biologists became interested in how cells in living
systems were able to differentiate. How could organisms, whose cells multiplied in
number from one to two, from two to four, and so on, doubling each time,
differentiate if their initial genetic material was identical? How could this identical
genetic material produce tissues as diverse as skin, muscle, nerve and bone?
Biologists were divided into two schools - vitalism and organicism. Although
both were opposed to a simplistic, reductionist understanding of biological systems,
they differed markedly in their proposed understanding of cell differentiation.
Vitalists thought that an additional non-physical force must be added to the physics
and chemistry of the cells to explain their ability to differentiate. The organicists
disagreed with this view, arguing that what was important was an understanding of
the relationships between the components - what they called their organising
relations (Haraway, 1976). In the early decades of the twentieth century the
biologist Joseph Woodger and the biochemist Lawrence Henderson made import
The non-linear tradition
63
ant contributions by introducing the terms 'organising relations' and "systems
thinking', respectively. What Woodger was emphasising by using the term 'organ
ising relations' was the idea that the essence of a system resided not in the structure
of each component, but in the way that each component in a system could interact
with, relate to and ultimately adapt alongside other components of the system
(Capra, 1996). The emergence of systems thinking had a profound influence on
scientific thinking generally in the Western world. Its central tenet was that the
essential properties of a living system are the properties of the whole, a property
held by none of the parts separately (Haraway, 1976). For systems thinking, the
context in which any system operated was of fundamental importance.
Heisenberg’s uncertainty principle
Developments in physics were informed by and advanced this emerging under
standing of complex adaptive systems. Heisenberg's uncertainty principle which,
put most simplistically, proposed that the more one measured the velocity of a
particle in a system, the less one could accurately determine the location of that
particle, and vice versa, had a dramatic effect on the conventional understanding
of what 'science' meant. In Heisenberg's own words, 'the foundation of physics
has started moving, and this motion has caused the feeling that the ground would
be cut from under science' (Heisenberg, 1971). The relevance of this uncertainty
principle is elegantly set out by Cohen and Stewart (1994): 'The answers we get',
they write, 'depepd on the questions we ask.' Consider, as Cohen and Stewart do,'
the analogy of a tree. We can interrogate the properties of a tree as a plant, or as a
boat, or as a pole for holding up telephone lines. The more we know about the
tree as a boat, the less we shall understand about its plant properties. The
uncertainty principle operates in this context, too. One cannot simultaneously
test a tree for its telephone-line-holding properties and its boat-like properties.
Heisenberg was commenting on a shift, as he saw it, from understanding the
parts to understanding the whole as part of a general conceptual revolution, so
much so that he entitled his biography Der Tiel und das Game - The Part and the
Whole - only to discover that his publishers, failing to realise the subtlety of the
title, had renamed the book Physics and Beyond (Heisenberg, 1971).
1940s: development of systems theory
Systems theory is conventionally associated with the work of von Bertalanffy, an
Austrian biologist whose contribution was to bring together developments in
biology, ecology, quantum physics and Gestalt psychology into a new way of
thinking which operated in terms of connectedness, relationships and context
(von Bertalanffy, 1968). The key characteristics of systems thinking, in relation to
living systems, are as follows.
• There is a shift in focus from the parts to the whole.
The essential properties of living systems are properties of the whole, and none
of the parts have these properties.
• These essential properties arise from the organising relationship between the
parts.
• These properties are destroyed by reducing the system to its component parts.
I
64
Complexity in primary care
• Systems nest within other systems - cellular systems nest within physiological
bodily systems, which nest within a human body, which nests within a person.
• At different levels within each living system, there is an increasing degree of
complexity. This is described as emergent, as it 'emerges' at different levels
within the system.
von Bertalanffy combined these insights from the first half of the twentieth
century with the process-oriented philosophy of Whitehead (1929) and with
Cannon's (1939) concept of homeostasis to create a theory of open systems. One
example of an open system is cellular metabolism - a continuous cyclical process
of synthesis, production of nutrients and excretion of waste that occurs within an
open environment in terms of its dependency on, and interrelationship with,
other provider systems (which provide the material for metabolism) and receiver
systems (which receive the metabolic products of that system).
1950s: development of self-organisation
i
i
1
Self-organising behaviour refers to the tendency within complex systems for
patterns of observable, coherent behaviour to emerge from what initially appear
to be random interactions. This was first observed by two chemists, Belousov and
Zhabotinski, in a very simple chemical reaction (which can be easily reproduced).
They prepared a mixture of citric acid, sulphuric acid and potassium bromate,
placed it in a shallow dish and stirred it. When this is done, bright blue dots
appear and spread, and then red dots appear in the centre of the blue dots,
forming expanding blue and red rings. When these rings run into each other, they
do not superimpose like waves, but form more intricate red and blue circular
patterns. This was the first (and is still the most easily reproducible) example of
spontaneous formation of patterns from a sea of chaos (Cohen and Stewart,
1994).
One might ask, so what? Self-organising features are emergent properties of
complex adaptive systems. They are emergent in the sense that their nature could
not have been predicted from a reductionist understanding of the separate
constituents of the system. For example, a wave is an emergent property of
water. Self-organising behaviour is a fundamental feature of complex adaptive
systems. And since complex adaptive systems are pervasive in biological, human
and organisational communities, it is important to understand the nature of self
organising behaviour in order to ascertain how those systems work (Cilliers,
1998). Self-organisation operates through positive feedback within a system. In a
biological complex system, activity that confers an advantage on the system, or
causes it to behave positively, tends to augment the influence of those agents or
activities associated with the desired state through positive feedback. Over time,
the system will preferentially weight the input of agents whose actions provide
positive output, thus establishing repeating patterns of behaviour, which the
system expresses as stable characteristics.
Computer specialists were able to reproduce self-organising behaviour in their
early modelling of binary systems. Binary systems are systems whose elements
can switch on and off, depending on the state of adjacent elements within the
systems. While the work was taking place with electrically lit binary computing
systems modelling neural activity in the brain, researchers noted the emergence,
72
Complexity in primary care
through which coherent action can be expressed, the presence of leadership .to
initiate complex responsive processes, and the potential to engage other agents to
co-create and adapt the system. The evidence from one national study of
healthcare organisations undergoing transformation change (Durie et al., 2004)
has identified a number of characteristics of receptive context. These include the
following, a recognition that ways of working need to be improved, and that
within the process of improvement, work practices may become quite different; a
iecognition that, in order to co-create new, different working practices, relation
ships are crucial, implying the need both to reconfigure existing relationships and
to create new ones; and a recognition, as a consequence of the first two
characteristics, that communication is the bedrock for initiating such change,
and that within communication the use of language (professional versus
informal, and specialist versus lay) is central.
Complex responsive processes
r.
The basic unit of activity within a 1human complex adaptive system is the
communication between individuals through
which those individuals
'
s co-create
and, in the process, make sense of the system. The basic unit of communication is
called a complex responsive process (Stacey, 2001). Through their patterning, the
system self-organises and develops its unique characteristics. 'The modelling of
complex systems', Stacey (2001) asserts, 'demonstrates the possibility that inter
action between entities, each entity responding to others on the basis of its own
organising principles, will produce coherent patterns with the potential for
novelty in certain conditions.' Stacey maintains that interaction through complex
communication constitutes a self-organising process, with coherence (an epi
phenomenon) as one of its emergent properties: 'There is no reason to look for
some kind of underlying blueprint, plan or predetermined mechanism other than
the interaction itself to explain coherence in human action, with its character
istics of continuity and potential transformation' (Stacey, 2001).
Three points need to be emphasised. First, complex responsive processes have
the potential to be transformational (Stacey, 2001). In participating in a complex
responsive process, the conversation is changed, each participant is changed, the
nature of their relationship can change and, by a ripple effect, the nature of the
participants' relationship with the larger system changes. Secondly, these pro
cesses are inherently unpredictable. During their course one participant issues a
gesture, which in turn calls forth a response from the other, in an iterative,
interactive and self-organising process. Thirdly, the patterning of such processes
and relationships, formed as a result, constitutes the self-organised characteristics
of the system, which confer a degree of stability, allowing it to be recognised and
described (Stephenson, 2004). This description of complex responsive processes
has two implications for any methodology that purports to analyse a complex
system (e.g. a healthcare organisation). First, it should alert researchers to the
importance of collating data about relationships - precisely who talks to whom,
where (not just formally but via the shadow organisation), and how those
discussions develop. Secondly, following on from this, is the recognition that
storytelling and narratives become a key data source, as it is within narratives that
participants in a system describe their formal and informal participation via
conversations. They can recount how these conversations changed the story-
The non-linear tradition
71
Characteristics of complex systems
At this point we can summarise the key features of complex systems before going
on to consider some examples of them in a range of disciplines.
Five key features can be identified:
1
2
3
4
5
sensitivity to initial conditions
complex responsive processes
self-organisation
adaptation (leading to co-evolution)
emergence.
Sensitivity to initial conditions
Lorenz's work on weather systems demonstrated that, in order to understand
how a complex system might evolve, it is crucial to know as much as possible
about the initial conditions under which it begins to operate. Lorenz showed that
even a tiny alteration in the initial state of one variable in a complex system with
a large number of variables (e.g. a weather front) can lead, through self
reinforcing feedback, to large alterations in the way that the system evolves.
The sensitivity of a system to its initial conditions is important because of the
nature of the feedback, both positive and negative, which influences the direction
that a complex system will take. In the conventional understanding of linear
models, the notion of negative feedback, leading to the 'desired' state of
equilibrium, dominates descriptions of complex systems. Examples in clinical
medicine include the effect of increased levels of thyroid hormone in feeding back
to the secretion of thyroid-stimulating hormone, to maintain equilibrium, or the
ability of insulin - whether injected, stimulated or secreted from the pancreas - to
equilibrate the blood sugar level. However, such biological systems operate with
positive as well as negative feedback. Thus, for example, in diabetes there is not
simply a failure of negative feedback. The system (that is, the patient) is unable to
detect swings in blood glucose levels, is unable to respond to those swings when
they occur, and is, during episodes of flux, prone to behaviour which reinforces
the direction of change of blood sugar level (away from normal) through
ethargy, inattention, and missed or inaccurate dosing of insulin. Thus the
system is 'encouraged' to move towards even greater disequilibrium. It is more
useful to consider homeostasis as the delicate balancing of a range of inputs, each
of which may act on others, rather than as a negative feedback system leading to
equilibrium. Holt (2002a) gives the analogy of balancing a snooker cue on the
palm of the hand - corrections to a potential imbalance are not undertaken in the
plane of the imbalance (the linear response), but by small, repeated corrections in
a wide range of directions.
In human systems, the equivalent of the sensitivity to initial physical condiuons that was described in relation to weather systems is termed enabling
framework or receptive context (Mitelton-Kelly, 2003; Durie tt al., 2004) The
term evokes the idea of an infrastructure of communicability - a set of conditions
that have the potential to facilitate the development of complex conversations
and actions whose patterning can, over time, constitute a complex system Thus a
receptive context implies a potentiality for coherent action, a set of values
I
j
2
70
Complexity in primary care
substituting slightly different values for each of the variables as he progressed
(Lorenz, 1963). Plotting the results in a notional three-dimensional graph on his
computer, he produced the eponymous graph with two lobes, shown in Figure
5.2. This is sometimes referred to as an 'attractor', a term which indicates a
representation of the behaviour of a complex system over time (Battram, 1998).
Figure 5.2 Lorenz butterfly.
Lorenz modelling revealed the inherent impossibility of predicting anything other
than over a very short range. The tiniest alteration in one of the values - for
example, changing a value at the third or fourth decimal place - could have a
dramatic effect on the direction of the system, as that result was re-introduced
into the modelling equations time and time again. Complex systems, it appeared,
were extremely sensitive to their initial conditions. If one tries to string together a
set of short-term predictions to create a long-term prediction, tiny errors creep in,
and these tiny errors, repeated iteratively as the modelling equations are run,
build in much larger errors (Stewart, 1989). However, the overall shape of the
Lorenz attractor also implied that the system - in this case broad patterns of
weather - will always remain somewhere within those boundaries. Thus one can
predict that given a broad set of initial conditions - British summer, say - the
weather is unlikely to produce temperatures below freezing or above 40 degrees
centigrade. However, it is much more difficult to say precisely what the weather
pattern will be in one location at a particular time.
Although it is beyond the scope of this book to explore in detail the
mathematical modelling of complex meteorological systems, it should be noted
that a firm mathematical basis has been established for analysing the nature and
evolution of such systems through, for example, the calculation of Lyapunov
exponents. A positive Lyapunov exponent is characteristic of chaotic systems, and
confirms that the system under analysis is sensitive to its initial conditions
(Schaffer, 1985).
The non-linear tradition
69
systems which exhibit period doubling. This is the pattern of doubling, which can
be charted at various stages in the system's evolution. When we see at the outset
the first period doubling cycle, we can visualise the system bifurcating so that it
diagrammatically resembles a tuning fork. As the system moves from a two-cycle
period to a four-cycle period, this bifurcation shape is reproduced, in smaller and
smaller iterations, as shown in Figure 5.1. The point at stake here is the self
similarity of the repeated patterning of these bifurcations. This patterning, known
as fractal patterning, is characteristic of chaotic systems.
Figure 5.1 Fractal patterning in period-doubling systems.
In the same way that Feigenabum's number is a universal constant, fractal
patterning has been found in all chaotic systems, and has been extensively
researched in analogue telephone signals, thermodynamics, geology, biology,
the music of Bach and the paintings of Jackson Pollock, the American abstrac
tionist, (Casti, 1995).
1960s: Lorenz Butterfly and the sensitivity to initial
conditions
Advances in computing during the third quarter of the twentieth century allowed
scientists to explore the nature of complex systems in more and more detail. When
exploring non-linear equations, especially equations with several variables which
were solved simultaneously, it became clear that the most minute difference in the
value of one of the variables at the beginning of a computation could make a huge
difference as those equations were solved iteratively, as one would do in order to
describe mathematically the evolution of a complex system over time. The classic
example of this is found in early attempts to model weather patterns.
In 1963, Edward Lorenz explored how weather systems might be modelled
mathematically to determine the extent to which the behaviour of such complex
systems could be predicted. He created a model of a simplified atmosphere using
just three variables which seemed to be crucial, namely the intensity of air
movement, the temperature difference between ascending and descending air
currents, and the temperature gradient between the top and the bottom of the
atmosphere.
These can be visualised as the three axes of a three-dimensional graph. For each
moment in time, one can plot a single point representing the combined functions
of the three variables. One can then imagine plotting serial points, showing the
location of the simplified atmosphere as time proceeds. Thus the development of
weather can be imagined as a tracing out of the single points over time.
Lorenz ran a series of equations, plotting each of these variables over time, and
68
Complexity in primary care
Bx - Bx1.
The equation will also involve feedback, as the output from the first iteration will
be the input to calculate the next step, or generation.
It is interesting, in terms of the application of non-linear mathematics to living
systems, to see what happens to the predicted population of insects when the
value of B varies. Consider the following few examples. When B is less than 1, the
population will die out, whatever the starting value of x. When B is greater than
1, but less than 3, this logistic equation settles down, after a sufficiently large
number of iterations (that is, in real terms, generations), to a fairly constant
population, with the equation solving itself at around 0.66, whatever value of x
between 0 and 1 is applied. This indicates a population that is fairly steadily settled
at around two-thirds of its maximum number. When B is greater than 3,
something quite different happens.
Once the iteration of this equation has been solved enough times, with values
of B just greater than 3, two different constant levels of population emerge, one
with a high population level and the other with a relatively low population level.
In real terms this makes sense. In one year there may be a large population,
which eats all of the food, and many individuals starve and die without
reproducing. The next generation therefore has a smaller population with
plenty of food. They all survive and lay eggs, so that the next generation is
larger, and so on. This phenomenon is known as period doubling, when a system
moves from a single equilibrium to a two-cycle steady state. Such an iterative
calculation of population levels can be continued. When the value of B (indi
cating the average number of offspring) is equal to 3.44, the period doubling of
the system moves from 2 to 4, and it jumps again, from 4 to 8, when the value of
B is equal to 3.56. When the value of B exceeds 4, there is no pattern to the cycle,
and the value of x, the total population, dots about randomly. This is chaos, and it
has been extensively studied (Tennison, 2002). It has been reached through the
iterative solving of a deterministic non-linear equation exposed to self-referential
feedback. This conjunction of determinism and positive feedback, with repeated
iterations leading to unpredictability, is characteristic of chaos (Gleick, 1998).
For mathematicians, the next step, which generated a huge amount of interest
in this type of mathematical modelling, was the calculation of the interval
between period doubling - that is, the numerical distance in the system between
period 1, fairly steady state, and period 2, oscillation between high and low, then
period 4, then period 8. Working at Los Alamos Laboratory in New Mexico, the
mathematician Mitchell Feigenbaum calculated that the period interval between
each bifurcation occurred at a constant ratio, namely 1:4.669. Researchers in
other fields soon discovered that this ratio, known as Feigenbaum's number,
applies to any such self-referential system (that is, one which feeds back into itself
via regular iterations), whether it occurs in biology, electrical circuits, geological
systems, oscillating chemical reactions or even, in principle, the business cycle of
the economy. Feigenbaum's number appears to be a universal constant, applied
to any such iterating self-referential system where period doubling occurs, in
literally any domain (Cohen and Stewart, 1994; Gleick, 1998; Gribbin, 2004). I
shall return to the potential importance of Feigenbaum's number when spec
ulating about the implications of non-linear mathematics in clinical medicine.
Feigenbaum's constant is related to another visible feature of self-referencing
The non-linear tradition
67
work of mathematician and scientist Henri Poincare, who reintroduced pictorial
representation into mathematics with his topological geometry a technique
whereby non-linear systems could actually be drawn as they evolved. Poincare
was interested in how systems evolved from the perspective of the whole rather
than by considering the parts of the system in isolation. He theorised about the
generic relationship of the whole of science to the facts of which it is composed:
'Science is built up with facts, just as a house is built with stones. But a collection
of facts is no more science than a heap of stones is a house' (Poincare, 1952). in a
later essay he writes, 'The aim of science is not things in themselves". . . but the
relation between things; outside these relations there is no reality knowable'
(Poincare, 1958).
The application of mathematics to non-linear biological
systems
Developments in the mathematics of non-linear systems initially improved
researchers' understanding of thermal and fluid dynamics (Gribbin, 2004). The
relevance of such mathematics to biological systems was subsequently explored
resulting in the ability to model the evolution of relationships between predators
and prey in defined ecological systems, such as a forest. For the purposes of the
arguments that are being developed in this book, it is worthwhile reflecting on
the nature of this modelling, as it was later applied to the spread of infectious
diseases, and it still holds out considerable potential as a research tool for
modelling other clinical conditions (Holt, 2002a).
Consider, in order to preserve clarity in the explanatory principles, a simple
exploration of the evolution of a population of insects, where the entire
population dies off in the winter, after laying eggs that will hatch out to provide
the next generation in the following season. We start with a population of x
individuals, each of which (again for ease of mathematical modelling) produces
an average of B offspring. We take into account the fact that some insects will die
before producing offspring - for example, if the initial numbers are large, and
there is not enough food for all the population members. This is accounted for by
setting an upper limit for the population, which can be done quite accurately
Cohen and Stewart, 1994), and then calculating the actual number, x, as a
raction of this, such that x will always lie between 0 and 1. Then, to take into
account premature deaths, the growth factor, Bx, is multiplied by (1 - x) in a
process termed renormalisation.3
We can then say that the population of such a system will rise and fall as a
unction of the birth rate, B. This is calculated, for varying values of B, by iterating
the following equation:
x(next) = Bx(] - %).
This is a non-linear equation, as it multiplies out to:
This works because if the population level at the outset of the analysis is very low all of
ie insects w.ll survive, (1 - x) will approximate to 1, and so the growth rate will be almost
exactly Bx Conversely, if the initial population level is high, x will nearly equal I and
- x) will approximate to 0, reflecting the fact that many members of the population will
starve or be eaten by predators.
1
L
66
Mill
Complexity in primary care
simple systems in real life. Their dilemma was best illustrated by their failure to
solve the problem of three celestial bodies under mutual gravitational attraction
(Stewart, 1989).
Stewart (1989) illustrates the conspicuous tendency of mathematicians to
linearise their equations and thus the solutions they could provide: 'it was a
linear world for most of the nineteenth and twentieth century.' However, in the
second half of the twentieth century there was a gradual acceptance in mathe
matics of the predominance of non-linear systems in nature. What mathemati
cians found when they applied their equations to non-linear phenomena was
striking. Simple deterministic equations produced rich and unexpected solutions.
Exact prediction, it seemed, was impossible, and self-reinforcing feedback
appeared to exert an important influence on such systems (Capra, 1996). In
non-linear systems, small changes could have dramatic effects as they could be
amplified by self-reinforcing feedback. A simple example will show this funda
mental characteristic of non-linear systems.
Mathematically, a feedback loop consists of a process referred to as iteration that is, repetitive solving of an equation, feeding back the previous solution to the
same function to obtain an iterated new solution, and repeating this process over
and over again. So if the function is to multiply the variable y by 4, shown by the
statement /(y) = 4y, then the iteration consists of repeated multiplications of that
function:
y
4y
4y —> 16y
16y —» 64y
z
and so on.
A very simple iteration in non-linear mathematics which illustrates this key
idea of non-linearity is derived from multiple (iterative) solving of the simple
function
y
ky(l -y)
where y lies between 0 and 1. Consider iterative solutions to this function where k
= 3, and y lies between 0 and 1. This can be worked out easily on a hand
calculator, but below I present a few solutions which show non-linear change.
Where y = 0
Where y = 0.2
Where y - 0.4
Where y = 0.6
Where y = 0.8
Where y = 1
0
0(1 - 0) = 0
0.2 -> 0.6(1 - 0.2) = 0.48
0.4
1.2(1 - 0.4) = 0.72
0.6 -> 1.8(1 - 0.6) = 0.72
0.8
2.4(1 - 0.8) = 0.48
1 -> 3(1 - 1) = 0
The numbers stretch out and then fold over, coming back to zero, in what is
known as the Baker transformation (Briggs and Peat, 1989). Importantly,
mathematicians pointed out that linearity was a subset of non-linearity - that
is, a special case of the simple non-linear equation
y = aZz + c
where a is a constant and n = 1.
A seminal contribution to the mathematics of complexity was made by the
L
I
The non-linear tradition
65
after a period of random flickering of the lights in the model, of a clear pattern of
repeated cycles. Even if the system was started randomly, an ordered pattern
would emerge. The process of ordered emergence of coherent behaviour was
termed self-organisation (Ashby, 1952). Within a decade of Ashby's report, Heinz
von Foerster proposed that in the process of self-organisation, systems increased
their internal order (von Foerster and Zoff, 1962), an observation which seemed
to counter the second law of thermodynamics. These ideas then gained wide
spread credibility and found increasingly subtle application in a wide range of
fields.
In thermodynamics, the Nobel-Prize-winning work of Prigogine (1998) showed
how an open system, far from equilibrium, had the capacity to respond to change
and disorder by re-organising itself at a higher level of organisation. Prigogine
made a series of observations about entropy, which can be loosely understood as
the amount of disorder in a system that is running down. The conventional,
Newtonian view was that the amount of entropy was increasing. That was the
basis for his Newton's second law of thermodynamics. However, Prigogine
measured not just the amount of entropy in a system but what happened to it.
He found that deterioration in systems was not inevitable. The disruption or
disequilibrium in a system, associated with entropy, need not inevitably lead to
dissipation (or equilibrium, the equivalent of dynamic death). Prigogine used the
term 'dissipative structures' to describe those systems which could give up their
original structures to recreate themselves in new forms. Such systems, according
to Prigogine, had the ability to self-organise.
Self-organisation became a central plank of the explanatory model in other
fields. The Gaia hypothesis of the English biochemist Lovelock, and the Chilean
neurophysiologists Maturana and Varella, all incorporate the notion of self
organisation in their explanations (Capra, 1996).
Developments in mathematics: self-reinforcing feedback
and non-linear equations
Parallel developments in mathematics and quantum physics fuelled the develop
ment of this non-linear paradigm. Towards the end of the nineteenth century,
mathematics had two sets of tools for solving problems, namely deterministic
equations and statistical analysis, for simple and complicated systems, respect
ively. Both shared the key feature of linearity, of which the equation
y = x 4- 1
is the simplest example.
Geometry, which was the original approach to mathematical solutions origin
ating in Greece, and algebra, which was introduced several hundred years later by
the Persians, had been unified by Descartes' analytical geometry, by which
technique mathematicians were able to represent linear equations pictorially,
using Cartesian coordinates in graphical form. Newton's subsequent contribution
was to develop differential calculus, which allowed mathematicians to represent
the motion of a body that was undergoing acceleration. What mathematicians in
the early twentieth century found, however, was that the exact solutions
provided by the elegant Newtonian mathematics applied to relatively few
The non-linear tradition
73
teller, and how the system might have changed as a result. This is the approach
now adopted by some organisational analysts, particularly those dealing with the
transformation of large international companies (Snowden, 2002; Mitelton-Kelly,
2003; Health Complexity Group, 2004).
Self-organisation
The essence of a complex adapting system is located in the basic interaction
between each of its agents. This interaction occurs at a one-to-one level, and is
then magnified through the interaction of other agents, all of whom interact with
each other, either directly or indirectly, through the process of adaptation or co
evolution. Thus the behaviour of complex systems consists of this myriad of local
interactions, the patterning of which constitutes the system's behaviour, allowing
it to be recognised and described. This process of co-creating coherent patterns of
behaviour is called self-organisation. Because complex systems exhibit non-linear
behaviour, the nature of such self-organising patterns cannot be predicted
precisely, and thus the product of such patterned behaviour - the emergent
properties of the system - cannot be anticipated either.4 Examples of self
organising behaviour include the flocking of birds and the behaviour of the
stock market (Battram, 1998). 'Flocking' is the self-organising process created by
a group of birds travelling together, and the resultant 'flock' is its emergent
property. It is an epi-phenomenon, neither planned in advance nor knowingly
constructed in a conscious, concerted effort by the agents (the birds).
Adaptation and co-evolution
Detailed studies of a wide range of ecosystems show clearly that relationships
between living organisms are, at their root, co-operative, characterised by coex
istence and interdependence (Capra, 1983). Although this view seriously chal
lenges the conventional Darwinian view of evolution, it is supported by a wealth of
evidence demonstrating the interplay of adaptation and creation in the process of
evolution. Kauffman (1993) illustrates this point by drawing the analogy of a small
forest-based ecosystem containing flies, frogs, fish and bears. There are many ways
tn which the frogs, who want to eat flies, and the flies, who do not want to be eaten,
interact. Frogs might develop longer or stickier tongues. Flies might develop more
slippery bodies to avoid capture, or an unpleasant taste to deter frogs when they are
captured. In a stable state, each of the frogs will eat a proportion of flies each season
- but this is a dynamic equilibrium, not a static state. Suppose that a frog does
develop a stickier tongue, and is able to catch more flies. At first a larger proportion
of flies will be eaten. However, the ones that aren't eaten are likely to be the ones
with the gene for a more slippery body, so that this advantage spreads throughout
the population of flies, just as the gene for a stickier tongue will spread through the
population of frogs. As a result, the system will settle down to a new state, where
roughly the same proportion of flies are eaten by the frogs. Although it may appear
from the outside as if nothing has changed, there has been a shift in the nature of
each of the agents in the system. And although each of the participants has
changed, so too has the nature of their interaction, the stickier tongue succeeding
in catching the even more slippery body. This is like the 'Red Queen effect' in Lewis
However, both can be described.
.1
74
Complexity in primary care
Carroll's Through the Looking Glass, where the Red Queen has to run as fast as she
can in order to stay in the same place.5
Emergence
Emergence is the key idea that holds together and unifies complex systems. It is
an epi-phenomenon, a higher-order feature of complex systems created by the
patterning of the interaction of its agents. The term refers to the potential within
complex systems, given appropriate initial conditions, to develop behaviours
(through self-organisation and co-evolution) which create emergent properties,
the nature of which could not have been predicted by knowing the components
of the systems at the beginning (the conventional reductionist approach).
Emergence is the product of self-organisation. Thus a wave is an emergent
property of water, a flock is the emergent property of birds flying together
(Battram, 1998), and temperature and pressure are emergent properties of
trillions of gas molecules in a box (Tennison B, 2004, personal communication).
Kauffman (1993), a theoretical biologist, has shown in a simple experiment
that complexity itself is an emergent property of complex systems. Kauffman
invites us to consider a system of a large number of buttons (say around 10 000),
laid out on a floor, which are increasingly connected simply by tying them
together with thread. You choose a pair of buttons at random and tie them
together. Repeat the process, not worrying if you choose, at random, a button
that is already attached to another one, as will increasingly happen as you
proceed. As the process continues, some buttons will become attached to more
than one other button, or to more than two or three others, and finally to more
than several hundred others (the vertical axis in Figure 5.3). Each button
represents a node in the system - that is, a point to which connections are
connected. Each such cluster of buttons can be termed a component of the
network. The number of buttons in the largest cluster (the largest component,
which may sustain 200-300 connections) is a measure of how complex the
system has become. Once the number of connections exceeds half the number of
nodes (the thread/button ratio, shown in the horizontal axis in Figure 5.3), it very
rapidly changes from one state (a large number of buttons with few connections)
to another one (a state in which almost every button is part of the network). This
relationship can be plotted quite precisely, as Figure 5.3 shows.
The relevance of emergence to living systems is well described by Gribbin
(2004) through the notion of autocatalysis. This term refers to a system that
develops the ability to continue to renew and generate itself, and it has been
postulated, again by the complexity theorist Stuart Kauffman, as a model to
explain the origin of biological life on earth (the model is speculative at present,
but is supported by a good deal of circumstantial evidence) (Gribbin, 2004).
Imagine, in the primordial chemical broth which existed shortly after the earth's
formation, that there developed some chemical substances which acted as
catalysts for other substances, like the catalytic process in the BelousovZhabotinski reaction referred to on p. 64. Suppose that chemical A catalyses
the production of chemical B. As the system develops, chemical B catalyses the
Indeed, the term 'Red Queen effect' was introduced into evolutionary biology by Leigh
van Valen at the University of Chicago in the 1970s.
The non-linear tradition
75
Size of largest cluster
300 -
200
100
0.5
1.0
1.5
Threads/Buttons
Figure 5.3 Phase transition and the emergence of complexity in a basic connected system.
production of C, and chemical C catalyses the production of D, and so on. If,
somewhere down the line of catalytic reactions, chemical X catalyses the
production of chemical A, the loop becomes self-generating and autocatalytic.
According to Gribbin (2004), Kauffman presents this model, with supportive but
not yet definitive evidence, as analogous to the connected-button model, namely
as a phase transition in a chemical system involving a sufficient number of
connections between the chemicals (analogous to the nodes in the button
model). This process of chemical autocatalysis is valuable because it illustrates
the idea of connectedness - the crucial interaction of individual components
within a system, whose iterative patterning forms a self-organising process with
the potential to create emergent properties.
Describing and recognising complex systems
At present there is no firm consensus about what constitutes the necessary and
sufficient conditions for a complex and adaptive system to be said to be present.
Current opinion (Mitelton-Kelly, 2004; Pisek P, 2004, personal communication)
suggests that the identification of a receptive context is a sine qua non, without
which there are no grounds to favour the patterning of complex responsive
processes or the self-organising of the system. Thus, contemporary experts agree,
the absence of receptive context implies that a system will be incapable of co
evolving and will, as a consequence, fail to develop any emergent properties by
which it might be recognised, described and explored. This analysis also implies
the existence of a temporal relationship between the features of complex systems
described above. If the presence of a receptive context is the necessary initiating
feature, then in human systems it is the enactment of complex responsive
processes which is the next sequential step. In an evolving complex adaptive
system, the patterning of such complex responsive processes will lead to self
organisation and, in turn, to the likelihood of co-evolution. Finally, in the
ji
76
Complexity in primary care
evolutionary sequence of a developing complex system, the pattern of self
organisation and co-evolution may lead to the appearance of emergent proper
ties, which tend to be stabilising features of such systems, allowing them to be
characterised and recognised - and researched (Stephenson, 2004).
As we shall see, this will have important consequences for researching such
systems. Although there is no agreement as to which of the features must be
present for a system to be said to be complex and adaptive, one can assume that,
in organisational research, for example, the absence of a receptive context is
sufficient grounds for asserting that the system under observation could not be
complex and adaptive. However, there is no consensus as to whether some or all
of the other four features described above need to be present. I speculate that it
would be premature to describe a system as complex and adaptive in the absence
of clear evidence of complex responsive processes, together with their patterning
in clearly identifiable self-organised entities. If a system had evolved to such an
extent that it had self-organised, then I take the view that, assuming the
continuing presence of a receptive context, co-evolution would be more likely
to occur than not. Thus the minimum conditions under which one could describe
a system as complex and adaptive would need to have evidence of a receptive
context, complex responsive processes and self-organisation. In the next chapter I
shall discuss a research methodology that incorporates an understanding of
complexity, scrutinising qualitative data in a second-level analysis from a
complexity perspective. In this chapter I use the criteria of a complex and
evolving system set out above.
1980s to the present: applications of non-linearity
in organisations
In the last two decades there has been increasing interest in the application of the
principles of complex systems to organisational change both in the commercial
sector (Stacey, 2000; Wheatley, 2000) and in healthcare (Pisek, 2000).
Applications in the commercial world
Wheatley (2000) contextualises her interpretation of complexity for organ
isational change consistently within the intellectual developments described
above. She cites Prigogine's understanding of self-organisation and its classic
illustration through the Belousov-Zhabotinsky reaction. She echoes Poincare's
view that 'relationships are not just interesting, they are all there is to reality'
(Wheatley, 2000). Her interpretation of the principles of complexity is informed
by quantum physics. Quantum physicists could identify a range of subatomic
particles, Wheatley explained, but these could not truly be understood in
isolation. They were particles in an intermediate state sustained within a network
of interactions (Zukav, 1979). 'Physicists can plot the probability and results of
these interactions, but no particle can be drawn independent from the others',
she observes (Wheatley, 2000).
Wheatley brings these notions together in a new model of change management
predicated on relational dynamics. She cites examples of her own fieldwork in
large commercial companies, where the reinterpretation of the principles of
complexity has been associated with large-scale successful transformational
r
The non-linear tradition
77
change. Oticom, the Scandinavian manufacturer of hearing aids, reorganised
their head-office space using self-organisation as its guiding principle, in what
amounted to a major de-structuring of their entire corporation (Pinchot and
pinchot, 1996). In an attempt to respond more swiftly and flexibly to the
changing environment in which they operated, Oticom employees literally
gave up their office space and furniture, swapping these for mobile essentials a cell phone, laptop computer, and file cart on wheels. So did the chief executive,
who located himself in marketing, finance or HR, depending on where an
immediate need had arisen (Pinchot and Pinchot, 1996).
Buckman Laboratories, a US-based manufacturer of speciality chemicals, have
reported an increased commercial capacity following their revised open distrib
uted approach to information - a prerequisite, according to Wheatley, for
effective self-organisation. The company recognised that information flow
could act as an organisational glue, encouraging richer connectivity between
the agents in their system - their employees - and they therefore introduced a
company intranet. One of the company's employees, challenged by some tech
nical information that was needed to close a business deal, made use of the
recently developed company intranet to request advice. Within hours he received
a range of replies from the company's centres in six countries. Not only did this
information help him to secure the deal, but also his technical query spawned a
further conversation between some of the respondents about the query, which
grew into an ongoing conversational resource - an interesting example of self
organisation (Willett, 1999).
Stacey diagram
Ralph Stacey, an organisational development specialist, has explored the
application of complexity principles in management theory. One particular
contribution that he has made is the agreement certainty matrix shown in
Figure 5.4 (Stacey, 2000). In this notional matrix the vertical axis represents
agreement - that is, agreement about the attributes of a system, and agreement
between the agents about an issue arising within that system. Certainty, repre
sented by the horizontal axis, is an indication of how sure one can be about the
cause-and-effect linkages within the system. Where one is close to certainty, one
can usually draw on previous experience of a similar issue in the past. New or
unexpected situations locate the agents far from certainty - towards the right of
this horizontal line.
This visual matrix can help managers to choose which approach might be best
suited to address issues which they can, by reflecting on their attributes, locate' at
different places within the notional space. For example, when operating in the
linear or simple zone, classic rational strategies such as process engineering can be
effective. However, in the zone of chaos they will not help. Here it is best to look
for patterns by continuously communicating with other agents in the system
before applying any coherent strategy. However, it is in the zone of complexity
that most of the issues facing large organisations lie. Building networks, enhan
cing communication, working collectively and allowing direction to emerge are
the guiding management principles here (Wheatley, 2000).
78
Complexity in primary care
Low
^Chaos
Agreement
Complex
Linear
High
High
Certainty
Low
Figure 5.4 Stacey agreement certainty matrix.
Applications in clinical care and healthcare policy
Pisek has recently incorporated the principles of complex adaptive systems into
his vision of how the US healthcare system should develop in the twenty-first
century. Like Wheatley and Stacey, Pisek focuses on the importance, in
healthcare systems, of the connections and interactions between components of
the system. 'A healthcare system' he writes, 'is a macro-system. It consists of
numerous micro-systems (doctors' offices, hospitals, pharmacies and so on) that
are linked to provide comprehensiveness of care.' Pisek distinguishes between
mechanical and adaptive systems:
In mechanical systems, we can predict what the system will do in great detail.
In complex adaptive systems, the parts (which in a healthcare system include
human beings) have the ability to respond to stimuli in fundamentally
unpredictable ways. For this reason, emergent creative behaviour is a real
possibility. (Pisek, 2000)
Pisek concludes that complexity provides a new paradigm to guide an under
standing of how systems work in healthcare.
Within the UK healthcare system, the principles of complex adaptive systems
have been reframed by Fraser et al. (2003) within the notion of 'agility'. An agile
system is one that can respond rapidly to a changing environment and markets.
The importance of rich interaction - a key feature of complex adaptive systems is stressed, as agile commercial organisations draw on their relationships with
suppliers, partners and customers to improve their practices. Using flexible
working patterns and virtual teaming, agile companies can, it is asserted, deliver
products more swiftly, with better quality and at lower cost.
The principles of agility, set out in Box 5.1, strongly evoke the key character
istics of complex adaptive systems, namely receptive context, self-organisation
The non-linear tradition
79
and co-evolution. In the UK there are examples in the NHS where these
principles of agility have been applied to healthcare organisations - for example,
in the redesign of older people's services in London.
Box 5.1 Attributes of agile systems applied to healthcare
I
I
Rapid changeover (e.g. in the use of operating theatres)
Doing today's work today (the basis of advanced access in general practice)
Co-operative rescheduling carried out with partners, stakeholders, patients
and carers
Flexibility, particularly in the constitution of teams
Synchronised scheduling
Care coordinated around a specific patient, not as part of mass customisa
tion
(Adapted from Yarrow et al., 2003)
I
Complexity and clinical medicine
Clinicians in medicine have also been slowly responding to the explanatory
potential of this non-linear relational paradigm. Non-linear systems have been
proposed as a better basis for understanding physiological and pathological states
in infectious diseases (Schaffer, 1985), cardiology (Goldberger and West, 1987),
neurology (Holland, 1998) and diabetes (Holt, 2002a). I shall discuss these
examples in more detail in Chapter 8.
Moving away from the purely clinical level, non-linear models have also been
postulated for education of healthcare professionals (Fraser and Greenhalgh,
2001), for understanding organisational change (Pisek and Wilson, 2001) both
in the NHS (Kernick, 2002) and in the North American healthcare system
(Zimmerman and Pisek, 1998), and for understanding the development and
embedding of clinical governance at the level of primary care trusts (Sweeney and
Mannion, 2002; Sweeney, 2003a). Hassey (2002) has proposed a theoretical
model for understanding the consultation in general practice, based on the non
linear principles of complexity.
Distinguishing between complex and chaotic systems
i
Both the Baker transformation and the non-linear equations that are the basis of
period doubling systems help to distinguish between chaotic and complex
systems. Chaos is concerned with those forms of complexity in which emergent
order coexists with disorder (Gleick, 1998). When a system moves from a state of
order away from equilibrium towards disorder, a new pattern of order can
emerge. This is what lies at the basis of the apparent paradox of order coexisting
with disorder, with determinism giving rise to unpredictability through iteration
and positive feedback.
Chaos theory is not the same as complexity, and it is helpful to distinguish
between the two, particularly in relation to social systems. Chaos theory describes
non-linear dynamics based on the iteration of mathematical formulae which, as
T'
80
Complexity in primary care
we have seen with the modelling of biological populations, can give rise to
unpredictable behaviour and the intricate patterning of fractals. However it is
within the repeated iteration of the constant formula that the inherent difference
between chaotic and complex systems lies. Complex systems may be capable of
adapting and evolving, and of changing the rules of their interaction - for
example, in relation to a major change in their environment. They are not
created simply by the iterative application of a formula. Thus one speaks more of
complexity when discussing human systems, as human behaviour allows for
choice, and the subsequent alteration of the nature of interaction - it does not
mimic mathematical algorithms (Mitelton-Kelly, 2003).
Summary
I
This chapter has described how in the twentieth century an intellectual tradition
predicated on non-linear relational dynamics evolved. Based upon observations
in biology, and assisted enormously by advances in mathematics and computer
modelling, complexity now presents itself as another explanatory model. This
chapter has laid out the key features of complex systems - sensitivity to initial
conditions (receptive context in human systems), self-organisation, co-evolution
and emergence - and has introduced the notion of period doubling in systems,
which can be described using self-referencing logistic equations. Two points still
need to be emphasised. First, there is sufficient evidence at present to allow us to
state that complex systems are ubiquitous. Secondly, they reflect the importance
of non-linear dynamics and organising relationships. The basic unit of activity in a
complex system involving human action is the complex responsive process, the
iterative patterning of which provides the system, over time, with its self
organising capability and potential for novelty. Emergence is the unifying feature
of such systems, and refers to the properties co-created by the interaction of the
components of the system in a deterministic but unpredictable way.
The application of the principles of complex adaptive systems is now to be
found in a wide range of disciplines, including ecology, thermodynamics,
meteorology, chemistry, and more recently management theory and clinical
medicine. Such a non-linear paradigm reflects a relational ontology, and con
structs its epistemological framework around the principles of complex adaptive
systems. Thus the description of complexity in this chapter contributes to the
propositions that are unfolding in this book by providing evidence of a third
explanatory model.
With regard to clinical medicine, there are two features of complexity that
demand serious reflection. These are the paradoxical juxtaposition of determin
ism and unpredictability, and the notion of complex responsive processes.
The analysis of the iterative patterning of non-linear equations introduced the
possibility of bifurcation - the possibility that a system could exhibit widely
swinging properties depending on the conditions present at the start of any
iteration. Remember what happened to our population of insects. We know that
their population will increase and decrease as a function of a fairly simple non
linear equation, but that as the birth rate ('B', the multiplier in the equation) just
exceeds 3, two different constant levels of population arose. Further bifurcations
arise as the value of B exceeds 3.44 and 3.56. This approach may help us to
understand more about the patterning of the incidence of some diseases, particu-
The non-linear tradition
8I
larly viral infections, which may exhibit a periodicity that can be explained by
straightforward non-linear mathematics. Indeed disease patterns, population
interventions and the attributes of a healthcare system may well interact in this
non-linear manner, constituting complex and evolving systems in their own right
(Tennison, 2002). At the root of this, in relation to the distinction between linear
and non-linear systems, is the notion of superposition. In linear systems, the effect
of the interaction of two variables, or two different causes, is merely the
superposition of the combined effects of those two causes (Tennison, 2002).
However, in a non-linear system, adding the effects of two elementary actions can
lead to dramatic, unpredictable and novel properties, as a result of the co
operativity between the two actions or causes. It is this co-operativity or
interaction that requires further research within the field of clinical medicine.
How do we understand the co-operative interaction between two or more
comorbid conditions? How much do we know about the interaction of medica
tions for different conditions when taken in conjunction for decades? Given the
description of complex systems presented here, and the evidence of their
ubiquity, it is not unreasonable to consider the need to explore such questions
from the non-linear perspective.
Complex responsive processes are a further useful notion within complexity
which may be relevant in deepening our understanding of the interaction
between doctor and patient during consultations. Remember Mrs B's consulta
tion at the beginning of this book. This was one of a set of consultations, in this
case going back a decade and a half, in which the participants - Mrs B and myself
- were changed, both in relation to each other and in relation to the other systems
(outside of the consultation in their own worlds) in which we participated.
Nothing could have predicted the outcome of that consultation, although the
conditions that led to it were all abundantly evidenced. My perception of the
outcome was that it combined a more profound interpersonal relationship with
this elderly woman, a greater respect for me (I sensed) on her part and, frankly, a
poor therapeutic outcome in terms of a clinical plan. The greater trust that we
both felt existed as a consequence of this event may in itself be seen as an
emergent property of the series of interactions leading up to, and including, that
consultation. Whatever conclusions one might reach, my argument is that the
principles of complexity provide a fresh and, I assert, more valuable set of
principles with which to explore and understand such consultations. In the
next chapter I shall consider some examples which show how the principles of
complexity have been deployed to address challenges in commerce, politics and
healthcare, before setting out more formally some thoughts on a methodology
that deploys those principles, with some examples of how they might be applied
to clinical activities.
r
3
Chapter 6
Developing an understanding of
chaos and complexity: implications
and examples
Introduction
I
The aim of this book is to support the proposition that the explanatory model in
contemporary medicine should be revised, and that this revision needs to
accommodate a plurality of world-views. To this end, I have presented a
conceptual exploration of the development and adequacy of medicine's con
temporary explanatory model. This consisted of a review of the history of
medicine, which showed how the model evolved and how science has come to
occupy its hegemonic position. This model was then assessed in the light of two
other intellectual traditions, namely the naturalistic and non-linear traditions,
from which a number of observations can be made. I shall argue that there is a
connection between an explanatory model, its epistemological framework and
the tacit acceptance of a related world-view, or ontological perspective. Viewed in
this way, it is argued that, in the practice of medicine, several ontological
perspectives are deployed. Those that are explored in depth reflect a positivist
reality (upon which the precepts of science are based), a socially constructed
reality (from which the fruits of qualitative research have developed) and a
relational reality (reflecting the dynamics of chaos and complexity). Each of these
traditions can be pressed into service in medicine to help to make sense of the
world from a particular point of view, by contributing through an expansion of
human knowledge. Each tradition operates in a preferred domain.
Complementarity in world-views
The implications of the relationship between the three intellectual traditions
described earlier are profound. Each has created a separate - but overlapping explanatory model. Each explanatory model is the product of assumptions about
reality (an ontology) and the way of knowing which makes sense of that world
view (an epistemology). This is how we each make sense of our world in our own
way. We have a view of how the world works, and we devise ways of creating an
understanding of it by developing types of knowledge that are compatible with
that view, which help us to make sense of the world accordingly. Thus, as the
outline of the three explanatory models described in this book implies, they are
the product of different world-views - which are not mutually exclusive, but
different. The scientific model is predicated on rational reductionism whose
epistemology retains the notion of linearity, expressed as regular, proportionate
J
84
Complexity in primary care
and stable relationships between cause and effect. The naturalistic model accepts
a more contextual, pluralist ontology in which a world-view is incrementally
constructed through experience. The third view, based upon complexity, implies
a world-view in which two principles, namely non-linearity and relational
dynamics, are fundamental. The devices of its epistemological framework there
fore populate that world-view and are compatible with it. The Stacey matrix,
which is presented in a modified form in Figure 6.1, helps to explain the
relationship between these traditions.
Low
Chaos
Agreement
Non-linear tradition
Complex
Naturalistic
tradition
High
Linear
Biomedical
tradition
High
Certainty
Low
Figure 6.1 Locating the three traditions on Stacey's diagram.
The dominant explanatory model in medicine operates best where there is
general agreement about the attributes of a system and a firm degree of certainty
about the causal links between them. Such systems are mostly linear, so the
explanatory model in medicine is seen operating best at or near the intersection of
the two axes.
However, the characteristics of the naturalistic tradition allow it to flourish in
less linear territory. Reflection upon the description by Smith and Heshusius
(1986) of the relationship between quantitative and qualitative research in the
latter part of the twentieth century suggests an evolving reciprocity or co
operation between these two approaches to investigation and knowledge. The
naturalistic tradition can shed light on systems whose attributes are poorly
understood, or on systems where the interactions between the components are
less predictable and regular. The relationship between the two is reciprocal and
can be cumulative. For example, we can describe the statistical benefits of
warfarin in patients with atrial fibrillation and then, drawing on the naturalistic
tradition, we can describe the struggle that doctors experience in implementing
this scientifically robust piece of evidence.
The third tradition of non-linearity operates at the edge of chaos (Zimmerman
and Pisek, 1998). Systems located in this space cannot be understood by applying
Developing an understanding of chaos and complexity
I
i
85
a linear notion of cause and effect. They will develop self-organising behaviours,
which will create emergent properties spontaneously, without a blueprint
(Battram, 1998). Agents that participate in these systems develop the systems
themselves, through co-creating processes of self-organisation (Wheatley, 2000).
This is what distinguishes systems around this notional space from chaotic
systems. Whereas the latter are the unpredictable outcome of the iterative
application of deterministic mathematical formulae, the former can, through
choice, change their rules of engagement and, as a consequence, the nature of
their interaction. In terms of organisations, the progress of such systems cannot
be managed by deploying command and control processes. Prediction is limited,
and uncertainty is inherent. The distinction between linearity and non-linearity
is not mutually exclusive - organisational systems can express themselves in
both linear and non-linear terms. In the National Health Service, this is seen in
hospitals that satisfy national targets by linearising their management approach
(e.g. to meet mandatory Government targets) while deploying the principles of
complexity (e.g. in transforming some of their services) (Pisek P, personal
communication).
•<41
1
Implications
What then might be the implications of this relationship - this complementarity between the three intellectual traditions described in the previous chapters? The
implications can be considered both theoretically and practically.
■r1
Theoretical implications
The influence of the scientific approach to understanding the world, described in
this book, has extended well beyond the fields in which it originated. Economics,
sociology, politics and most recently international relations have all been heavily
influenced by the scientific paradigm (Fukuyama, 1993). I have described how
healthcare policy in the NHS has also been heavily influenced by a reductionist
approach, expressed in the Taylorist predilection to view organisations as
machines (Taylor, 1911). I argue, on the basis of the observations and evidence
presented in this book, that healthcare policy should no longer predicate its
policies on this reductionist, mechanical view of organisations. I have described a
number of successful, explicit applications of an approach to organisational
development and healthcare policy based on a clear understanding of complex
adaptive systems (Zimmerman and Pisek, 1998; Pisek, 2000; Wheatley, 2000).
To this can be added the firm theory of the complexity of enabling infra
structures provided by Mitelton-Kelly (2003), which has progressed discussion
about the relevance of complexity to organisational development (Durie et al.,
2004). Emphasising that complexity provides a conceptual framework for thinking
about the world, as opposed to seeing the world, Mitelton-Kelly has reframed the
principles of complexity for the context of organisational development. This
reframing is summarised in Table 6.1. Although many of the implications and
their consequences may be fairly familiar, Mitelton-Kelly's contribution has been
to consolidate a theoretical basis for them.
The increasing acceptance of complexity in organisational research, both in the
commercial world (Wheatley, 2000) and in the healthcare sector (Pisek, 2000;
86
Complexity in primary care
Sweeney and Griffiths. 2002). will have implications for methods of research At
take is the ability of research programmes to assist in contemporary sense
makmg as new complex and adapting systems evolve. In terms of data collection
this wi afford a greater interest to storytelling and narrative analysis than before,'
trip- WL-nie 7 m°^e °n experts on discourse and conversational analysis to apply
their skills to analysing the co-creating conversations in healthcare (Snowden DV
2004, personal communication).
Table 6.1 Reframing the principles of
complexity for organisations (Mitelton
Kelly, 2003)
Complexity principle
----- --------r Complex responsive
(processes
Application to organisational development
The basic unit for the co-creation of a complex and adapting
system. The outcomes of such conversations, in terms of actions or
decisions, will ripple out and affect other parts of the system. Thus
any 'improvement' arising in one part of the system may impose
either benefit' or 'costs' on other related parts of the system
Relational dynamics
All the multiple dimensions of complex systems interact with each
other. In human terms, this means that interpersonal, social,
technical, economic and global dimensions may impinge upon and
influence each other. Alterations in relational dynamics may
change the rules of interaction. Agents that create such change do
so by acting on limited local knowledge, not a comprehensive
understanding of the whole system
Adaptation and
co-evolution
The connectedness between individuals is not uniform, and it
occurs both within and between systems. Individuals and their
organisations exist in an ecosystem, in which adaptation by one
part of the system alters the nature (fitness) of the system for other
parts. Organisational thinking alters when one considers the
possibilities of 'evolving with', rather than 'adapting to'. There is
no hard boundary between a system and its environment
Self-organisation
The patterning of complex responsive processes can co-create
relatively stable features of an organisation, as a result of
adaptation and co-evolution. These structures cannot be predicted
in advance, and do not develop as a result of a pre-ordained
blueprint. The products of self-organising processes become the
emergent properties of the system^
S«ondly. it should also endorse aa layered
layered approach
approach to methodology. In this
approach initia qualitative and quantitative
quantitative data
are ianalysed for emerging
data are
second-level analysis from complexity is
ducted on the themes that are identified. The report on the Pursuing Perfection
<a.pIrogramme of transformational change in healthcare services in
health and social care communities) described below serves as an example of this
from°clini " 1
Tn ChaPtCr 1 ShaU iUUStrate thiS W”h SOme more eMnlples
from chmcal research. Dune et al. (2004) have proposed a third-level analysis
ncorporatmg workshops and exploring the implications for policy and practice to
mplemenl the first two levels. The report on the Pursuing Perfection Pro
gramme is the first example of the application of such a method.
Developing an understanding of chaos and complexity
87
Practical applications of principles of complexity
A number of examples of the application of complexity principles to organ
isational change have already been cited (Zimmerman and Pisek, 1998; Wheatley, 2000). Three further examples of the application of complexity theory to
transformational change are presented in this section, the first from the com
mercial world, and the other two demonstrating the relevance of complexity to
national healthcare policy.
The European bank case study
I
Ii
Working with Papefthiniiou, Mitelton-Kelly advised a large European bank on
the reconfiguration of its entire information system, to prepare the organisation
for the arrival of the new European currency - the euro (Mitelton-Kelly and
Papefthimiou, 2000). The principles of complexity theory formed the basis for the
advice that they gave and the research they conducted.
For the bank, the challenge was that the legal and regulatory frameworks for
the introduction of the euro had been laid out, and their deadlines were clear.
These arrangements constituted a necessary but in themselves insufficient set of
conditions for change within the bank to occur. A raft of other conditions had to
be created internally to establish a receptive framework, or enabling infrastruc
ture, through which the bank could evolve broadly in a way that was consistent
with achieving its goals. Prior to this change, the internal structure of the bank
was such that systems developers, IT professionals, business managers and
operations personnel rarely met each other, and as a result they simply did not
talk to each other. Thus, although all of the agents in the system, namely the
bank's employees, agreed on the nature of the goal (preparing the bank for the
introduction of the euro), there was no effective enabling framework in place
through which those agents could act together to achieve this goal.
The leadership necessary to facilitate coherent action around the agreed set of
values (the bank had to maintain its reputation as it embraced the new currency
regulations) was provided by the manager who oversaw the project. This
manager:
• set up a programme of monthly meetings for all professional groups in the
bank
• supported these meetings by providing weekly information updates for all
personnel.
As a result of this:
• cross-dependencies between the participating groups were gradually identified
(after a period of indifference to the programme itself)
• once these dependencies had been identified, new forms of communication
emerged and new groups self-organised.
These processes and structures provided an enabling framework that was based
on trust. Trust arose from the conversations, or complex responsive processes,
whose patterning co-created the system's self-organisation, which is expressed in
the new conditions for joint working. It grew out of the cross-dependencies,
encouraged by the monthly meetings and weekly updates. From this perspective.
I
I
88
Complexity in primary care
trust became an emergent property of the system. In turn, the trust fed back
positively and iteratively, into the cross-dependencies, strengthening the con
versations through which they were expressed. Sufficient continuity wL ensured
throughout the project by the monthly meetings, and the subsequent self
organisation of the professionals into autonomous groups, with the authority to
experiment and take decisions within their domain of competence. Interestingly
project manager introduced into the organisation an 'interpreter'- literallyVn
agent who mediated dialogue between the domains of expertise that were
represented at the meetings. However, the communications were never either
managed or controlled from the top of the organisation
the enmnk
°rganisational 80als in P^ty of time for the introduction of
LH v
LI I Lv .
Pursuing Perfection Programme
*»>
tTohoZCOndtCOmr1lted StUdy WhiCh SUPP°rtS the relevance of complexity theory
o organisational change, this time in healthcare, describes the Pursuing7Perfec
tion Programme - a programme of transformational change in health and social
care communities, which was conducted at four sites in the NHS in EnXnd
(Sweeney, 2003b,.
essence, (one beahh and social care com™ „«es £
se’ .lLs7i„lXet'„° ?rt“pa'' ln ,his
agreed »
sense
services initially two, then five more) around patients. The idea was to be radical
nd
plan services an,and pa.lenis' Ilves, no. .be o.ber way “„d (Beva
named bv ."il “"”,"u'’ica“on>- Fro">Incep.ion, .be programme was anconspamed by a constructive enquiry, a process of research, which collated and
commented upon the main themes constituting the programme's activities at
each site. In summary, the enquiry consisted of a three stage analysis In tile firs!
stage, a standard case study was conducted (Creswell, 1988). In the second stage,
nersnerC1 eS ai?Sln8 T™ thlS lnitlal Qualitative analysis were interrogated from the
fevefanalysi” took’thT7 (3S f" the
°f
PaperS
ChaPter 7>- The
tevei analysis took the form of workshops with fieldworkers, NHS professionals
two levekf atlOnt eXPartS'
WhiCh the imPlications of the findings of the first
two levels for policy and practice were debated. The methodology of this enauirv
SXL <swXX“*ly!,s ,l" '"’p'ri“‘ ““ pres'n“d ” ch*p“r 7
The final report on this study, documenting the
progress of two of the sites
which succeeded in transforming their selected services, identified eight princioal
conditions which constituted the receptive context for whole-syster/transforma
tional change within the participating organisations (Durie et al 2004) These are
hown m Table 6.2, where each of the conditions identified through the firs deve
reThrnd'coL16^1?^
PUrSUing PerfeCtiOn p™gramme fs set out in the
In .he
emZ' rela'ed
»'
’dap“”
h >how„
V
Developing an understanding of chaos and complexity
89
Table 6.2 The eight conditions that constitute the receptive context for whole
system transformational change (Durie et al., 2004)
Condition for receptive context in the Pursuing
Perfection Programme
Feature of complexity
Recognising that things are not working well Sensitivity to initial conditions. The
enough, or could be done differently, with necessary but in themselves insufficient
better outcomes for patients
conditions for change to occur
Leadership, demonstrating genuine
Leadership to facilitate coherent action, as a
commitment to aspirational goals. Visible
prerequisite of an enabling framework
behaviour change by leaders, indicating
genuine commitment to the programme and
to projects, with flexibility and comfort with
ambiguity and emergence
Behaviour change by the agents.
Reconfiguration of relationships/creation of
new relationships among staff, and between
staff and patients
Reconfiguring relationships to co-create
fresh complex responsive processes, whose
patterning will help the system to self
organise
' ul
t
Encouraging a culture of experimentation
and supported risk taking
Acceptance of inherent unpredictability
in the system. Permission to allow
interdependencies to coalesce into small
self-organised experimenting groups
Accepting the possibility that different ways The set of values which constitute receptive
of working and thinking will be better for
context
patients
Genuine and meaningful patient
involvement
The set of values that constitute receptive
context
The importance of language (including the
challenge of professional language) and
communication (between and within
organisations)
The importance of communication as the
bedrock of the complex responsive processes
whose patterning co-creates coherent
behaviours and outcomes
Pursuing Perfection as a 'Way of Working'
The emergent property of coherent action.
Not just 'another project', but a new way of
working
5
ir
There are similarities between the themes in this study of the Pursuing Perfection
Programme and the study by Mitelton-Kelly and Papefthimiou (2000). Both
accepted the external environment as a necessary but insufficient condition for
change, both identified the need for new relationships (achieved by the monthly
meetings in the European bank case study and by the reconfiguration of
relationships in the Pursuing Perfection Programme) as a way of co-creating
new complex responsive processes, and both recognised a set of values as the
'glue' for holding the emerging behaviour of the system together. The culture of
experimentation that was fostered in the Pursuing Perfection Programme and the
trust that was engendered in the European bank case study constitute the
respective emergent properties of the system.
J
90
Complexity in primary care
A strategy for containing AIDS: the experience of Brazil
The third example of the application of complexity theory to 1healthcare
' '
policy
describes the Brazilian government's strategy to contain the AIDS epidemic in
Brazil.
In the 1980s, Brazil had one of the highest infection rates for AIDS in the world
(Darlington, 2000). With an accelerating infection rate for the virus far in excess
o that in South Africa, coupled with an annual per capita income of less than
$5000, the World Bank predicted disaster, calculating that Brazil would have 1 2
million cases of AIDS by 2000 (World Bank, 1997). In fact, in that year 0.5 million
cases were reported to the World Health Organization, representing an infection
rate of 0.6%, compared with a rate of 25% in South Africa (World Health
Organization, 2002).
Gloubermann and Zimmerman's (2002) careful analysis of the World Bank's
appraisal reveals how the Bank's predictions were predicated on a linear
modelling of the issues. This led them to develop complicated rather than
complex assumptions about Brazil's predicament, from which an apocalyptic
picture emerged, predicting disaster for the country. The World Bank's assump
tions included the following.
Effective treatment, in the form of antiretroviral treatment, is too demanding
of clinical services in poor developing countries.
• Poor countries rapidly realise that they cannot sustain the cost of effective
treatments, so they concentrate exclusively on prevention.
• Even if drug treatment is available for some AIDS patients, the ill-educated,
barely literate people, who are typical of the AIDS patients in poor countries'
cannot possibly manage their own complicated drug regimes.
• The way to implement effective prevention is to scare people - fear of death
will limit the spread of the disease.
• Effective prevention still results in huge losses to the <current adult generation,
and its benefits will take two or three generations to• accrue.
An integrated programme of prevention and treatment in combination is
beyond the organisational capacity of poor and developing countries.
Analysing the Brazilian authorities' approach to this challenge, the authors
describe what happened, before examining those actions from the perspective
of complexity. The Brazilian government's actions included the following.
They gave the drugs away free. They took a risky decision to manufacture their
own generic brands of the antiretroviral preparations, which up to that time
had been produced expensively, mostly in the USA, by huge international
pharmaceutical companies. The risk lay in a legal stand-off with the pharma
ceutical giants, who after nearly two years decided not to pursue the
government for breach of patent. By 2000, eight of the 12 available prepar
ations were produced generically in Brazil, and consequently the costs of
treatment turned out to be between 65% and 90% less expensive than in the
USA, upon whose figures the World Bank had based their calculations.
• They used treatment as a part of the prevention strategy, figuring that, when
people know that they will receive free treatment, they will be more willing to
attend for therapy. Those patients who did so received preventive advice, spread
1
Developing an understanding of chaos and complexity
91
the word to the close communities in which the disease was rife, and felt their
decisions reinforced as the progression of their symptoms slowed down.
• The Brazilian authorities accepted poor literacy and numeracy in their target
population as a challenge, and developed a huge number of creative ways of
tackling it. Doctors and nurses co-opted other healthcare workers, lay people
and patients themselves to produce their own ideas of how to get the key
messages across. Drawings of food were used to remind people when to take
the pills. Arrangements were made for food to be provided free through schools
and churches, giving a further point for compliance messages to be reinforced.
Humour became a key ingredient of billboards advertising free condoms.
• The Brazilian government seized upon the AIDS epidemic as an opportunity to
strengthen its healthcare infrastructure, rather than simply seeing it as the
reason for their failure to control the problem. They deployed over 600 pre
existing non-governmental and community organisations to access hard-toreach groups, and they established a network of over 130 testing and
counselling centres (Centre for Disease Control, 2000).
I
From the perspective of complexity, what the Brazilian government did was to
accept and make use of the messiness in their healthcare system, by maximising
the connectivity of the existing informal, social and community relationships.
They knew that they had a receptive context - a focus on potentially shared
activities (the care of AIDS patients), which all constituents agreed was main
stream business (no one was in any doubt about the fact that the country faced a
catastrophe). By facing up to the pharmaceutical giants, the Brazilian government
exhibited sufficient leadership to facilitate the interaction of the other agents in
the system. As a result, the system self-organised - with AIDS patients redefining
their informal groups as patient groups receiving treatment and preventive
advice. And the system co-evolved - churches and other non-governmental
organisations became actively involved as agents of the healthcare system.
Positive feedback, in the form of descriptions of symptomatic improvement by
AIDS patients themselves, reinforced the self-organising processes and supported
the emergent approach taken by the authorities.
In summary, as Gloubermann and Zimmerman (2002) have expressed it, they
reframed complicated questions as complex challenges. When the World Bank
had asked 'Who can you afford to treat? What will you have to cut back on to
afford this?' - linear questions - the Brazilian government asked 'How can we
reduce the cost of treatment so that we can provide it for everyone?' - a question
that demands a non-linear, emergent solution. 'What infrastructure do you need?
And from what existing service will you take the money to pay for it?', the World
Bank asked. The Brazilians transformed this into the following questions: 'Where
and what are the pre-existing informal arrangements that we can deploy as part
of an emerging infrastructure and how can we strengthen them?'. Complicated
became complex, predetermined became emergent, and a shortage of resource
was redefined as potential abundance.
Implications for practice: epidemiology and public health
The distinction between chaos and complexity explains why early research into
the implications of non-linear systems for clinical medicine focused on the
MM
•j
J
r
i:
I
92
Complexity in primary care
relevance of chaotic mathematical modelling to clinical issues. Where systems can
be described numerically, sometimes with great precision, their iterative pattern
ing becomes a real possibility, holding out the potential to elicit features of chaotic
systems, such as period doubling or bifurcation. A number of the examples from
clinical medicine illustrate this point.
One of the first examples of this was an important paper by Schaffer (1985) that
modelled the spread of infectious disease using simultaneous non-linear equa
tions. Although the mathematics in Schaffer's paper are beyond the non
specialist, the author presents a complete non-linear mathematical modelling of
infectious disease spread, showing that it exhibits fractal patterning and has a
positive Lyapunov exponent, which shows sensitivity to initial conditions (see
Chapter 7), both of which are essential features of chaotic systems.
Here Schaffer is exploiting the analogy between the biological species model
ling described in Chapter 7 and the infectious agent process. More recently,
Tennison (2002) has shown how the non-linear equation used to model species
survival can, with sufficient simplifications, be applied to the spread of infectious
diseases. Thus the equation becomes:
^r+i = ^xt (1 —
fI
I
r
where t denotes time, x denotes the point prevalence of the disease at time t, and a
denotes the virulence of the infecting agent, which is a variable, not a constant,
and can change over time. Such variation may, for example, be due to climate
change, genetic drift or the effect of other organisms. Iterative solving of this
equation, for variable values of a, produces the same bifurcations and period
doubling as were discussed on p. 68 in relation to species survival. But how does
this help public health specialists? Epidemiologists have long been familiar with
the cyclical patterns of incidence of certain viral infections, including measles and
pertussis. They are also familiar with diseases such as psittacosis and plague,
whose incidence seems to be quite random, and they recognise that yet other
diseases, such as multiple sclerosis, exhibit long-term trends (called secular
trends) for no obvious reason. The value of non-linear modelling, and the
attendant possibilities of period doubling (sometimes with long period intervals),
allow epidemiologists to speculate that these diseases might have an underlying
dynamic, with an oscillation in infectivity over long time periods, similar to the
period-doubling models of species survival. Speculating about the potential of
non-linear modelling in infectious epidemiology, Tennison (2002) acknowledges
that 'even a highly simplified non-linear model exhibits remarkably complicated
behaviour, similar to that seen in the real world.' Some parts of complexity
theory, Tennison concludes, have great potential relevance for epidemiology - for
example, in the more accurate planning of vaccination campaigns.
Example: cardiology
Cardiology provides an example of the potential of chaos and complexity theory
to lead to a revision of basic assumptions about the underlying conventional
models. Bifurcation is now a well-studied phenomenon in cardiology, and
Goldberger and West (1987) have contributed to an understanding of period
doubling in sick sinus syndrome. Fractal patterning (the tendency for self
similarity within the period-doubling trends in chaotic systems) occurs at the
Developing an understanding of chaos and complexity
93
His-Purkinje conduction network. Healthy heart rate variability has also been
shown to have a fractal structure, the loss of which appears to be associated with a
poor prognosis, suggesting that chaos underpins normal cardiac function. In their
review of the applications of non-linear dynamics to clinical cardiology (Goldberger and West, 1987), the authors speculate about fractal patterning (they use
the term 'fractal anatomy') in the pulmonary, hepato-biliary and renal systems.
Example: diabetes
In diabetes, the degree to which chaos and complexity have been applied is less
advanced than in cardiology or epidemiology, but non-linear models have been
suggested, and considerable speculation now underpins the debate about optim
ising diabetes management, drawing on non-linear mathematics. Holt (2002a)
disputes the conventional biomedical explanation of diabetes as a state of relative
or absolute insulin deficiency, with or without insulin resistance, arguing that
any predictive modelling for a patient with diabetes that is based upon linear
thinking has the ability to predict fluctuations in blood sugar levels for a period of
only 15 days (Liska-Hackzell, 1999). This analysis raises serious questions about
the routine staging of appointments at diabetic clinics at the current 3-monthly
intervals.
Holt (2002a) presents the conventional understanding of blood glucose behav
iour as:
141
I
3
G(postprandial) = G(preprandial) + a(C- I)
I
where C is the amount of carbohydrate eaten at the last meal, and / is the amount
of insulin injected or secreted since. Conventionally, 'a' is taken to be a constant
that is determined by patient-specific parameters such as body mass index or
insulin sensitivity. This has given rise to a negative feedback, equilibrium-based
model of diabetes in which, so long as C 'balances' / at each meal, the blood
glucose behaviour will remain stable. However, as Baxt (1994) pointed out 10
years ago, it has been a common misconception in pathophysiology generally to
mistake a variable for a constant, and it is Holt's contention that 'a' in the above
equation is better understood as a variable, in which case the equation becomes
non-linear. The additional features which contribute to 'a' and render it variable
include the impaired ability of the system (that is, the diabetic patient) to detect
movements of glucose levels away from normal, the inability to respond to such
swings when they occur, and the presence of positive feedback. The source of the
positive feedback consists of the behavioural mechanisms that affect some
patients with diabetes, which tend to move the system further away from
normal - for example, inactivity, lethargy, and inaccurate or missed insulin
dosage. Further positive feedback, in the longer term, comes from the cycle of
increased insulin levels leading to weight gain, which leads to an increased
insulin dose, and so on. Holt is careful to state 'some' patients, as others can sense
these fluctuations and react appropriately, suggesting that there is a balance
between negative and positive feedback, which is likely to vary between
individuals. Drawing upon the insights from the work of Garfinkel et al. (1992)
m cardiology. Holt speculates about the potential for a non-linear adaptive
algorithm to help patients with diabetes to manage fluctuations in blood sugar
levels. Any such adaptive algorithm, Holl suggests, would require attention to the
’j
.:y
r
94
Complexity in primary care
behavioural characteristics of the individual, reflecting the contribution of
positive feedback within the system, and awareness of blood glucose levels
without testing (to enable reaction to small perturbations), and residual endo
genous insulin secretion, the presence of which might allow the system to settle
at a displaced stable position, say with a slightly higher 'normal' stable range of
ood sugar concentration. Although as yet there is no firm empirical basis for
con rming or refuting Holt s speculation, the author commends the use of
prolonged one-dimensional time-series data, plotting blood sugar level at time
G,, and against blood sugar level one time interval later at 'G/+1', and repeating
this over many hundred data points, on the grounds that chaotic systems can
sometimes be revealed in this way (Rossler and Rossler, 1994)
Implications for research
J
Researching health technology as a complex system
Support for the analytical approach (Sweeney, 2003b; Durie et al., 2004) and
aPP lea to the empirical data is found in Griffiths' description (2002) of the
methodology in her research on the impact of new technologies. In the course of
t is research, Griffiths explored in detail the impact of two new technologies,
namely mammography and bone densitometry, on middle-aged women.
Griffiths argues at the outset that her subject matter can only be explored by
seeing it as a complex system, in this case bounded by the policy, practice and
consequences of technological innovation. The impact of such technology is
predicated on complex responsive processes and characterised by iteration
feedback and co-evolution. In her research she adopts a dual approach, termed
fine and coarse grain', to her exploration of the impact of new technology,
acknowledging that health technologies are shaped by society and that they in
turn shape society. Further interactions and feedback within the system become
the focus for the research's interest at three levels:
i those developing and providing the technology
2 those using the products of their innovation
3 those observing and commenting on it publicly (government reports) and
privately (e.g. carers and colleagues of users).
At the 'fine-grain' level of research, Griffiths and her colleagues focus on the
interactions between those providing new technologies (in this case mammo
graphy and bone densitometry) and those consuming them (middle-aged
women). These can be seen as the complex responsive processes that co-create
the impact of the technology. Data at this level are collated from descriptive
quantitative data (what is being used?) and detailed qualitative data (how is it
being used?). At the 'coarse-grain' level, the focus becomes the macro-issues,
such as the cultural, policy and organisational issues at regional or national level
that form the system's (loose) boundaries. Material from this level of research is
drawn from government reports, professional guidelines, guides from self-help or
voluntary groups, or commentaries in the media. This level of the research
accepts that the environment in which the technology is provided will impact
on, and in turn be affected by, the experience of providers and users. The
researchers look for patterns in their data - patterns of response, use, comment
Developing an understanding of chaos and complexity
95
and user experience - which constitute the system's emergent properties
(Griffiths, 2002). Their search for patterning parallels the second-level analysis
of the empirical data described in Chapter 7. Griffiths adopts a co-evolutionary
approach to her material, assuming that change is occurring in all of the
interacting populations of the system under scrutiny, and accepting that, as a
consequence, change can be driven in both directions by both positive and
negative feedback, and between the participants at the 'coarse' and 'fine' levels
of activity.
The research described here, together with the example from Durie et al. (2004)
described earlier, has implications for this type of health services research
generally. Henceforth, research methods will benefit from considering both the
linear and non-linear features of any domain. In clinical research it will always be
necessary to explore as robustly as possible the basic science of a problem, its
pathophysiology and any proposed technical intervention (the linear side of the
domain). What we learn from complexity is that this is not enough. The
subsequent interaction, between innovators, developers, providers, users and
their carers, needs to be rigorously described in order to capture the reality of a
system in which use (or disuse) of an innovation is accepted as an emergent
property of the system, co-created by the complex responsive processes of its
agents (technicians, guideline developers, patients and carers). This aspect of
health-related research will combine both conventional quantitative data and
methods (how much of this innovation is being used?) and qualitative data and
methods (how is this being done?), looking for patterns within the data which
represent iterative feedback and co-evolution within the system that is under
scrutiny.
Implications for healthcare policy
The ubiquity of complex systems seriously calls into question the rationale behind
command and control management policy in healthcare. Identifying what they
call co-evolutionary dynamics, Volberda and Lewin (2003) call on policy makers
to commit themselves rather to guiding the evolution of behaviours that emerge
in the course of the interaction of independent agents within any system. Policy
makers, these authors argue, need to encourage self-organisation, recognising the
potential of organisations always - through the patterning of their complex
responses - to find order, however complex or convoluted the environment.
Healthcare research needs to take account of co-evolution, focusing on the
emergent properties of a system (that is, a healthcare organisation or commu
nity). These emergent properties arise from the micro-state adaptations that are
co-created by the system's internal complex responsive processes, and the macro
state adaptations that reflect the community's interaction with the wider
environment. The authors' recommendation for the more widespread use of
longitudinal time-series data sets has implications for the funding of healthrelated research. If one cannot predict either how long it will take for a system to
evolve, or what direction that evolution might take, how can one frame research
projects that are 'good enough'? One response, currently supported by the
Modernisation Agency of the Department of Health, and applied to the Pursuing
Perfection Programme in the NHS in the UK, is to undertake policy innovation
and research concurrently, speculating initially about how much time will be
F
96
Complexity in primary care
'enough', before gradually building up experience of just how much time is likely
to be needed to make intelligent observations.
Summary
ar
We are now in a position to restate the key themes in the debate about
explanatory models, and their relationship to ways of knowing. This book asserts
that the conventional explanatory model in biomedicine is predicated on
scientific positivism but that, in practice, professionals deploy a range of 'ways
of knowing' which take advantage of appropriate features of the three traditions
described in the preceding chapters, pressing them into service where appro
priate. My contention is that practitioners do this tacitly and subconsciously.
However, when the features of and relationship between the three traditions are
set out, one is forced to consider more explicitly the relationship between ways of
knowing and the nature of clinical practice. Mrs B's consultation illustrated the
tension that is created when two traditions clash. In these situations, one is
compelled to ask which tradition should predominate. How do we judge when
best to deploy one way of knowing against another? And who decides this? The
centrality of complex responsive processes to the interaction between doctor and
patient, and the subtlety of the inexorable changes that emerge from such
iterative interactions, compel practitioners to focus very precisely on a number
of issues. These include the way we explain things, the reciprocity of the doctor
patient-doctor relationship (which is mutually changing, not the one-way traffic
suggested, for example, by the term 'compliance'), and the way we as practi
tioners change as a consequence of our interaction with a whole range of patients
-as well as our interaction with the other systems in which we participate. From
Chapter 5, it is clear that the other sciences have accepted the challenge of
modifying their explanatory models as a result of their learning about the
principles of chaos and complexity. We await such a response from the
biomedical community, where the algorithmic bastions of explanation remain
unbreached.
unuicdciieu. However,
tiowever, before
betore considering the future.
future, I want to develop some
ideas about the implications of complexity for research in medicine, by consider
ing - from the perspective of complexity - some examples of a second-level
analysis of data that were collected during four standard research projects
conducted in primary care. The purpose of the next chapter is to demonstrate
the practical applications of complexity principles to the contexts described in the
examples that are presented.
Chapter 7
Using complexity principles in
healthcare research: examples of data
analysis using complexity principles
Introduction
This chapter presents four original papers, published in peer-reviewed journals in
the fields of health and social care, which show how the principles of complexity
can be incorporated into a research methodology in order to deepen our under
standing of some common challenges in general practice. Each of the papers is
reproduced in full, as published, in the Appendix, and a short resume is presented
here.
Thus the aims of this chapter are as follows:
• to present empirical data which develop the conceptual exploration of the
adequacy of the explanatory model in contemporary medicine
• to interrogate the main themes in four empirical papers from the perspective of
complex adaptive systems
• to identify any emerging hypotheses suggested by this analysis
• to commend possible avenues of future research based on the foregoing.
Four papers (in all of which the author was involved) are presented. In each case
a short resume is set out, to which are added some reflections on the research
projects, enabled by re-presenting some of the original data which were omitted
due to the editorial constraints of the journals in which the papers appeared. I
hope that by introducing this additional original data, the second-level analysis,
drawing on the principles of complexity, will be made clearer.
The methodology: second-level analysis from complexity
The methodology used in this analysis has been developed and deployed
successfully in research that I have conducted with the Health Complexity
Group of the Peninsula Medical School of the Universities of Exeter and Ply
mouth. The method has been described theoretically (Sweeney, 2003b), and was
implemented recently in a national study of transformational change in NHS
health and social care communities (Durie etal., 2004). In principle, this approach
involves a detailed mixed qualitative and quantitative case study at its first level.
At the second level, the themes identified via the first-level analysis are explored
for any insight which either confirms or disputes key features of complex systems.
These features, which were explained in Chapter 5, are as follows:
IM4
jI
-.11
4
if'
98
Complexity in primary care
• receptive context (an enabling framework or infrastructure of communic
ability conducive to an agreed purpose)
• complex responsive processes (the conversations which constitute the primary
unit of a complex adaptive system involving people, through which they co
create the system)
• self-organisation (the evolution of coherent units or behaviours consistent
with that purpose)
• co-evolution (identified through sustained and adaptive patterns of behaviours
consistent with the overall purpose)
• emergence (the unifying feature, an epi-phenomenon created by the pattern
ing of interaction of a system's agents).
I
Great care is exercised at the second level not to 'read into' the data possible
connections with complexity. The aim is to reflect on the data as they stand,
drawing on the principles of complexity, to determine the value of any insights
(from that perspective) for transferability and learning. Should the themes
identified in the four papers in this chapter not yield any insights consistent
with the features of complex systems, this would be revealed and discussed.
The criteria for identifying the presence of a receptive context are set out in
Chapters 5 and 6. Thus the absence of a receptive context is taken as sufficient
grounds for assuming that the system under scrutiny was not complex. The
presence of a receptive context, coupled with evidence of self-organisation
through the patterning of complex responsive processes, is accepted as the
minimum condition under which it becomes justifiable to describe the system
under scrutiny as complex. In the papers analysed in this chapter, these criteria
were applied after the data had been collected and had undergone first-level
qualitative analysis.
Paper I
Evidence-based practice: can this help joint working?
Published in 2000 in Managing Community Care. 8: 21-7.
Background
The first paper captures the findings of a small qualitative study which compared
the way in which participating health and social care professionals conceptualised
the medical model, in its contemporary form of evidence-based medicine, and
how they compared that with what they called the social model. The paper
capitalised on a rare opportunity in which a range of health and social care
professionals had come together for a study day to learn about and discuss
evidence-based medicine. The purpose of the study day was to provide a work
shop introducing the basic principles of critical appraisal, and to explore, in small
group work, the participants' attitudes to and beliefs about evidence-based
medicine. The study day took place against a backdrop of little evidence, within
that ------health__1
and social
care community, of joint working in the context of
evidence-based practice.
Using complexity principles in healthcare research
99
Method
This was a small opportunistic study comprising three focus groups consisting of a
mixture of health and social care professionals. The health professionals were
drawn from primary care, and included general practitioners and practice and
district nurses, as well as representatives of the professions allied to medicine. A
grounded theory approach to data analysis was adopted (Creswell, 1988).
(
Main results
Two main themes emerged. These related to views on evidence-based practice,
and perceived barriers to working together. Under the latter theme, three sub
themes were identified, namely operational matters, the metaphysics of health
care, and philosophical differences in the conceptual modelling of health and
social care.
The participants said that the prospect of health and social care professionals
working together more closely was welcomed, and that evidence-based practice
should be encouraged and financially supported. There were difficulties in
applying evidence derived from population studies to individuals, and also in
applying evidence from a study undertaken in a locality that may be quite
different from the participants' home territory. Evidence-based practice might
also present a threat. For example, one healthcare professional remarked, 'Do I
really want to accept I've been doing something futile for 20 years?'.
Although the participants regarded joint working positively, they bemoaned
the difficulties of getting 'the right people round the table.' All three groups
agreed that the public's rising expectations of health and social care services, and
their increasing demands, were stressful. In addition, patients presented not with
discreetly packaged issues, but with undifferentiated problems whose origin lay to
a large extent in their unique personal and social circumstances.
Two of the groups identified serious conceptual difficulties in working together,
based on a perceived dichotomy of approach between health and social care: 'The
basis of the assessment is different - GPs think in terms of treatment' and 'We've
got the problem of them working most specifically with the social model as
opposed to the medical model.'
All three groups explored perceived differences between the medical and social
models of practice. The data are best illustrated by the following extract:
'Medicine is much more easily definable . . . with medication you either take it
or you don't; with social services you are talking about people and there are an
infinite variety of variables.' This perceived difference in professional approach
extended not only to the assessment of individual cases, but also to the
application of research evidence, and to treatment decisions generally.
Conclusion
There is both a general willingness among these health and social care profes
sionals to work together, and enthusiasm for evidence-based practice. Factors
over and above the research evidence have an impact on the willingness of the
groups to work together. These include metaphysical factors, by which is meant
the personal unhappiness or insuperable disadvantage of some patients. Distinct
differences existed in the ways in which each of the two groups (health and social
care) conceptualised the other's discipline.
M
I
T
I
100
Complexity in primary care
Commentary on Joint Working Paper
This is clearly a small and opportunistic study, and as such its findings must be
viewed as a snapshot of how a mixed group of health and social care professionals
express then views about the nature of their professional practice, rather than
anything more definitive. On the other hand, it is extremely unusual for such
diverse groups to come together for such a discussion - sufficiently rare for the
editor of this respected peer-reviewed journal of social care to be anxious to
present the findings. It seemed to be too good an opportunity to miss.
he mam concern about the publication of this paper was the shortage of
i ona space afforded by the editor. Clearly this constrained the way in which
we could present our data and methods. It also had the effect of abbreviating
some of the results sections in a way that reduced the impact of three key themes
These were to do with context, and with what we called the metaphysical and
philosophical levels of interpretation. In order to gain a clearer understanding of
ow a second-level analysis from complexity might proceed. I have gone back to
the original data, to expand on these themes, before commenting on their
relevance.
Context
The original data set contained a separate category entitled 'geography'. This
category contained participants' comments on their own locality - that is the
physical context in which they operated. The importance of this category lay in
the way in which the participants appeared to be saying that where they worked
affected how they worked. As such this seemed to be significant, as it limited the
way in which they approached the whole notion of applying the scientific
method (m the shape of evidence-based medicine) to their routine working
P«ctlce- •n the final draft only six lines were retained by the editor in this area
egory °Wm8 Paragraph re’Presents the relevant original data under this catThe groups spent some time describing how their own working practice was
affected by the specific geographical characteristics of their localities. For example,
one seaside town appeared to be 'detached' or 'peripheral' and 'very rural and
spread out'. Poor transport connections were blamed for creating this 'out here'
ee ing. Demography also had an effect on working practice. These localities had a
ugh rate of unemployment that was weakened by a seasonal influx of a
population seeking part-time work. This seasonal influx was not wholly wel
comed for other reasons: 'the influx of some of our visitors leads to drug and
alcoho problems'. One participant summed up these themes as follows: 'A lot of
the links we have to make tend to be in [name of town], a lot of resources are
centralised that actually is quite a barrier when it comes to working at a local
level.""
’
Thus
described, the locality context appeared to have an impact
impact not
not simply
simply
on participants'
own
work,
but
also
*
----, —v
on their enthusiasm for working together.
Metaphysical level
Although this heading was retained by the editor, the final paper omitted what
seemed, from the poignancy of the data, to be an important reference to the very
p"Vay Vlew ParticiPants held about the nature of their own work, and the way in
which working models, which gave rise to their professional boundaries,
Using complexity principles in healthcare research
101
appeared at times to be confusing. The following extract presents the original data
in this area:
A set of data was identified under the broad heading of working together which
seemed to transcend mere day-to-day operational issues and dwelt upon the
inherent nature of the problems faced by these professional groups which in
itself might render joint working difficult. Some of these were generic. 'We all
carry a fantasy of what's going on in our heads', observed one participant.
Another felt the fact that 'we are all faced with personal unhappiness and
distress' could be dispiriting and as a result dampen enthusiasm for col
laboration. The group also recognised that professional demarcation, in terms
of responsibility, did not always appear sensible: a dichotomy of medical versus
social model emerged. 'Take baths, for example: is this a medical or social
bath?’
Philosophical level
Although the final paper retained almost all of the original interpretation of the
participants' data, the following section highlights the consistency of this theme
of perceived difference in the approach to assessing patients held by these groups
of combined health and social care professionals:
This perceived difference in professional approach extended not only to the
assessment of individual cases but to treatment decisions generally and to the
application of research evidence. Tt would be interesting to see, working with
the social model or the medical model, what those differences are.' One
participant argued that the medical model takes away responsibility, 'you
know using the medical evidence of the clinical diagnosis', while the social
model was characterised ‘by working with the person.' Judging by the amount
of data collected under this heading, this theme was seen as central.
-
J
if ‘
Summary of Joint Working Paper
Two main observations can be made from this small study. First, given the
opportunity, these health and social care professionals were able to express
profound views about the nature of their responsibility, which went beyond
the simple application of clinical evidence to reflect on the impact of metaphysics
- patients' unhappiness, suffering and despair - on their enthusiasm for working
together. Secondly, they held contrasting views about the nature of medical and
social care, and they framed this comparison with fairly adversarial metaphors 'the problem' or 'polarisation'.
Paper 2
A preliminary study of the decision-making process within general practice
Published in 2000 in Family Practice. 17: 428-9.
Background
This study sought to establish an empirical basis for the notion of personal
significance, described in Chapter 3, by exploring the factors that contribute to
■
pHG-lOO
lO'tV'J
poG
)
I 02
Complexity in primary care
the process of decision making within general practice, over and above evidence
based information. Awareness of the latest scientific evidence, as well as the
ability to critically appraise literature and assess its generalisability, have been
identified as integral to the practice of evidence-based medicine. However, the
evaluation of evidence in general practice is often illogical and irrational
(Sweeney, 1996), and it cannot be assumed that GPs practise the principles
underpinning evidence-based medicine in their decision making.
Method
A qualitative study was conducted using semi-structured interviews on a
purposive sample of five GPs, based in south-west England. Each interview was
tape-recorded and transcribed verbatim.
Main results
Six themes emerged from the data, namely practitioner, patient, practitioner
patient relationship, verbal and non-verbal communication, evidence-based
medicine and external factors. These are addressed in sequence.
All of the practitioners described how previous clinical experiences and their
own clinical beliefs had an impact on clinical decision making: 'doctors also have
their own philosophy of health' (Interview 1) and 'it's the things that go wrong
that imprint on your memory . . .' (Interview 3). Equally, the participants
recognised that understanding patients' cultural beliefs, background and attitudes
is integral to the decision-making process. One general practitioner remarked
'You have to know where the patients are coming from . . . and what their beliefs
are' (Interview 1). These two features constituted the bedrock of the practitioner
patient relationship, which the participants described as constantly evolving, and
for the sake of which the practitioners were at times able to bow to the patient's
expectations for the sake of maintaining good relations: 'the nature of the
relationship is one that continues and goes on and there may be far more
important issues coming up than this trivial issue of whether or not you prescribe
penicillin . . .' (Interview 1).
To develop the relationship, the practitioners acknowledged
acknowledged the
the importance
importance of
of
communication, not only in terms of their own language - 'you've got to pitch
what you say at a level that the patient will understand . . .' (Interview 3) - but
also by sensitively observing sensitive cues from patients - 'patients do give quite
strong messages, without necessarily expressing them verbally, about what they
want (Interview 1). As far as evidence-based medicine was concerned, the
participating doctors accepted its contribution in a qualified way: 'EBM measures
the things that can be measured . . .' (Interview 2). However, external factors
influenced the decision-making process as well: 'GPs are conscious of society's
views, but particularly cost' (Interview 3), 'time is critical, we don't have very
long, that's the problem' (Interview 5) and 'the media are more powerful than
anything else'(Interview 5).
Conclusion
Consideration needs to be given to the way in which the nature of the decision
making process impacts on the way that 'evidence' is constructed and promoted
in general practice.
Using complexity principles in healthcare research
103
Commentary on the preliminary study of the decision-making process
within general practice
As its title suggests, this was a preliminary study, and its results and contribution
should be considered alongside the next paper in this chapter (Freeman and
Sweeney, 2001). The enquiry was sparked off by a discussion with the co-author
about the meaning of the term 'linear' when applied to explanatory models.
Evidence-based medicine had been criticised for being algorithmic. There was, it
had been argued, a linear sequence in its five-stage approach. But was this, we
thought, the way it actually happened in practice? How did decisions evolve
within consultations? If evidence-based medicine was linear, this assumed the
existence of an opposite, non-linear approach. What, then, might be the
attributes of such a non-linear system? This was the discussion which provided
the impetus for this small study.
Comments on the data
As the editor of the journal in which this paper was published presented it as a
short report, much of the original data, as well as clarification of the methods, was
excluded. As the themes in this short study are relevant to the argument, I have
gone back to the original data to report them in greater depth. The headings that
were used in the published paper have been retained.
I
I
Practitioner
When using the phrase 'philosophy of health', the practitioners in this study
appeared to mean their own values - 'you make value judgments all the time'
(Interview 5) - and this sometimes conflicted with professionals' values. One
practitioner said 'I'm not quite sure where to draw the line between what I
believe in and what is acceptable as a GP' (Interview 2). Several participants
stressed the importance of previous clinical experiences as an influence on the
way in which they consulted: 'If you've had a patient who has a problem, and
dies from a stroke, a bleed due to warfarin, then you're going to be cagey about
putting other people on to it' (Interview 4).
Patient
The implications of accommodating the patient's health beliefs, referred to in the
published text, were spelled out in several interviews. 'The evidence is always
tempered by the patient and the doctor,' reflected one general practitioner, 'but
particularly by the patient' (Interview 3). Sometimes this might mean that robust
research evidence would be ignored 'if the patients don't believe it works'
(Interview 3). Another practitioner summed up the tension as follows: 'The
evidence is cold and we are about managing patients' (Interview 1).
Practitioner-patient relationship
The use of the word 'Technicolor' to describe the doctor's relationship with a
patient is intriguing, and suggests an entity that is rich, multi-faceted and difficult
to quantify. 'There is some logic in it,' one GP said, 'but you certainly can't
quantify it and write it down in a flow diagram' (Interview 2).
UN
•i
I
J
104
Complexity in pnmary care
Verbal and non-verbal communication
i'
I
)
In addition to recognising the importance of semantics in conveying important
messages and receiving verbal clues from patients - referred to in the published
text, the general practitioners stressed the importance of reciprocity in these
consultations by adding, for example, that 'You can make a remark or a question
which shows that you're on line, then you're away . . . similarly if you ask a
question which shows you're not, they'll stay dumb and blocked' (Interview 4).
atients pick up on doctors' 'styles' (Interview 4) and, as one doctor admitted
they can select themselves out and shop around ... I do appreciate that there are
Untervi^w 4)tS Wh° 316
SUitable for my aPProach than are suitable'
Evidence-based medicine
A perceived advantage of evidence-based medicine for the doctors in this study
was that it provided a yardstick to use when coming to an agreement about when
a new piece of evidence should change their own, and their partnership's,
practice: 'we would normally need something of the level of ... um ... a BMJ
article or Lancet editorial' (Interview 3). In several interviews, the doctors at this
point often reflected on the tension between clinical significance and personal
significance. One doctor commented about patients who might say: 'Oh, he's
taken me off my X drug which I've been perfectly happy with . . .' to a new
(evidence-based drug) which 'provokes anxiety and people feel nervous about it
■ ■ . saying look at all these side-effects' (Interview I). An interesting distinction
was made by one doctor between vigorously treating risk factors and relieving
suffering: 'when somebody is having a clear-cut piece of suffering which we're
trying to relieve and the further away we get from that the harder it is to fit the
evidence with a person and a practical situation' (Interview 4).
Summary
The central observation from this study relates to the importance of the evolving
relationships that doctors elaborate with their patients, and vice versa (as we are
reminded by one participant in this study). This idea of two people contributing to
and being mutually influenced by their evolving relationship echoes the defini
tion of complex responsive processes that was developed in Chapter 5 I shall
return to this later, in the second-level analysis of all four papers in the final
section of this chapter.
Paper 3
Why general practitioners do not implement evidence: a qualitative study
Published in 2001 in BMJ. 323: 1100-14.
Background
This paper extends the line of enquiry that was initiated in the previous paper by
exploring, through the medium of Balint groups, the reasons why general
practitioners do not always implement best evidence. The aim was to supplement
Using complexity principles in healthcare research
105
the empirical basis supporting the relevance of personal significance to clinical
practice, by reflecting on how doctors crafted their decisions during consultations.
Method
A total of 19 general practitioners took part in 13 Balint group meetings, for 11 of
which data were available for analysis.
Main results
The process of implementing clinical evidence is affected by the personal and
professional experiences of the doctor. For example, one participant remarked
that: T actually had two 50-year-olds who had strokes from atrial fibrillation
because they didn't get warfarin . . . that really hit me.' Others described how,
having initially been less than enthusiastic about anticoagulating patients with
atrial fibrillation, subsequent positive clinical experience could change their view
- T'm back on it.' The main point of this theme is expressed thus: 'We are
influenced at least as much, if not more, by the experiences of individual patients
as we are by the evidence.' This suggests that the relationship that the doctor has
with individual patients also affects the process of implementing evidence, a
second strong theme that emerged from the data. 'Even if the evidence was
extremely good,' one general practitioner said, 'most of us would only ever
interpret it in the context of the patient.' This was not a one-way process, and
patients could, at times, and as a function of their relationship with the doctor,
influence the doctor's decisions. This is summed up by the following quote: 'Well,
he's a farmer, so every time he calls the vet he gets antibiotics.'
When describing the clinical applications of evidence-based medicine, these
participants depicted a tension between primary and secondary care. The doctors
thought that specialists approach evidence-based practice differently. For
example, they do not realise how tricky it is to control some common conditions.
'You get stroppy letters from the clinic saying your patient's blood pressure is still
160,' said one participant, 'and 1 go . . . yes, yes, I know. You feel under pressure
from the guidelines, but you know it's not for want of trying.'
In addition, the practitioner's feelings - not just about their relationships with
patients, but also about the evidence itself - modify the way in which clinical
evidence is applied. The very presence of 'evidence' - for example, in key journals
- could 'make me feel anxious,' said one doctor, with the result that: 'With me
messing about with his medication and trying to practise evidence-based
medicine, I found it was making him [the patient] feel more anxious.'
When discussing the process of coming to a decision, the doctors in this study
clearly held the view that their choice of words in consultations could sway
patients to accept or reject clinical evidence. Doctors realise this, and can use it to
pre-empt patients' decisions. 'It's how you put it over,' said one group member. 'It
depends on how you feed information to people,' said another. Finally, although
the evidence might be strong, and its relevance in certain cases clear, logistics
could still act as a barrier. Referring to the anxiety about a patient bleeding while
on warfarin (a recognised side-effect of this drug), one practitioner remarked that:
'It's not a minor bleed if your patient is 30 miles from the nearest transfusion
service.'
?
106
Complexity in primary care
Conclusion
These general practitioner participants appeared to act as a conduit within the
consultation and to regard clinical evidence as a square peg to be fitted in the
anda^aptive1
PatientS Ufe' The process of implementation is complex, fluid
Commentary on why general practitioners do not implement the evidenceevidence:
a qualitative study
This study was prompted by two conversations. The first arose from a discussion
about the study by McColl et al. (1998) referred to in the introductory section of
the paper. This paper had suggested that there may be unique reasons within a
general practice setting that might constitute barriers to the implementation of
evidence-based medicine. The second conversation, which followed from this
was a further exploration of the notion of personal significance referred to earlier
there were unique barriers to implementing evidence-based medicine, what
might they be? Could they be partly explained by the idea underpinning personal
significance, namely the opaque activity of transferring ideas from professional to
pauent, and vice versa? The area remained under-researched, and this study held
out the possibility of extending the findings of the preliminary study published in
annmreV1°US
presented as PaPer 2 in this chapter (Mears and Sweeney.
Commentary on the data
The data that appeared in the final published draft of this paper provided a fair
and accurate reflection of the analysis of the original focus groups' transcripts
The editors of the BMJ requested a redrafting of the discussion section of the
paper, but were happy with the presentation of the results as we had initially
suggested. No further elaboration of the results is required here.
'
Summary
The mam findings of this study extend and support the relevance of personal
significance to clinical practice. One observes the practitioners distilling clinical
evidence internal y, as it were, weighing it up against their own experience both
personal and professional. One senses that, having come to a view themselves
the doctors know that they can - and indeed have to - 'sell' that view to their
bo h n 't,n
u°nteXt 01 311 e,aborate' fluid- and evolving relationship to which
both parties - the patient and the doctor - contribute. And all of this proceeds
°f anX1Cty Created by the very Presence of the evidence (in
all the BMJs, all the rags’), as well as the logistical challenges inherent in its
anflvsTamhe11' JhefSe, obs^rvations wil’ be developed during the second-level
analysis at the end of this chapter.
Using complexity principles in healthcare research
107
Paper 4
A comparison of professionals' and patients’ understanding of asthma:
evidence of emerging dualities?
Published in 2001 in Journal of Medical Ethics: Medical Humanities. 27: 20-25.6
Background
The purpose of this paper was to extend the debate about the nature of medicine s
explanatory model, which had been developed in the first three papers, into the
patient's domain. The first paper tentatively explored whether, when discussing
the nature of clinical practice in the context of joint working, health and social
care professionals expressed some views about the deeper nature of clinical
practice and the context in which they practised. The next two papers collated
empirical data about 'personal significance', the nature of which is described in
Chapter 5. This final paper focuses on another aspect of personal significance,
which is enshrined in the reciprocity that lies at the heart of its definition. At stake
is the debate about modelling a disease - in this case asthma - on the basis of a
biomedical model predicated on scientific positivism. If the contemporary ex
planatory model was to dominate clinical discussions, one might expect a
reasonable degree of sharedness of that understanding, the absence of which
could constitute an impediment to joint 'shared' decision making. This is what
this paper explores.
Method
Two sets of focus groups were convened in parallel, four consisting of profes
sionals (doctors and nurses) and four comprising patients with asthma. The
professional groups consisted of one separate group each of specialist doctors,
secondary care nurses, general practitioners and practice nurses. Patients with
asthma were identified from general practice disease registers. To obtain a
sufficient spread of patients with the type of asthma that is seen routinely in
general practice, the sampling frame was stratified by age and by use of inhaled
steroids (which was used as a proxy indicator of asthma severity).
Main results
The healthcare professionals and patients who participated in this study showed
broad agreement in their explanations of the aetiology and drug treatment of
asthma. However, the data suggest a lack of congruence in the development of
treatment strategies and locus of control. In summary, the doctors and nurses
constructed management plans prospectively - based, for example, on their
theoretical knowledge of the effects of inhaled steroids over time. When referring
to Ventolin and Becotide (standard symptomatic and prophylactic treatments,
respectively, for asthma), one general practitioner said: 'one makes you better at
the time, the other keeps you better for tomorrow.' The patients formulated such
plans retrospectively, based on their previous experience of various treatment
modalities. For example, a younger male patient who frequently used an inhaler
6 Authors: KG Sweeney, K Edwards, J Stead and D Halpin, University of Exeter and North
and East Devon Health Authority.
w
3
I 08
Complexity in primary care
said: T take it if I get bad or get a cold ... on to my browns three or four days then
it works. All the bad stuff comes towards it and bounces off it or gets eaten
possibly. Maybe absorbed.' This sometimes led individuals to act against medical
advice to use Becotide continuously: 'When you haven't got a tight chest you
haven t got asthma. I just forget to take Becotide' (older male patient who was an
infrequent inhaler user).
The healthcare professionals and patients in this study used different metaphors
to conceptualise asthma. The former group more frequently used metaphors that
evoked ongoing processes. Consider, for example, this doctor's comment: 'it
involves probably formal components of your inflammatory pathway, so prob
ably certain parts can be switched off. . . some are more prominent than others.'
Compare that with the following lay participants' models, which represent the
tendency of the patients to visualise the chest (in their use of metaphor) as a static
container, emptying and filling throughout the course of the disease: 'It's like a
windsock' (younger, infrequent user) and 'hubbly-bubbly pipes' (younger.
frequent user).
I
I
■
'■I
It
Conclusion
The analysis supports the view that there is an <epistemological
’
\
2;;
difference
between
doctors and patients in this context. We postulate that the two) groups professionals and patients - draw on different types of knowledge wheni constructing their model of asthma. Doctors, it is asserted, draw more on theoretical
knowledge, whereas patients make use (of" their
’
lived experience, drawing on
’ ;e.'
what Piaget (1932) calls 'figurative knowledg<
Comment on the data
Very few constraints were placed on the presentation of the data by the editors of
this journal. The published paper was 4500 words in length.
Summary
I
I
There does not appear to be a single, uniform conceptualisation of asthma that
was shared and explored by the professionals and patients in this study. If the
biomedical model, in this case of asthma, predicated on scientific positivism, was
indeed hegemonic, and constituted the dominant currency in discussions, one
might reasonably expect, if not an equivalence, at least a conspicuous degree of
overlap. We observe here an ongoing tension between two different types of
knowledge - referred to as operational (theory based) and figurative (derived
from lived experience) - which, the paper argues, suggests that the two groups
are drawing on different epistemological frameworks. This theme will be
elaborated when the second-level analysis reflects on the relevance of complex
responsive processes to clinical conversations.
Second-level analysis of the papers
In this section, the main themes distilled from the four papers presented in this
chapter are analysed from the perspective of complex adaptive systems. The five
key features of complex adaptive systems, which were set out earlier in this
chapter, are held as the compass points for the analytical framework.
Using complexity principles in healthcare research
109
■«.-awi> 'an
Preliminary observations on the data collection in the four papers
As the four papers presented in this chapter represent 'one-off' data collection
exercises - snapshots, as it were, of the experiences reported - they are unlikely to
yield primary evidence of features of complex systems that evolve over time.
Time-series data, which collect evidence at various points throughout the
evolution of a system, would be more likely to achieve that (Holt, 2002b).
Thus one would not expect to see primary evidence of self-organisation and co
evolution which, de facto, occur over time. Participants may refer to these
features of complex systems when they describe the context in which they
were operating, thus providing some indirect evidence of their influence. The
other two features of complex systems, namely receptive context (also termed
enabling context; Mitelton-Kelly, 2003) and complex responsive processes, are
more likely to be identified.
Given the criteria for a complex system that were set out in Chapter 5, one can
then postulate that there should be evidence of the presence of an enabling
framework in systems which evolved creatively, or evidence of its absence in
more dysfunctional systems. And if complex responsive processes are funda
mental to human adaptive systems, one should discern evidence of these, or at
least reference to them. Thus this second-level analysis is undertaken tentatively
in relation to these four papers, as the research that they report was not
conducted primarily with the two-level analysis in mind (Sweeney, 2003b).
5
£
Receptive context
Consider the definition of receptive context that was presented in Chapter 5. It
embraced the following notions:
I
• an enabling framework, or an infrastructure of communicability
• a set of conditions that have the potential to facilitate the development of
complex conversations and actions, whose patterning can, over time, consti
tute a complex system
• a set of values through which coherent action can be expressed
• the presence of leadership to initiate complex responsive processes
• the potential to engage other agents to co-create and adapt the system.
In Paper 1 (on joint working), we are told in the background information that the
study day on evidence-based medicine was held on behalf of a community in
which there was little evidence of joint working in the context of evidence-based
medicine. We learn that the participants bemoan their inability to secure meet
ings at which all the 'right people' turn up, and that their quite distinct - indeed
opposing - views on what constitutes the medical and social models reduced their
enthusiasm for joint working. So did the physical location, which they described
as 'out here', evoking a feeling of being marginalised, and peripheral to a more
central place where financial decisions were made. Some of the participants saw
evidence-based practice, which was considered as one mechanism for facilitating
joint working, as a threat, making them realise that they had been 'doing
something futile for 20 years.'
This constitutes evidence which suggests the absence of a receptive context, or
that the participants were not operating within an enabling framework. I
II
I I0
Complexity in pnmary care
postulate the existence of a link between the absence of such a framework and
the absence of coherent action, leadership to encourage it, and other agents
enthusiastically co-creating it. This hypothesis is supported by the evidence
describing the barriers to joint working, and makes sense in the light of the
reason for holding the study day in the first place. Given the criteria for a complex
system that were described in Chapter 5, one concludes that the system in which
the participants in this study operated - a health and social care 'community' was not complex or adaptive.
A similar analysis can be applied to the background to Paper 4 (on asthma) The
background to this paper involves the recognition that, despite a clearer under
standing of the pathogenesis of asthma and an increase in services provided to
patients with the condition, morbidity and mortality remain high. We learn in the
paper that, although both professionals and patients agree about the basics of
what causes the disease, there is less congruence of opinion about how to manage
it, particularly with regard to the role of inhaled steroids. Locus of control is also
disputed, with a tacit assumption among many of the healthcare professionals
t lat they can delegate control to patients (implying its prior location within their
sphere of influence). The patients' response is typified by the comment made by
one sufferer that: T know my asthma better than anybody.' The analysis at the
end of the paper supports the deployment of different epistemological frame
works by the two participating groups.
There is no direct evidence in this paper that links morbidity in these patients to
the absence of a shared understanding of asthma or its management, but one can
speculate about such an association. Accepting the features of receptive context
described above, one r
' conclude
’ ’ that
’
is ’led’ to
there is evidence of some but not all
the features of an (enabling
’ ”
framework. The set of conditions, namely the
professional-patient interactions
—in the paper, have at least the
------- described
potential to construct an tenabling
’ ’
'
framework,
and one can infer from the
participants' reports that the two groups shared some values - f._
for example, the
desire to manage asthma successfully. However, against this the presence of
_f two
different models for the disease, reflecting the deployment of operational and
figurative knowledge, does not support the presence of a shared, complementary
set of values through which complex conversations may lead to coherent action
And there is no evidence in the study that the interactions between the
professionals and the patients satisfied the description of complex responsive
responses. The nature of any leadership associated with the initiation of coherent
action is disputed. Who is to lead - the patient or the doctor? Thus one can
conclude that although there is some evidence of some of the features of a
receptive context, one cannot confidently discern its presence and its influence
on the evolution of the system, which consequently cannot be described as
complex or adaptive on the basis of this evidence.
Second-level analysis of Papers 2 and 3
In this section, the second-level analysis of Papers 2 and 3 focuses on the three
features that together constitute the minimum conditions necessary for a system
to be considered complex and adaptive.
J
Using complexity principles in healthcare research
I I I
Complex responsive processes in Papers 2 and 3
The definition of a complex responsive process that is offered by Stacey (2001)
suggests that:
• it involves interaction through complex communication
• one participant issues a gesture, which in turn calls forth a response from the
other, in an iterative, interactive process
• this interaction is self-organising, and has the property of emergent coherence
• the nature of such processes has the potential to be transformational. In
participating in a complex responsive process, the conversation is changed,
each participant is changed, the nature of their relationship can change, and by
a ripple effect, the nature of the participants' relationship with the larger
system can change
• these processes are inherently unpredictable.
There is some evidence to support the presence of complex responsive processes
in both Papers 2 and 3, which for the purposes of this second-level analysis are
considered together, as they explore similar systems, namely the doctor-patient
dyad. The notion of a gesture calling forth a response is supported by the
description of conversations by one of the participants in Paper 2, who commen
ted that: 'You can make a remark, or a question which shows you're on line, then
you're away . . . similarly if you ask a question which shows you're not, they'll
stay dumb and blocked.' Indeed, patients 'shop around' to select doctors with
whom they can conduct these delicate exchanges of gesture and response. The
doctors in this study agreed that 'you have to know where patients are coming
from, what their beliefs are . . .'. In the context of complex responsive processes,
this is a prerequisite of an enabling framework within which such creative
conversations can take place. These relationships, expressed through complex
conversations, 'continue and go on', and need preserving in order to mature hence the flexibility to concede on trivial issues, such as the occasional penicillin
prescription.
Did the participants in these two studies declare that they might be changed
themselves in the process of these conversations? It seems so. 'We are influenced
at least as much, if not more,' asserted one general practitioner in Paper 3, 'by the
experience of individual patients as we are by the evidence.' Did the nature of
their conversation change in the course of their evolution? The evidence does not
dispute this. One doctor, we learn, built up the relationship with the patient by
initially not following the guidelines, and then, when they were in a position of
greater trust, they were able to implement the guidelines. T have followed the
guidelines, of course,' this doctor stated, 'but in a sneaky way, and it's taken about
three months to do it.' Accordingly, one can cautiously conclude that the
interactions which these participants were describing were complex responsive
processes.
Receptive context and self-organisation in Papers 2 and 3
Is there evidence from the first-level analysis of these two papers to support the
presence of a receptive context? In Table 7.1 the five features of receptive context
are shown in the left-hand column, with a second-level commentary for each
feature in the right-hand column.
2
I 12
Complexity in primary care
In order to qualify as a complex and adaptive system, we have agreed that three
features should be evidenced, namely a receptive context, complex responsive
processes and self-organisation. When analysing these papers, which reflect 'oneoff snapshots of a system, it will be difficult to discern evidence of the last of this
triad, namely self-organisation. However, some preliminary observations can be
maae.
Self-organisation is about the creation of coherent patterns of behaviour. One
can speculate from the doctors' descriptions in these two papers that there are
some patterns to the behaviours that were acted out in their consultations with
pa Liciits.
Table 7.1 Receptive context in Papers 2 and 3
Feature of receptive context
Comment on evidence in Papers 2 and 3
Infrastructure of communicability
The nature of the consultations to which they refer
suggests the presence of an appropriate infrastructure
in which the doctor-patient dyad has the potential to
evolve
A set of conditions to promote
complex responsive processes
There is some indirect evidence from the doctors of
the influence of their previous experiences on
evolving relationships. They also refer to patients'
beliefs as an important component of their
interaction
A set of values for coherent action The data imply an aspiration that the doctors provide,
to be expressed
and the patients receive, the best possible care. In
their interaction with patients, the doctors refer to
their own philosophy of health, and the way patients
develop their own beliefs about health and disease.
Although there is no direct evidence linking these to
coherent action, the conditions for such action may
be said to be present
The presence of leadership to
There is indirect evidence of this. The doctors
initiate complex responsive
appeared to be disinclined to apply the rules of
processes
evidence-based medicine slavishly, and they
acknowledged that sometimes patients lead (e.g. in
cases where the patient doesn't believe that a
treatment will work). This evokes the notion of
dispersed leadership described by Durie et al. (2004)
The potential to engage other
agents to co-create and adapt the
system
One can infer that other agents influence the nature
of the interactions between doctors and participants.
The doctors refer to 'society', and one can assume
that the patients' beliefs are created within their own
life worlds in which they interact with significant
others
Thus they talk of 'knowing where the patients are
evidence being 'tempered " by the patient 'and “the Tctor.'"one panicipam
and the doctor.'
escribes implementing guidelines in a 'sneaky' way, which implies some form
oi pattern in a series of interactions that are spread over time. More broadly, the
Using complexity principles in healthcare research
113
doctors refer to a continuously changing, moulding and evolving relationship
with their patients, whose patterning allows some decisions to be taken which,
although not technically justifiable (like the prescription of an antibiotic), are
deployed in order to maintain the pattern of communication (the ongoing
relationship with that patient).
Thus one can conclude that there is some evidence to support the view that the
participants in these two studies did operate within a receptive context, and that
the nature of their interaction satisfied the definition of complex responsive
processes. The evidence for self-organisation is much less clear, is indirect, and is
based on inferences that were made on the basis of the participants' descriptions.
Summary of second-level analysis
It must be emphasised that this second-level analysis is undertaken with great
caution as, at the point of data collection, no attention was paid to the possibility
of collating material that could directly support (or dispute) the relevance of these
features of complex adaptive systems.
In Paper 1 (on joint working) there are enough data for it to be accepted that a
receptive context was not present, and that consequently the system under
scrutiny was not complex. One can speculate that the absence of a receptive
context might have been related to the reason for holding the study day to which
the participants contributed, namely the absence of good collaborative working
between local health and social care professionals. In Paper 4 (the asthma study)
only some of the features of a receptive context are supported, and it is not clear
that the nature of the interaction between the professional and patient groups
constituted a set of complex responsive processes. One could have collected data
that might have clarified the nature of these interactions, but the data as
presented do not allow this.
However, there is stronger although inconclusive evidence for the presence of
complex adaptive systems operating in the interactions described in Papers 2
and 3. The interactions appear to constitute complex responsive processes, and
there is some, albeit indirect evidence of a receptive context. Although there is
not clear prima facie evidence of self-organisation, one can cautiously make some
inferences about this from the descriptions given by the participating doctors. This
analysis permits further reflection on the notion of personal significance
(described in Chapter 3).
Given what is now understood about the nature and importance of complex
responsive processes, personal significance may actually consist, at its root, of
such complex processes. Personal significance can be considered as the emergent
property of the patterning of complex responsive processes. Overall, the analysis
supports the view that further research looking more directly for evidence of
complex systems might usefully be undertaken.
Summary: reflections on second-level analysis and the
direction of future research
The main purpose of exploring these four small studies was to illustrate how the
principles of complexity might be incorporated into a usable research method
ology, and to demonstrate what kind of conclusions one might draw using this
£
I I4
<
■
!
[
I
3
Complexity in primary care
method, and given the definitions of the features of complex and evolving
systems set out in Chapter 5. In short, it appears to be possible to deploy the
principles of complexity as an interpretive framework at a second level, once a
conventional first-level analysis of collated data has been undertaken using the
basic principles of qualitative research. Although there is always a danger of
reading into the data to gather evidence from this interpretive framework, this is a
danger inherent in any focused qualitative analysis. Marxist, feminist or racial
perspectives are perfectly acceptable frameworks to deploy in qualitative analysis,
and they run the same risk. In this example, one takes the principles of complex
systems and scrutinises the assembled themes for data which constitute evidence
that they were or were not in place. One cannot create evidence to support such
conclusions, but one can identify such evidence and comment on its merits.
So what kind of domains might benefit from the kind of second-level analysis
set out above? Inherent in the notion of complex adaptive systems is time - time
for the systems to evolve and mature, expressing, over time, characteristic
features through the patterning of iterative processes that give rise to the system's
emergent properties through self-organisation and co-evolution. This gives us a
clue as to which topics might benefit from such an analysis.
Accordingly, the principles of complexity may usefully be deployed in research
ing programmes of organisational change in healthcare systems. Complexity
could be used to understand the challenges to clinical practice, which involve a
cultural or work practice change - for example, the implementation of a National
Service Framework, or the introduction of a new technology. And a framework
based on the principles of complex adaptive systems could form the basis of an
assessment of major shifts in the deployment of human resources in healthcare for example, the impact of the work of general practitioners with a special
interest, or the impact of employing nurses as first-contact care clinicians in
primary care.
In addition, complexity may be a useful framework for gaining a clearer
understanding of patients' narratives (Snowden, 2002). Here the use of a
reflective diary which is updated iteratively over time, and which recounts in
detail the chronological sequence of events in an illness experience, could be
analysed from a complexity perspective in order to understand how the patient
interacted with the system, and how the system responded, affording opportun
ities to reflect on the evolution of the system over time. In future studies in which
the perspective of complexity is to be applied, it will be useful to collect
participants' stories, as they may shed light on the nature of the evolving
processes that constitute the system under scrutiny, and they may also illustrate
how interactions within those systems come to be patterned in a particular way.
In the previous chapter I discussed the opportunities afforded to healthcare by
the developments in non-linear mathematics for understanding change in some
physiological and pathological systems. Although the details of the mathematics
involved are beyond the scope of the non-specialist, it is appropriate to speculate
how our understanding of the evolution of pathological systems over time can
be assisted by these developments. It is likely, as Holt (2002a) argues, that
models of diabetes based on non-linear mathematics will help patients to
manage that condition better; and Tennison (2002) argues that we will be
able to understand the cyclical infectivity of some pathogens by drawing on
non-linear mathematics.
Using complexity principles in healthcare research
«i<li ir nrnm—iwiirmiiiiiMm iiiimrii.ii i j i
115
--n,—r- ■iT-u.iirrmiiiir:
In summary, the principles of complexity can assist healthcare research in
domains of transformational change, whether that change occurs in organ
isations, in cultural shifts in clinical practice, or in the narrative recounting of
illness experiences. So far as human physiological and pathological systems are
concerned, the ability to collate accurate numerical data over time opens the
interpretation of these systems to the benefits of non-linear mathematics. I shall
develop these themes in the next and final chapter.
r
I
?
i
Chapter 8
rxrr»-W^^-SSS«MW«WUWS
Complexity and medical practice:
prospects for the future
Introduction
There really is only one question to address when evaluating the potential
benefits of utilising the principles of complexity in healthcare: will this make
me a better doctor? Related to this is the form of the question at the organisational
level: will understanding complexity make ours a better health service? In turn,
this invites the same question at a clinical level: will this make us better
clinicians? This chapter will attempt to answer these questions.
Asking questions
What the preceding chapters have done is to ask questions about the questions
that doctors ask. First, we have asked how we know things in medicine. We have
an explanatory model, I have argued, that is rooted in scientific positivism. This
link between the explanatory model and its related world-view (or ontological
perspective) has been explored, to argue that a world-view is expressed in a
predilection for a certain type of knowledge, which we have called an epistemo
logical perspective. When we speculate that this triumvirate, ontological per
spective, epistemological framework and explanatory model might serve as a
generic model, we are permitted to ask whether there are other ways of knowing,
expressed in different types of knowledge, leading to distinct explanatory models.
I have argued that the naturalistic tradition is predicated on a world-view that
differs from scientific positivism and that deploys a socially constructed epistemo
logy, from which has arisen an explanatory model that concerns itself with
attitudes and beliefs, with intention and action. One way in which this is used in
healthcare is through the principles of qualitative research.
In Chapter 3 we reflected on the contemporary form of the explanatory model
in medicine, by setting out the principles of evidence-based medicine, and then
reminding ourselves of the criticisms that it had attracted. At the outset, one could
be forgiven for asking why a model which simply exhorted practitioners to apply
the best evidence to solve clinical questions, and in so doing equipped those same
practitioners with useful critical skills, could attract such an enthusiastic critique.
The answer lay in the nature of the critics' concerns. First, evidence-based
medicine involves an initial abstraction - an extraction of a suitable biomedical
question from the patient's narrative. If a story of sleeplessness, fear, uncertainty
and malaise is called 'anxiety', then the hounds are out on the course, and the
search for best evidence can proceed - we have a 'case' for valium. If a naturalistic
interpretation is applied, an array of personal, behavioural and structural
I I8
■
i!
I
!
Complexity in primary care
impediments to well-being is exposed, leading to a different strategy for resolu
tion. These interpretations are not mutually exclusive, and good clinicians will
pursue both. However, in so doing we have supported the key proposition in the
first part of this book, namely that there is more than one way of knowing. Good
clinicians will deploy two ontological views, and in turn will seek out the
preferred knowledge to populate each view.
Mrs B s consultation showed us how this can happen in the most mundane of
consultations. Here we had an intersection of ways of knowing, a clash of the
positivist and naturalistic ways of knowing. What happens, we can ask, when
such a clash occurs? Does one view dominate? If so, who decides which view does
dominate? And if some coalition of world-views is to be fashioned, by whom is
that coalition created? This raises two questions about the nature of these
common but profound consultations. First, where does the locus of control lie
in these consultations? Is it always driven by the doctor, an assumption driven by
a conventional understanding of the doctor as expert? Or is it more desirable to
see the role of the doctor as that of the informed servant, like the butler in the
Death of Ivan Illyich, supporting, informing and guiding the patient towards their
own decision which suits the particularities of the context rather than the
generalities of the evidence? The second question is when is enough, enough?
The fact that many individuals in middle or late-middle age are surviving what a
mere decade and a half ago would have been a fatal clinical event means that they
survive into later life, susceptible to a constellation of comorbid conditions which
develop as a consequence of their longevity. Mrs B is an example of this. Her
comorbidity was amenable to treatment of all of its constituent conditions (with
the possible exception of her macular degeneration). In that consultation, I
offered evidence to support tighter control of her blood pressure, her glycosylated
haemoglobin and her dyslipidaemia, and could have justified more active
intervention for her intercurrent depression and background arthropathy. In
her historic riposte, 'Jack's dead and the boys have gone', she asked the question
which will challenge the next generation of practitioners more than my own
generation: when is enough, enough? We are encountering a new kind of
coercive public health, where 'health' is seen as a minority sport. Around 90%
of the adult European population, we learn, is now considered to be 'at risk' of
cardiovascular disease. Virtually the entire continental population will be asked to
submit to monitoring, to vigilant personal surveillance and often to medical
intervention. It is when we articulate our response to that challenge that we must
begin to ask questions about the questions that doctors ask.
None of this is really new. Thus presented, in the context of complexity, the
observations are a restatement, and a contemporary attempt to address the ageold problem, for clinicians, of balancing the general with the particular. With the
advent of evidence-based medicine, clinicians were encouraged to interpolate
from population data to individuals. In so doing, however, we were at the mercy
of the ecological fallacy - assuming that any and all conclusions derived from
population data could be applied to all individuals in the data set. A number of
attempts have been made over the years to integrate these two seemingly
irreconcilable perspectives. Greenhalgh (2002) has called for intuition in general
practice. She describes this as the rapid, unconscious process of integrating
multiple complex pieces of data, where causality is set aside and selective
attention is paid to fine detail. Importantly, clinicians become increasingly skilled
Complexity and medical practice
I 19
at exercising intuition over time, re-establishing the importance of experience, or
what Aristotle would call phronesis (wisdom) in the process (Ross, 1988). More
recently, Gillies (2005) has described deliberative specification as the means ot
synthesising evidence and intuition. This allows Gillies to accord appropriate but
proportional weight to the evidence, while also paying attention to the unique
particularities of the patient's context. The contribution of the principles of
complexity is to extend our ability to theorise about how, in practice, this process
of deliberative speculation might express itself. In the notion of complex
responsive processes, we have a model of communication that confirms the
continuous, subtle co-evolution of the relationship between doctor and patient.
Note the use of the term 'co-evolution', which implies that both participants are
changed during the process of their interaction. Doctors are not immune from this
change as a result of their status, but are both contributors to and alumni of that
change process. It is in that iterative and recursive interaction, leading to change,
that we find the roots of the experience at the heart of intuition. Complexity
helps us not just by legitimising, theoretically, the need to reconcile evidence and
context, but also by showing how pervasive and inevitable, that process is. No
matter how good the evidence, or how sound the clinician's judgement, if the
patient is not engaged in a meaningful, changing interaction, they will count for
nothing. Ultimately, in the clinical encounter, it is the patient who decides what
the nature of the outcome will be, through either concordance or its absence. A
new metaphor is needed to describe this relationship between doctor and patient
- the doctor as the informed servant.
The principles of complexity, set alongside the analyses in the preceding
chapters, tell us one thing above all, namely that there are several ways of
knowing. Doctors, I argue, deploy not one but a whole range of ontological
perspectives when they consult, drawing on a range of epistemological frame
works to populate their explanations. In general practice, above all, it is important
not to let one perspective form the unique basis of the interaction. Attention to
particularities, recognition of uniqueness, and a greater willingness to accept
intuition as an asset that matures with age, combine to encourage flexibility in
the consultation and a recognition that we, as clinicians, are ourselves part of that
change process, and will be subtly affected by it. This is the framework within
which the three dimensions of generalism were presented in the introduction to
this book. Technical generalism is the conventional understanding of the skills of
the general practitioner - the ability to deploy diagnostic and managerial skills
over a wide range of the partialist specialisms of secondary care. Contextual
generalism conveys the importance of a second dimension, recognising that the
dynamic of a consultation can change profoundly and swiftly. Drawing on the
principles of complexity, we can theorise that it is in recursive interaction of
complex responsive processes within the consultation that such contextual
changes occur, and are recognised and legitimised. Now we can understand
that evidentiary generalism is a third dimension which is implied by the first two,
namely the need to understand that when context shifts in a consultation, the
index paradigm and its related epistemological framework shift, too. The general
ist, drawing on the skills of evidentiary generalism, recognises the legitimacy of
utilising a range of evidentiary frameworks, will come to an understanding with
the patient about the balance of the biomedical with the biographical, and will be
prepared to accept, in certain circumstances, the supremacy of the latter over the
122
Complexity in primary care
analysis helps us understand why change imposed from outside a system is often
unsuccessful. Unless those participating in a system first accept that things have to
change, and secondly discover for themselves how change might occur, the
change process will not constitute a visceral part of their interaction. Think of
how 'gaming' happens in organisations - the response to change which accom
modates it, according to some rules of engagement, but with no internal accept
ance of its need or value.
A second insight that is provided by the principles of complex systems involves
the notion of 'adjacent possibles' (Durie R, 2004, personal communication).
Adjacent possibles are notional 'spaces' in which change can occur, as a
consequence of interaction between the agents in a system, leading to self
organisation. These spaces can be said to be 'adjacent' in that they are within
organisational reach for the participants, that is not too different to destabilise the
system, but sufficiently different to encourage change, initially usually on a small
scale. Such changes, if successful, can then be patterned over the system to create
change on a wider scale. The opposite of this is also true - if they are unsuccessful,
but on a small scale, they do not destabilise the system, although they don't
change it either. Consider, as a theoretical example, a healthcare community that
wants to set up a primary-care-based, intermediate service for a chronic disease,
such as diabetes. The idea of adjacent possibles suggests that such a change should
be sufficiently within the organisation 'reach' of all the significant participants in
the system, but. sufficiently different to constitute a change (in the form of an
improvement) in the provision when established. In planning such a change
process, one must first accept the possibility of failure, as unless the system is
coerced into changing, the agents will, through their own interaction, determine
the outcome of the process themselves. Secondly, one sees the need to engage all
potentially significant participants early on, if only to initiate the interaction, and
hopefully to assist in the co-creation of some complex responsive processes,
through the sharing of common values and aspirations, which constitute the
receptive context of the system at its outset. Any programme of change should
then be small enough to be discernibly different, while not destabilising the
system if it fails. In the example of an enhanced diabetes service, this might look
like joint working between primary and secondary care providers in a locality, or
through a cluster of general practices, which, if successful, could be patterned
subsequently through the wider health and social care community. The change
would then be 'adjacent' organisationally, and 'possible' inter-professionally.
What would not constitute an adjacent possible would be the unilateral establish
ment of a stand-alone intermediate service, no matter how well provided, as it
would have failed to engage the key players, and it would run the risk of failing to
be patterned - that is, accepted by the local community. One can theorise that, in
the setting up of these adjacent possibles, the planning might have to be quite
linear. In the diabetes example, this might involve the decision to set up the
service to serve, say, a specific number of patients over a specific, short period of
time, with an agreed set of actions to be provided and reviewed - for example, to
manage 30 patients from three practices for one year (an agreed interval for a full
diabetic review) and scrutinise the clinical and patient-based outcomes. At that
point, if successful, the patterning of the change, in the form of its wider adoption
by the whole community, might flow in a less controlled, non-linear way,
allowing the participants to self-organise around a small number of agreed
Complexity and medical practice
I 23
strategic values. It might also co-evolve, as needed, with systems either nested
within that community (other general practices) or outside the immediate system
(national bodies or other healthcare communities).
Summary
The main proposition in this book is about ways of knowing. I have argued that,
by exploring the glorious intellectual history of western civilisation, one discerns
a trend which gave rise to the scientific positivism at the heart of the clinical
method, and a parallel trend, described as naturalistic, from which a more
grounded understanding of the human condition arose, and out of which the
principles of what we now call qualitative research emerged. In the last third of
the twentieth century, many of the fundamental sciences - thermodynamics and
biology in particular - revised their explanatory models to accommodate the
insights provided through the principles of chaos and complexity. These insights,
which were presented in Chapters 5 and 6, have implications for medicine. At the
clinical level, the applications of the mathematical models of chaos are yielding
new understanding of physiology. At the organisational level, the principles of
complex and adapting systems are helping us to understand how large organ
isations work. This in turn will help us to understand how policy might be better
crafted, how change might be better shaped, and how uncertainty is an
inescapable (and perhaps healthy) part of that process. At the theoretical level,
chaos and complexity can help us to synthesise evidence and intuition. They
dignify the notion of intuition, and re-establish the importance of experience and
wisdom, seeing them as emergent properties of the thousands of iterative,
recursive interactions in consultations. Are chaos and complexity the answer to
life, the universe and everything? Probably not, but they do help us to ask better
questions about the questions that doctors ask.
I 26
Complexity in primary care
Paper I
research into practice
Kieran Sweeney
LECTURER IN GENERAL PRACTICE
AND HEALTH SERVICES RESEARCH.
EXETER UNIVERSITY
Jonathan Stead
Liz Cosford
Evidence-Based Practice:
Can This Help Joint
Working?
NORTH AND EAST DEVON
HEALTH AUTHORITy
flBSTRA 39
This article presents a qualitative
analysis of three focus groups
convened during a study day for
health and social care professionals,
which reveals a strong perception of
a philosophical difference in
approaches to professional practice.
The prospect of health and social
care professionals working more
closely together is welcomed, and
evidence-based practice should be
Introduction: collaboration and evidence-based care
Collaboration is the new deity in health and social care. Encouraging the
two groups of professions to work together is politically correct,
professionally desirable and likely to lead to greater integration in the
delivery of health and social care services. Actually, the mantra of
collaboration has been sounded for over three decades, but advocating
working together has been a lot easier than realising it in the field
(Clarke, 2000). A myriad of impediments have been cited to explain this
failure, the three most frequent reasons being differing professional
perspectives on problems, different occupational cultures and confusion
over professional roles (Sheldon, 1994; Dailey, 1991; Abramson &
Mizrahi, 1996). Leedham and Wistow (1992) argued that differences
in values between general practitioners and social workers
actually produced conflict which undermined the best-laid plans for
collaboration.
encouraged and financially supported.
While established educational
strategies can be deployed to respond
to the conventional perceived
barriers to working together, more
innovative models are needed. The
authors commend the model of
Significant Event Auditing.
Managing Community Care
Both health and social services are also required to deliver evidence-based
care. Could training in evidence-based practice through acquiring
critical appraisal skills achieve what other strategies have so far failed to
do? There is little direct evidence to raise hopes, but many policy-makers
have identified joint training as a means through which many of the
impediments to joint working could be overcome (Department of
Health/Social Services Inspectorate, 1991). Training together would
dispel mistrust, Gambetta (1998) argued, and governments tend to
agree (Department of Health/Social Services Inspectorate, 1989).
Volume 8 • Issue 5 • October 2000 © Pavilion Publishing (Brighton) lid
21
Appendix
research into practice
This article reports on a joint study day held in April
1998 which provided an opportunity for health and
social care professionals in one locality to express their
views on joint working, in the context of a joint
training exercise in evidence-based practice. A group of
professionals from health and social care were invited
to attend a study day to discuss the principles of
evidence-based practice, to receive an introduction to
critical appraisal skills and to reflect on perceived
barriers to working together. As part of the study day,
three small groups of mixed professionals were
convened for one hour, during which their views on
these topics were explored and recorded. Efforts were
made to continue with the groups in their localities
after the study day. This report presents the analysis
of the data collated from these groups, postulates
some theory to explain the analysis and offers a
strategy for acting on the findings.
Methodology
Following a formal presentation on critical appraisal
skills, three focus groups (Morgan, 1988) consisting
of a mixture of health and social care professionals
were convened. The facilitator explored how such
skills might promote evidence-based practice (EBP) in
day-to-day work. Would EBP help address and/or
resolve some of the problems which the groups met
regularly? Could an evidence-based approach
encourage joint working? Finally, the groups were
encouraged to express their own views on the barriers
to joint working. All three facilitators received the
same briefing notes, so that the same topics were
discussed in all three groups.
A grounded theory approach to analysis of the raw
data was taken (Creswell, 1998). Thus, fresh
transcripts were read and each freestanding idea or
independent contribution was annotated and identi
fied as a theme. These themes were individually
reviewed and drawn together into higher-order codes
22
Managing Community Care
which were denoted categories. These categories
constituted the higher-order level of analysis from
which the theory - grounded in the testimony of the
individual themes - was postulated. This approach
has a long tradition in qualitative research (Glaser &
Strauss, 1967), but takes a more Glaserian approach
(Glaser, 1992).
Results
The main categories derived from the emerging
themes were entitled evidence-based practice and
working together.
Evidence-based practice
In general, the whole idea of using evidence to inform
practice was welcomed by all the groups, on the
grounds that it could help manage resources. There
was a consensus that money should be prioritised to
encourage evidence-based practice and an expectation
in some groups that such activity could increase costs
and involve a large time commitment. All groups
recognised the importance of key players being
involved in this activity.
The data reveal widespread concerns in the groups
about applying research evidence in general, despite
their overall enthusiasm for the activity. There were
two strands of thought here. One described the
difficulty of applying national data to a specific
locality, and the second the generic difficulty of
applying population evidence from randomised
controlled trials to individuals - 'every person is
different'. Some participants were more enthusiastic
than others about using evidence. ‘Clinical practice does
change with really sound studies', one person argued.
But applying research evidence can be tricky,
cautioned another, for example where a large
randomised controlled trial had selected patients who
simply were not representative of the population seen
Volume 8 • Issue 5 • October 2000
© Pavilion Publishing (Brighton) Ltd
127
I 28
Complexity in primary care
research into practice
regularly in routine practice. One particularly
poignant intervention recognised the threat posed by
evidence-based medicine: 'Do 1 really want to accept I've
been doing somethingfutilefor 20 years?'.
The groups spent some time describing how their own
working practice was affected by the particular
geographical characteristics of their localities, such as
isolation or seasonal working. One participant
summed up these themes thus:
'A lot of the links we have to make tend to be in
[name of town), but a lot of resources are
centralised - that actually is quite a barrier when
it comes to working at a local level'.
Thus described, the locality context appeared to have
an impact not simply on participants' own work, but
also on their enthusiasm for working together.
The groups agreed that restricted financing could pose
problems, not just to working together but also
within the professional groups separately. While
evidence-based practice might produce quality work,
'it is all overshadowed with thefinance of it all'.
None of the groups had any difficulty in identifying a
range of problems which could benefit from joint
working. Judging by the amount of data collected for
each topic, care of the elderly came out easily as the
key issue to tackle, in terms of rehabilitation, respite
care and the mental health of the elderly. Dementia,
falls, bed-blocking and the reason for hospital
admissions were other topics the groups wanted to
tackle in this area.
Working together
Each group separately and spontaneously identified
the theoretical advantages of professionals from
health and social care working together:
it s a great idea just to be able to communicate leaving aside the fact that there's yet another
meeting and it's time that we have to give up'.
Managing Community Care
The participants perceived an advantage in the team
approach to sharing knowledge and tackling
problems. ‘1 guess, what we've been talking about, profes
sional groups working together... mmm... those issues are
for me about relationships', said one participant.
Data which described barriers to working together
were identified at three levels. At the first level, profes
sional or operational difficulties were discussed.
Beyond this, data at a second, deeper level described
problems about the very nature of health care and
social care; this level has been called metaphysical. And
finally a philosophical level was identified, where data
relating to (perceived) fundamental differences in the
medical and social models of practice were collated.
Barriers to working together - professional level
The operational difficulty of convening multi
professional groups was described in a large number
of contributions from all three groups.
'We never get the right people around the table.'
Within this data set, general practitioners specifically
and repeatedly came in for criticism for their failure to
attend multidisciplinary groups. Some other barriers
to the idea of working together were identified. These
include divisions between health and education, which
were described as being 'poles apartConfusion about
the roles and responsibilities within non-doctor
professional groups was identified.
Barriers to working together metaphysical level
A set of data was identified under the broad heading of
working together which seemed to transcend mere
day-to-day operational issues and dwelt upon the
inherent nature of the problems faced by these profes
sional groups, which in itself might render joint
working difficult. Patients did not come describing
literal, discrete and identified health or social care
problems, but with diffuse, undifferentiated concerns.
The groups recognised the difficulty general
Volume 8 • Issue 5 • October 2000
® Pavilion Publishing (Brighton) Ltd
23
Appendix
research into practice
practitioners experience facing such 'undifferentiated
problems', which they tend to see more than others
because they are so accessible.
Within separate professions, some individuals recog, nised the changing nature of the demands on their
professional group. 7 don't think I do what I was
trained to do’, commented one social worker. Some
were simply 'so busy we can't do this job properly',
while others cited the problems associated with rising
expectations as chronically stressful.
Barriers to working together philosophical level
Two of the groups recognised serious conceptual
difficulties to working together based on a perceived
dichotomy of approach between health and social
care:
‘the basis of the assessment is different, GPs think
in terms of treatment’
‘we’ve got the problem of them working most
specifically with the social model as opposed to the
medical model '.
i
i
I
The groups considered that there was at least the
'potential for argument from polarisation of views’ and
included what they described as 'a huge variation in the
CPs’ practice of medicine’ as a potential impediment to
working together.
All three groups explored perceived differences
between the medical and social models of practice. The
data are best illustrated by this extract.
‘Medicine is much more easily definable.... with
medication you either take it or you don't; with
social services you are talking about people and
there are an infinite variety of variables.'
This perceived difference in professional approach
extended not only to the assessment of individual
cases, but also to treatment decisions generally and to
the application of research evidence.
24
Managing Community Care
Discussion
From these data, there appears to be a general willing
ness among these participants from health and social
care to work together. They felt that it would improve
relationships, improve practice and ultimately lead to
better service provision. Adopting an evidence-based
approach seemed compatible with collaboration;
money should be diverted to encourage it. At the least
it was another avenue for better communication. An
evidence-based approach could help joint working by
encouraging a joint approach to problem-solving,
which in turn might lead to a joint improvement in IT
skills, and maybe, as a consequence, better library
facilities. But enthusiasm for an evidence-based
approach was not unrestrained. At a personal level,
there was the threat that the evidence would reveal
that customary practice was ineffective (or worse),
and at a methodological level, there was the problem
of relating population research to individuals. These
concerns have been rehearsed elsewhere, and standard
educational strategies can be adopted in response to
them (Daws, 1996).
As far as pitfalls to working together were concerned,
the data show that there are concerns on a number of
levels. Issues of finance and geography were strong
themes. Evidence-based practice might increase costs,
some thought, although this view has been rejected by
the 'high priests' of evidence-based medicine (Sackett
& Rosenburg, 1995).
At a much more profound level, the groups expressed
two further sets of perceived barriers to working
together. They related to the metaphysics of working
together and philosophical differences in the way that
the participants described the medical and social care
approach to professional practice. Of all the themes
which emerged in the analysis, these represent the
most serious challenge for senior people in health and
social care who want to encourage collaborative
evidence-based practice.
Volume 8 • Issue 5 • October 2000
© Pavilion Publishing (Brighton) Ltd
129
130
Complexity in primary care
research into practice
Two lines of thinking emerged. On the one hand, the
participants identified deeply distressing parts of their
own work, like the inescapable personal unhappiness
or insuperable disadvantage which many of the clients
experienced. Here, the groups seemed to be saying
that, despite a professional preparation for their work
based on a professional analysis of the problem, many
of the issues which presented to them were existential,
insoluble and utterly dispiriting for the individuals
involved. In itself, the groups were suggesting, this
could diminish enthusiasm for evidence-based
approaches and working together.
F
The second level of problems in this category may
partly explain this. The participants from social care
thought that there was a strong difference between
their own philosophical attitude to practice and their
ideas of how, at the same level, those in health care particularly doctors - approached their work. There
could be differences, for example, in the way that
research was applied, they argued, depending on
whether the social model or the medical model of
practice was being applied. Medicine was perceived by
those in social care to be much more definite, inter
ventionist and standardised than social care. And this
view was reciprocated by some of the general
practitioners, who described the approach of social
care workers as diffuse.
r
Importantly, the relationship between the two
approaches was not seen as potentially complemen
tary or symbiotic. Rather, the groups chose adversari
al metaphors like ‘the problem', 'argument' or
'polarisation' when discussing how the two approaches
related to each other. And this distinction did not have
simply an intellectual relevance; those in social care
argued that the medical approach to service delivery
could explain the variation in clinical medical practice
that the social care professionals perceived in their
locality. But what is also interesting about this set of
data is the absence of any consideration of variation in
Managing Community Care
practice by other professionals in the groups, either
social workers or nurses.
Future prospects
The results of this research fit with other reports on
joint training (Hunter, 1993; Corney, 1995), and the
authors acknowledge that the data do not demon
strate improved collaboration. We cannot assume, as
Loxeley (1997) has warned, that skills from joint
working will just emerge -through a sort of osmosis' as
professionals learn together about a specific topic.
While the operational problems to working together
seemed predictable and relatively easy to respond to,
the philosophical distinctions which appear to be
firmly held by the participants in these groups do
represent a substantia] barrier to working together.
There is a recognised theoretical basis in medicine on
which this antithesis may be founded, reflecting either
a traditional biomedical approach to clinical practice,
or a more interpretative, contextual view of clinical
work (Evans & Sweeney, 1998). Such a distinction
could also be interpreted as a continuation of the
ancient tension between the Hygeian and the
Aesclepian traditions of medical practice. While the
former seeks to ensure the inner and environmental
equilibrium within a individual, the latter, which
conventional medicine tends to embrace, sees an
inherent disorder in sickness and seeks to bring
measured external forces to bear to restore order
(Greaves, 1996).
How can we tackle this? While the social services
staff's concerns about the nature of medical practice
are not new, it is curious that within medicine itself
there is a vibrant debate about just how robust,
scientific and black-and-white medicine is (Dixon &
Sweeney, 2000). It is perhaps a lack of familiarity,
then, which leads to these conflicting views. If each
professional group understood more about how their
Volume 8 • Issue 5 • October 2000
© Pauihon Publishing (Brighton) Ltd
25
1
Appendix
research into practice
colleagues worked, how they took decisions, the
carry out any decisions arising from a significant
elusive and uncertain way in which many of these
event audit, and these actions should be audited
regularly too.^)
decisions had to be taken, there might be greater
motivation to collaborate. Looking at existing
evidence which is relevant to common problems faced
The data from this study suggest broad enthusiasm
might underpin this.
for an evidence-based approach to practice, and a
willingness to prioritise it as a learning need. But
Within medicine, a working model exists which could
collaboration means more than just sharing a model
address these issues. The Significant Event Audit (SEA)
of good practice; it means talking together, reflecting
has been developed from industry and tailored to
and a willingness to understand how other profes
primary care (Pringle et al, 1995). It has gained in
sional groups in health care operate and improve.
popularity within general practice and is now one of
Significant Event Auditing has begun to achieve that
the compulsory criteria for a higher qualification of
for primary care teams. It is to be commended to social
general practice (Fellowship by Assessment of the
care, and should be developed as a joint effort. Maybe
Royal College of General Practitioners). In SEA, a
the key to the solution was best expressed by this
practice team meets, usually monthly, to address any
participant:
issue regarded as significant by any member of the
practice team. This could be a major clinical issue, for
'It's people who communicate, organisations don't
communicate'.
example how someone who collapsed in the surgery
was managed, or an important managerial issue, like
how to handle influenza vaccination of 1,000 at-risk
individuals. But it is also a forum where a team can
congratulate themselves, for example by discussing
an improved repeat prescription service or reducing
delays for patients seen by the nurse. Of course, it will
be the forum where complaints will be aired and
References
Abramsom, J. S. & Mizrahi, T. (1996) When social workers
and physicians collaborate: positive and negative inter
disciplinary experiences. Social Work 14 (3) 270-182.
Clarke. N. (2000) Improving the Performance of Social
Services. PhD Thesis (pending). University of Exeter, Exeter.
investigated, but the approach here is not judicial or
blaming, rather what the practice can learn from and
do about a particular complaint. Any staff member
Corney, R. (1995) Mental health services. In: P Owens,
J. Carier & J. Horder (Eds) Interprofessional Issues in
Community and Primary Health Care. London: MacMillan.
can bring an issue to the forum, confidentiality is
stressed, minutes are recorded, and an audit of actions
based on the decisions thus minuted is undertaken
Creswell, J. W. (1998) Qualitative Enquiry and Research
Design. London: Sage Publications.
regularly.
Dailey, B. (1991) Beliefs and behaviour: professionals and
the policy process. Journal ofAgeing Studies 5 (2) 163-180.
Recent evidence for primary care suggests that SEA
Daws, M. G. (1996) On the need for evidence-based general
and family practice. Evidence Based Medicine 1 68-69.
improves understanding of others' roles, assists team
working, helps to develop basic problem-solving skills
and encourages an atmosphere of reflective practice
(Westcott et al, 2000). To be successful, SEA needs
good leadership, has to establish a non-judgmental
atmosphere and should encourage equality of
participation. Individuals should be nominated to
26
Managing Community Care
Department of Health/Social Services Inspectorate (1989)
Working with Child Sexual Abuse: Guidancefor Training Social
Services Staff London: HMSO.
Department of Health/Social Services Inspectorate (1991'
Care Management and Assessment: Managers' Guide. London:
HMSO.
Volume 8 • Issue 5 • October 2000
<D Pauilion Publishing (Baghion) Ltd
131
I 32
Complexity in primary care
research into practice
Dixon, M. & Sweeney, K. G. (2000) The Human Effect in
Medicine: Theory, Research and Practice. Oxford: Radcliffe
Press.
!
Evans, M. & Sweeney, K. G. (1998) The Human Side of
Medicine. Occasional Paper 72. London: Royal College of
General Practitioners.
Gambetta, D. (1998) Trust: Making and Breaking
Co-operative Relations. Oxford: Blackwell.
Glaser, B. (1992) Emergence Versus Forcing Basics of Grounded
Theory Analysis. California: The Sociology Press.
Glaser, B. & Strauss, A. (1967) The Discovery of Grounded
Theory. Chicago: Aldine.
Greaves, D. (1996) Mystery in Western Medicine. Avebury
Publishing: Aidershot.
Hunter, D. (1993) ‘It's not the knowing, it’s the doing'
Health Service Journal 21 October/.
Leedham, 1. & Wistow, G. (1992) Community Care and
General Practitioners. Leeds: Nuffield Institute.
Loxeley, A. (1997) Collaboration in Health and WelfareWorking with Difference. London: Jessica King Publishers.
Morgan, D. L. (1988) Focus Groups as Qualitative Research.
London: Sage Publications.
Pringle, M., Bradley, C. P, Carmichael, C. M., Wallace, H.
& Moore, A. (1995) Significant Event Audit. Occasional
Paper 70. Exeter: Royal College of General Practitioners.
Sackett, D. L. & Rosenburg, W. M. C. (1995) The need for
evidence-based medicine. Journal of the Royal Society of
Medicine 88 620-624.
Sheldon, B. (1994) Biological and social factors in mental
disorders: implications for services. International Journal
of Psychiatry 40 (2) 87-105.
Westcott, R.. Sweeney, G. & Stead. J. (2000) Significant
Event Audit in practice: a preliminary study Family Practice
17(2) 173-179
Correspondence to KGSweeney@exeter.ac.uk
A Pavilion Premier Event:
STANDARDS FOR REGULATING
AND COMMISSIONING SERVICES
Examining the implications of national standards and their regulation
on local decision-making and commissioning of social care provision
13 December 2000
-
Royal Society of Medicine, London
'-phis conference breaks new ground by providing an opportunity to consider the implications of the national
1 regulation of care provision and of national required standards for local decision-making. It will look at the extent
to which national requirements impact on matching local services to local needs, the implications for local contracts
and whether individual needs get lost in the bigger picture.
Speakers include:
Chris Hume, Programme Manager,
National Care Standards Commission
Chris Johns, Project Manager, SSI, Department of Health
TYish Davies, Department of Health lead on standards
Hugh Gardner, Association of Directors ofSocial Services,
Wales
Bob Lewis CBE, Chair CCC Consumers' Issues Group
Who Should Attend?
Social Services and Health
Chief Officers and Managers
Commissioners
Regulators
Registered Home
and Social Care Providers
Pavilion
PCTs
For a full conference programme or any further information please contact:
Pavilion. 8 St George s Place. Brighton. East Sussex BN I 4GB Tel 01273 623222 Fax 01273 625526
Email pavpub@pavilion.co.uk Book online at www.pavpub.com
Managing Community Care
Volume 8 • Issue 5 • October 2000
© Pauthon Publishing (Brighton) Ltd
27
Appendix
133
Paper 2
Family Practice
© Oxford University Press 2000
Vol. 17, No. 5
Printed in Great Britain
A preliminary study of the decision-making process
within general practice
Rebecca Mears and Kieran Sweeney8
Mears R and Sweeney K. A preliminary study of the decision-making process within general
practice. Family Practice 2000; 17: 428-429.
Objective. The aim of the present study was to explore the factors that contribute to the pro
cess of decision making within general practice, over and above evidence-based information.
Methods. A qualitative study was conducted using semi-structured interviews on a purposeful
sample of GPs, based in the South West of England. Each interview was tape-recorded and
transcribed verbatim.
Results. Five broad categories emerged from the data: practitioner; patient; practitioner-patient
relationship; verbal and non-verbal communication; evidence-based medicine; and external
factors.
Conclusion. The nature of general practice is such that the process of making clinical decisions
is complex. In an era when GPs are being overwhelmed by evidence-based information, con
sideration needs to be given to the implications that the nature of the decision-making process
has upon the way 'evidence' is constructed and promoted within general practice.
Keywords. Clinical decision making, decision making, evidence-based medicine, general
practice, GP.
I
I
Introduction
Methods
Defined as the “conscientious, explicit and judicious use
of current best evidence”,1 evidence-based medicine
(EBM) has emerged as a new paradigm for medical
practice. Awareness of the latest scientific evidence, the
ability critically to appraise literature and assess the gen
eralizability have been identified as integral to the prac
tice of EBM. However, the evaluation of evidence within
general practice is often illogical and irrational2 and
it cannot be assumed that GPs practise the principles
underpinning EBM in their decision making.
The purpose of this study was to conduct a preliminary
investigation of the factors that contribute to the clinical
decision-making process within general practice, over
and above the assumptions underlying evidence-based
information.
Semi-structured interviews were conducted using a
‘convenient’, purposeful sample of GPs, who were either
based in a research-based general practice or involved in
continuing medical education.
Data collection
RM conducted semi-structured interviews with each
practitioner for ~1 hour. Each GP gave both their written
and verbal consent. Interviews were audio-taped and
transcribed verbatim and, to ensure reliability, each
transcript was read independently by RM and another
researcher who was blinded to the study aims. Their
results were compared in order to establish a degree of
congruence and disparity. A degree of congruence of
80% was deemed acceptable.4 Respondent validation
was also used to ensure that the data analysis and inter
pretation were an accurate reflection of the views of the
practitioners.
Received 4 November 1999; Revised 6 April 2000; Accepted
16 May 2000.
Department of Epidemiology and Public Health, Institute
of Child Health, Great Ormond Street Hospital, 30 Guilford
Street, London WC1N 1EH and “Research Development
Support Unit, Postgraduate Medical School, University of
Exeter, Exeter, UK.
Results and discussion
All five practitioners were based in practices within the
South West of England. One of the practitioners was
428
134
Complexity in primary care
A preliminary study of the decision-making process within general practii
ice
female. Their mean age was 47 years (range 40-54 years),
and three of the practices were training practices.
The results indicated five broad categories that con
tribute to the decision-making process within general
practice: practitioner; patient; practitioner-patient re
lationship; communication; EBM; and external factors.
Practitioner
All of the practitioners described that previous clinical
experiences, and their own philosophy of health and clin
ical beliefs had an impact upon clinical decision making:
"doctors also have their own philosophy of health, and
there’s no reason why they shouldn’t...” (Interview 1)
and “it’s the things that go wrong that imprint on your
memory..(Interview 3).
Patient
Among all respondents, there was also recognition that
understanding patients’ cultural beliefs, background and
attitudes is integral to the decision-making process: “you
have to respect their beliefs and values ...” (Interview
5), “you have to know where the patients are coming
from ... and what their beliefs are” (Interview 1).
Practitioner-patient relationship
Respondents all referred to the importance of main
taining good relations with patients: .. patients are
techmcolour and actually the relationship is technicolour” (Interview 1). One practitioner described a
situation in which he had bowed to the expectations of
the patient for the sake of maintaining good relations:
‘the nature of the relationship is one that continues and
goes on and there may be far more important issues
coming up than this trivial issue of whether or not you
prescribe penicillin ..(Interview 1).
Verbal and non-verbal communication
All of the respondents mentioned being aware of the
language that they used during a consultation: “you’ve
got to pitch what you say at a level that the patient will
understand .
(Interview 3). Two respondents cited
non-verbal cues as informing the clinical decisions that
they made, e.g. whether or not to prescribe: “patients do
give quite strong messages without necessarily express
ing them verbally, about what they want” (Interview 1).
Evidence-based medicine (EBM)
There was a pervading feeling among respondents
that the EBM approach to clinical decision making
did not allow for the complexities inherent within the
429
decision-making process in general practice: “only a
proportion of clinical decision making is ever to do with
research ..(Interview 4),
. EBM measures the
things that can be measured ..(Interview 2).
External factors
Time, cost and the media were three factors extraneous
to the practitioner and patient most commonly cited
as influencing the decision-making process: “GPs are
conscious of society’s views but particularly cost ..
(Interview 3), “time is critical, we don’t have very long,
that’s the problem” (Interview 5), .. the media are
more powerful than anything else” (Interview 5).
Study limitations
Due to both time and financial constraints, the sample
size m this study was very small. This factor represents a
major limitation for this study; the results are therefore
presented as preliminary.
Conclusion
The findings of this preliminary investigation suggest
in support of previous studies,3-5 that the approach to’
clinical decision making within general practice is multi
faceted. The complexities inherent within this process
are not reflected in the ‘linear’ approach of formulating
a clear clinical question, promoted within the EBM model.1
In an era when GPs are being overwhelmed by evidence
based information, consideration needs to be given to
the implications that this has upon the way in which
evidence’ is constructed and presented to GPs. Any
evidence-based model aimed at general practice needs
to be compatible with its complex, and often irrational,
illogical nature.
References
1 Sackett DL Richardson WS, Rosenberg W, Haynes BR. Evidence
based Medicine—How to Practice and Teach EBM. Churchill
Livingstone, 1997.
2 Sweeney K MacAuley D, Pereira Gray D. Personal significancetne third dimension. Lancet 1998; 35L134-136
Tomlin K, Humphrey C, Rogers S. General practitioners’ percep‘ Miu
J 1CCl,Ve hea'th
Br MedJ 19";31fc 1532-1535.
Miles MB, Hubennan AM. Qualitative Data Analysis. 2nd edn.
Thousand Oaks, CA: Sage, 1994.
5 Jacobson LD, Edwards AGK, Granier SK, Butler CC. Evidence449^52 edlC'ne and genCraI Practice BrJGen Tract. 1997; 47:
1
1
Appendix
135
Paper 3
Downloaded from bmj.com on 28 November 2005
Primary care
Why general practitioners do not implement evidence:
qualitative study
A C Freeman, K Sweeney
Abstract
I
Objectives To explore the reasons why general
practitioners do not always implement best evidence.
Design Qualitative study using Balint-slyle groups.
Setting Primary care.
Participants 19 general practitioners.
Main outcome measures Identifiable diemes dial
indicate barriers to implementation.
Results Six main themes were identified dial affected
die implementation process: die personal and
professional experiences of die general practitioners;
die patient-doctor relationship; a perceived tension
between primary and secondary care; general
practitioners’ feelings about dieir patients and the
evidence; and logistical problems. Doctors are aware
dial their choice of words with patients can affect
patients' decisions and whether evidence is
implemented.
Conclusions General practitioner participants seem
to act as a conduit within the consultation and regard
clinical evidence as a square peg to fit in the round
hole of the patient's life. The process of
implementation is complex, fluid, and adaptive.
In a recent questionnaire study of general
practitioners' attitudes to evidence based medicine,
answers to an open question suggested that there are
unique baniers to implementing evidence in geneial
practice within a patient centred context" This study
set out to explore the issues raised by these responses.
We used a qualitative approach to explore the reasons
why and circumstances in which doctors had not
implemented evidence they knew about.
Participants and methods
Evidence based medicine is based on tuniversally
----appealing ethical and clinical ideals in that it- promotes
the identification of die best methods of health care
and helps patients and doctors to make better
informed choices.' Its fiamework for searching out and
critically appraising evidence helps doctors ask
answerable questions to help patients make appropri
ate decisions.’
Although evidence based medicine has heightened
awareness of the most effective management strategies
for many conditions, much of the evidence is not acted
on in everyday clinical practice.’ Numerous strategies
to improve implementation of such evidence have
been tested,’ and various impediments have been iden
tified? General practitioners have been cautious about
Three focus groups of established general practition
ers were set up in three areas, each located around a
different district general hospital. The hospitals were in
the south west of England and covered the area served
by a single primary care research network Each area is
geographically separate by about 80 km and tends to
develop its own medical community. The groups did
not contact each other throughout the study and were
not in regular social or professional contact outside the
study. By using these separate groups, we aimed to
improve the trustworthiness of the data.
Participants were asked to discuss their behaviour
in individual cases, which could be seen as sensitive. We
therefore adapted the standard focus group techniques
to use a Balint-style model. This style of group work is
widely recognised in general practice, and derives from
the work of the psychotherapist Michael Balint1" The
focus groups were not pure Balint groups because they
did not include a psychoanalyst However, a widely
used modified form of these original Balint groups has
become common in general practice." The particular
Balint-style feature of these groups that distinguished
them from standard focus groups was that each meet
ing focused around the case notes of a particular
patient the doctor-patient relationship, and the
feelings that were generated. Basic rules of confiden
tiality are a prerequisite for convening the group, and
the participants agree not to discuss material raised in
the group outside. The same group of doctors met on
several occasions in the hope that as the group
the evidence based model generally.6 In one study that
matured, they would feel more comfortable about
asked general practitioners why they depart from
evidence based practice, the commonest reason was
reluctance to jeopardise their relationship with the
patient? Apparent hesitation in applying evidence in
specific clinical areas such as atrial fibrillation has been
attributed to patients’ unwillingness to take die drugs."
exploring honest reasons behind their failure to
implement evidence.
The groups consisted of six to eight volunteer gen
eral practitioners, each led by an experienced group
leader. The group leader was given an honorarium to
lead and administer die groups and operate the tape
Introduction
BMJ VOLUME 323 10 NOVEMBER 2001
bmj.com
Somerset and
North and East
Devon Primary
Care Research
Network, Institute
of General Practice.
School of
Postgraduate
Medicine and
Health Sciences.
Exeter EX2 5DW
A C freeman
genrral practire
reseanh facilitator
K Sweeney
general practice
researchfacilitator
Correspondence to:
A C Freeman
PCRN@exeter.acuk
BMJ 20013231-5
1
136
Complexity in primary care
Primary care
Downloaded from bmj.com on 28 November 2005
recorder. The plan was to have the groups meet about
once a month on six occasions, each meeting lasting
Main themes from data
about two hours. Two of the groups consisted of
doctors from different practices and one group
The process of implementing clinical evidence is
affected by the personal and professional experiences
comprised doctors from one practice. Participating
of the doctor
doctors represented a mix of urban, rural, and semiruThe relationship that the doctor has with individual
ral practices. There were a total of 19 doctors; 13 men
patients also affects the process
and six women. Their lengdi of time as a principal var
There is a perceived tension between primary and
ied from diree to 25 years. Fourteen held the member
secondary eye: die doctors diought that specialists
ship examination of die Royal College of General
approach evidence based practice differently
Practitioners, and seven were general practice trainers.
The practitioner’s feelings about their relationships with
At each meeting, a group member was asked to
patients and about the evidence have an important role
in modifying how clinical evidence is applied
present the details of a case in which he or she had
The doctor’s choice of words in consultations can sway
knowingly not followed evidence based practice.
patients
to accept or reject clinical evidence. Doctors
Participants were advised to anonymise the patient
realise this and can use it to pre-empt patients' decisions
details and not present any material that could lead to
Implementation comes up against logistical problems,
the identification of a particular patient We asked the
which affect how evidence is applied
groups to discuss die case and explore the implemen
tation issues arising from it as well as die doctor’s feel
ings about these issues. The local research ethics
Personal and professional experience of
committee approved the study.
practitioner
The researchers were not part
[
' group, but
of" the
Our data show that doctors’ personal and professional
before the first meeting of each group a researcher experiences influence how clinical evidence is impleattended and explained the research agenda. We mented.
homogeneous
mented. Despite
Despite being
being aa relatively
j ’ ‘
'
*
group, the----------general1 —
practitioners’ enthusiasm°for (he
explained that the individual doctors would be anonymous. We had no further contact with the groups. We
evidence and the way in which they implemented it
varied. This seemed to be partly explained by tlieir
returned copies of the transcripts to the groups, and
each member understood that if diey were not happy
previous experience of clinical practice.
Two influences were relevant: the doctors’ life
with the content dial transcript would not be used.
experience
of hospital medicine as
stuThe meetings were taped, and die tapes
deliveredand experience
...
-----------------------to us. The tapes were transcribed, and each researcher
denU or juniors doctors. “My grandfather died when
separately analysed the transcripts. Each researcher
he was
WaS shocked,
shocked” recalled one participant, discussing
anticoagulation
used a grounded dieory approach in developing theoan,icoaguladon in atrial fibrillation,_ “so ,I reach for a
decent dose of warfarin and digoxin no hesitation at
retical principles (or at least explanatory principles).12
-------UUUU1K
w
alt" Another said: “I actually had two 50 year olds who
This was
to ensureU1Vthat
the coding
of themes
consistently and robusdy followed grounded theory
had strokes from alrial fibrillation because they didn’t
rules and that all the emerging themes were directly gCt warfann ''' tha‘ real,y hit me" In another group,
supported by verbatim data from the meetings We did ST"?1 Practitioner said. “ I lost a patient as an
not set out with tire overarching aim of
SH°‘ S^that PUls ™ off warfarin ”
not set out with die overarching aim of generating
Accidents, mishaps, or spectacular clinical suc
theory from the findings.
cesses have a direct influence on subsequent practice.
We met to compare analysis and identify common
Commenting again on anticoagulation in atrial
themes. To ensure compatibility of analysis, we each
fibrillation, a participant exclaimed, “I’m back on it”
analysed three transcripts joindy and the others
This doctor had previously been uneasy about antico
separately. For the separate analyses, we were given the
agulating patients in atrial fibrillation but had recently
transcripts recorded out of our own area to minimise
seen one of his patients who was not given warfarin
the recognition of names, accents, or circumstances
have a cerebrovascular event. This theme was taken up
that could lead to the identification of patients or par
m another group: “But I suppose if we had a run of
ticipating doctors.
people who ... then had terrible hemiplegias and
ended up being a huge workload on die community
if we saw the ones the papers were talking about, we
would probably be warfarin zealots, wouldn’t we.” One
Results
doctor summed up this view, thus: “We are influenced
Transcripts for 11 meetings were available for analysis.
at least as much, if not more, by the experiences of
Two of the groups met six times each, and the diird
individual patients as we are by the evidence.”
°^e onty-that
is,, —
13 .meetings.
recordings
oftwo
two r» .
,
—vu,,6i The1C
cuiuuigs oi
...
of tlie groups could not be used because of poor sound “oc‘°^s relationship with individual patients
Implementation was influenced by the relationships
quality.
that doctors developed with their patients. "Even if die
The main clinical areas die general practitioners
evidence was extremely good,” one general prac
discussed included hypertension, ischaemic heart
titioner said, “most of us would only ever interpret it in
disease, and anticoagulation. Other topics developed in
the
context of the patient.” Perceived patient characterthe groups -discussion included diabetes, chronic
istics could have a positive or negative effect on imple
obstructive pulmonary disease, menorrhagia, choles
mentation. “Of course, if they're the sort who always
terol, and the use of investigations. Six main themes
want die specialist, then you follow their [the
emerged from the data (box).
specialist’s] advice.” Anotiier explained, “ I think you
2
BMJ VOLUME 323 ] 0 NOVEMBER 2001
bnij com
Appendix
ailiravast
Downloaded from bmj.com on 28 November 2005
have to judge how people feel about it I try to get
patients to reveal to me where they lie in the game ...
from I want it mate to I don’t want to know nothing
about it doc ... I make tremendous judgments.”
Patients could influence clinical decisions as a
result of their own experiences. “Well he’s a farmer, so
every time he calls the vet he gets antibiotics.” Another
patient reportedly said, “My brother died on warfarin,
I’m not taking rat poison." Some doctors found that
personal relationships tended to make practising
evidence based medicine “harder because you have a
close relationship with them.” At other times patients
could simply block a doctor’s attempts to practise
evidence based medicine: “Sod that, says the patient.
I'm fine."
The assumptions doctors made about their
patients seemed at times paternalistic. Some were
described by their doctor as “ the type who did not
want to rock tire boat," others as “depressive cum fatal
ist” “Somatisers,” declared one doctor, “eventually get
something.” By using these descriptions, the contribu
tors were suggesting that their view of the patient
modified how and when they applied the evidence.
One doctor built up the relationship with the
patient by initially not following the guidelines and
then, in a position of greater trust was able to
implement the guidelines properly. “I have now
followed die guidelines of course, but in a sneaky way
and it’s taken about three months to do it”
I
Perceived tension between primary and secondary
care
The general practitioners talked at length about their
relationships with secondary care doctors. They felt
that specialists approached evidence based practice
differently, treating “diseases rather than patients” in a
context that they perceived as much more controlled
than the “real life” of general practice. On the whole,
the relationship was described in pejorative terms.
'They do seem a slightly different breed," one general
practitioner said, referring to cardiologists. A doctor in
another group described cardiologists as “being a bit of
an evidence based mafia.”
Specialists were accused of failing to realise just
how tricky it was controlling some common diseases.
“You get stroppy letters from the clinic saying your
patient’s blood pressure is still 160, and I go ... yes, yes,
I know. You feel under pressure from the guidelines,
but you know it’s not from want of trying.” In one
group, quite a fundamental difference in approach to
clinical practice between primary and secondary care
was described. “A few hypertensives, without any
symptoms, they're well. They're just running a risk We
give them a drug and a side effect—change the quality
of their life,” said one doctor. A female participant in
the same group agreed, saying, "Show me one GP who
doesn’t think like this, show me one cardiologist who
does. I mean, this is the problem, isn’t it?”
Clinical evidence can evoke feelings among doctors
and patients
For the doctors in our study, clinical evidence is notjust
an intellectually celibate commodity that is lifted out of
medical journals and transferred to a patient. It has an
emotional impact on practitioners and patients. “Yes it
does make me feel anxious ... all the BMJ&, all the rags
... these people must be on warfarin." “With me messBMJ VOLUME 323 10 NOVEMBER 20(11
L
bmj.com
137
ajwaarasanKBM*
Primary care
ing about with his medication and trying to practise
evidence based medicine, I found it was making him
[the patient] feel more anxious.” Sometimes the knowl
edge that the evidence existed, waiting to be applied,
was seen as a burden in itself: “We get bogged down
with perhaps putting the evidence first and consecrat
ing it."
Another aspect of this theme reflected the doctors’
feelings about the consequences of failing to act on
clinical evidence. One participant poignantly
described how, after the death of a young man who had
been inadequately anticoagulated for a venous throm
bosis, he felt unease “standing behind his widow in the
greengrocer queue.” Another group, taking up this
theme, distinguished between probability and cer
tainty, reflecting the tension general practitioners feel
about predicting the clinical course in any one person:
“You don’t know, do you? You just don’t know."
The group discussions also produced data that
indicated doctors’ familiarity with the evidence and a
positive attitude to it They described its importance to
everyday practice: “I think it’s always the basis for most
of what I do ... it’s fundamentally evidence based but
it's tailored completely." They recognised that evidence
based medicine gives new emphasis: “That is the one
that I have been hammering, the diabetic blood
pressures, to try and get them to 140/80, and I am cer
tainly getting them better titan I was but it is hard
work.” For some of the general practitioners evidence
based medicine was revolutionary: “I think that is the
first time I have become aware of one study, or group
of studies, that has actually changed my practice within
a week."
Words used by doctors can influence patients’
decisions
Doctors realised that the words they chose to present
the evidence could have a strong influence on the
patient’s decision. They effectively limited the options
while seeming to invite the patient to make the
decision. The contributors framed these themes with
phrases such as “It’s how you pul it over,” and “It
depends on how you feed information to people.” The
semantics then affect the way in which evidence is
implemented by swaying the patient in a particular
direction. “There is a reasonable chance of you having
a stroke in the next year or so if you don’t do
something about your blood pressure ... I'm as
barbaric as that,” commented one participant
The participants realised that this in effect
“pre-empted" the decision that they were encouraging
patients to take during consultations. Some talked of
“selling" a particular view on clinical evidence. This
tension between encouraging autonomy and effec
tively limiting options by the slanted presentation of
relevant material was a relatively strong theme: “I make
these judgments in theory with tire patient but
probably on my own.” Another contributor described
the problem as, “How much are we obliged to
persuade people, or do we let them make up their own
minds?”
The choice of words or the use of metaphors like
“slanting" or “selling" were mechanisms the doctors
used to influence patients to make a decision about
their treatment that was consistent with what the
doctor had decided was appropriate. Doctors would
3
I 38
Complexity in primary care
Primary care
Downloaded from bmj.com on 28 November 2005
refer to “rat poison” when describing waifarin if they
felt its use would be difficult or inappropriate, or
describe pills as “having been shown to keep the heart
young" when diey wanted a patient to agree to
treatment When a doctor argued that it “depends on
how you feed information to people," other members
of the focus groups debated the issue body: doctors
might influence decisions, they said, but patients can
refuse to accept advice too.
Logistics of general practice
The doctors in this study described some tricky logisti
cal problems that made them less endiusiastic about
implementing clinical evidence. “Risky," “hard work,”
and a “hassle” both for doctors and patients were typi
cal descriptions of the problems of starting treatment.
One doctor said, "The problem is starting him on the
ACE because he is very anxious about any medication
change, and every time you change die medication it
entails another four or five visits to go and see him and
to try and reassure him that he is on the right
medication."
Complications always tended to happen “over die
weekend,” and those practitioners who, for example,
did not always have nursing staff to help do blood tests
seemed to be less enthusiastic about implementing evi
dence on anticoagulation. When discussing the poten
tial side effects of warfarin, one participant said, “ It’s
not a minor bleed if your patient is 30 miles from the
nearest transfusion service."
Knowing the patient's personal situation influ
enced implementation too. Doctors took into account
the patient’s behaviour, capabilities, or rural location
when making decisions. One doctor felt reluctant to
anticoagulate one 88 year old woman because “she had
an alcohol problem, kept falling. She was forever in
casualty being stitched up, bandaged up, whatever.”
■
Discussion
This study suggests that the general practitioner acts as
a conduit in consultations in which clinical evidence is
one commodity. For some doctors die evidence had
clarified practice, focused clinical effort, and sometimes
radically altered practice. But a stronger theme from
our data is that doctors are shaping the square peg of
die evidence to fit die round hole of the patients life.
The nature of the conduit is determined pardy by the
doctors’ previous experiences and feelings. These feel
ings can be about the patient, the evidence itself, or
where die evidence has come from (the hospital
setting). The conduit is also influenced by the
doctor-patient relationship. The precise words used by
practitioners in their role as conduit can affect how evi
dence is implemented. In some settings, logistical
problems will diminish the effectiveness of the conduit
Strengths
The strengths of our study derive from the fact that
three groups were held separately (enhancing the
trustworthiness of identified themes). There was good
concordance in the analysis of joindy reviewed
transcripts, and validation by respondents did not show
serious disagreement with die analysis. One group
could not continue in the study, and dropped out This
group consisted of doctors in a single practice; one of
the partners was enthusiastic about the project but was
What is already known on this topic
General practitioners do not always act on
evidence in clinical practice
General practitioners are reluctant to jeopardise
their relationship with the patient and sometimes
feel that patients are unwilling to take drugs
What this study adds
Implementation of evidence by general
practitioners is a complex and fluid process
Decisions are influenced by the doctor’s personal
and professional experience as well as by their
knowledge of and relationship with the patient
Doctors’ choice of words can influence patients'
decisions about treatment
unable to sustain the other partners' interest Because
the group consisted of doctors in a single practice, the
discussions involved the whole practice allocating time
whereas in the other groups, individual geneiul practi
tioners made their own arrangements to attend.
For the two groups that met six times, the Balint
format seemed to work well. The doctors spoke
honestly about difficult clinical situations in which their
practice was incompatible with the principles of
evidence based medicine. Over the course of the meet
ings, doctors developed sufficient confidence in die
confidentiality of the group to allow them to speak in a
way that probably could not have been captured as well
by another qualitative instrument Semistructured
interviews might have offered an alternative: but
careful listening to these tapes suggests that the honest
interaction among group members encouraged
individuals to be more explicit about their experiences
than they might have been in a one to one interview.
Implementation of evidence
Doctors in the groups were talking about situations in
which they already knew the evidence but had not
implemented it. Although the groups did not confine
their discussion exclusively to incidents in which the
clinical evidence was not applied, the data focus wholly
on implementation issues. We felt that if a wider brief
had been given to the groups—for example, to discuss
implementation generally-the detail of die difficulties
these practitioners had implementing evidence would
have been less likely to come up. There was plenty of
evidence that the doctors were implementing evidence
and were happy to do so. The data also indicated that
doctors were working together with patients and for
the benefit of their patients. Sometimes these factors
and the doctor’s experience lead to the conclusion that
strictly sticking to the rules of guidelines is not appro
priate. Whether that is the strength of individual
doctoring in a long standing and trusting relationship
with a patient or a weakness remains open to debate.
The doctors associated evidence based medicine
with randomised controlled trials and systematic
reviews. There was no data to show that they were
aware of evidence from qualitative or observational
research, although such studies are beginning to
inform evidence based medicine.
4
BNQ VOLUME 323
10 NOVEMBER 2001
binj.com
Appendix
Downloaded from bmj.com on 28 November 2005
Put together, these themes illustrate the complexity
of implementing evidence from well structured clinical
trials in individual patients. Our findings are supported
by other studies tn the United Kingdom,11 ” the Nether
lands,’ and Australia.14 In some ways, our study
illustrates what Kernick has described as the parallel
universes of scientific research and general practice.15
We argue that the doctors in litis study were exploring
personal importance—that is, the “key to the transfer of
an idea to and the evaluation and interpretation of an
idea by the doctor and patient together.”'6 Evidence is
not implemented in a simple linear way, as some defi
nitions of evidence based practice imply, but in an
evolving process whereby reciprocal contributions
from the doctor and the patient over time influence
how evidence ultimately is used.
Wc thank tiic general practitioners who gave their time to help
in this research.
Contributors: ACF conceived the idea for this project, was
involved at every stage of the study, and contributed to the
analysis and all sections of the final paper. KS was involved at all
stages of tire study, and contributed to the analysis and all
sections of the final paper. ACF is the guarantor.
Funding: This research was supported by a grant from the
NHS South West Research and Development Executive.
Competing interests: None declared.
Kcnidge I, Lowe M. Henry D. Ethier and evidence based medicine. UMj
1998316:1151-3.
139
Primary care
2 Strauss SE, Sackett DI. Using researrh findings in clinical practice. RMJ
1998317:339-42.
luyatt G, Cook D. Transferring evidence from
3 Haynes RB. Sackett D, Cu
coming barriers to application. Etidenced-Based
research to practice: overct
Medicine 19975:68-9.
4 Oxman Al), 'Ihomson MA, Davis DA Haynes RH. No magic bullets; a
systematic review of 102 trials of interventions to improve professional
practice. CanMedAssocJ 1995:153:1423-31.
and
t
5 Budd J. Dawson S. hiftucnrmg clinical prndice.implementation of'rtuarrh
deielopment results. London: Management School Imperial College of
Science Technology and Medicine, 1994. (Report to North Ihames
Regional Health Authority.)
Jnty. In: Marinker M. cd. Seme and sm(i Sweeney KG. Evidence an uncertaint
sitnlih m health care. London: BMJ_ TPublishing, 1996:59-87.
7 Veldhuis M. Wigersma I..
L. Okies I.L Deliberate departures from
'
good gen
eral practice: a study of motives among Dutch general practitioners. BrJ
Gen had 1998:48:1833-6.
8 Howitt A, Armstrong D. Implementing evidence based medicine in gen
eral practice: audit and qualitative study of antitlirombotic treatment for
atrial fibrillation. BMJ 1999318:1324-7.
9 McColl A, Smith H, White P, Field J. General piactinoneris perceptions of
the route to evidence based medicine: a questionnaire survey. BMJ
1998316:361-5.
10 Balint M. The doctor, hu patient and the illness. London: Pitman. 1957.
11 SalinskyJ. Psychoanalysis and general practice: what did the Romans do
for us? BrJ Gen Brucl 2001 ;51:506.
12 Glaser B, Strauss A The diicmery of grounded theicory. Chicago: Aldine. 1957.
iractitioners' pcrccptioiu of
13 Tomlin Z. Humphrey C, Rogers S. General pr
effective health care. BMJ 1999318:1532-5.
14 Mayer J, Piterman L The attitudes of Australian GPs to evidence-based
medicine: a focus group study. Tam Prod 1999:16:627-32.
15 Kernick DE Muddling through in a parallel track universe (letter]. BrJ
Gen Prod 200030325.
16 Sweeney KG. MacAuley D. Gray DP. Personal significance: the third
dimension. Ijmat 1998351:134-6.
(Aarplal 6 August 2001)
!
I
BMJ VOLUME 323 10 NOVEMBER 2001
bmj.com
5
140
Complexity in primary care
Paper 4
J Med Ethics: Medical Humanities 2001 ;27:20-25
A comparison of professionals’ and
patients’ understanding of asthma:
evidence of emerging dualities?
K G Sweeney, Karen Edwards, Jonathan Stead and David Halpin University of Exeter and North and East
Devon Health Authority
Abstract
I
’j
I
number of general practices have responded by
providing an increasing number of clinics, often led
Despite an increase in the provision ofservices to
by specialist practice nurses.4 ’
patients with asthma, morbidity from the disease
Despite this increase in service provision, recent
remains high. Recent research (outside asthma) has
surveys
have shown that about half of all asthmatics
raised the possibility that patients may develop a
continue
to have night symptoms, and of that
conceptualisation of illnesses which is not entirely
group, about half have such symptoms most
compatible with the prevailing biomedical view. This
nights.4 There is no good indicator which can accu
paper compares the way in which health care
rately predict adverse outcome in asthmaA’ Nei
professionals and patients with asthma described
ther the peak expiratory flow rate (PEER),
various aspects of the illness, using an approach which
symptom scores used in isolation, nor amount of
considered the type of knowledge which might be used
bronchodilator use has been validated as a reliable
to construct the respective conceptualisations of
predictor of outcome.’10
asthma. A qualitative method is empliyed, using focus
Thus, the situation is paradoxical: health service
groups. Eight focus groups were convened, four of
provision for asthmatic patients is greater, and
professionals and four of patients with asthma.
effective treatments are available, but morbidity
Following the initial data analysis, the results were
from the disease remains high. A number of factors
reviewed linguistically, with particular attention to the
might contribute to this situation. There is
use of metaphor.
published evidence which suggests low levels of
The health care professionals and patients
adherence to seemingly logical, rational medical
participating in this study agreed broadly in their
advice, particularly about the prophylactic use of
explanations of the aetiology and drug treatment of
steroids, and considerable dissonance between their
asthma. The data suggest lack of congruence in the
recommended and actual use." A person’s attitude
development of treatment strategies and locus of
to and beliefs about asthma can influence the way
control. Health care professionals and patients in this
treatments are used.'* Thirdly, there could be mis
study used linguistically different metaphors to
understandings between the doctor or nurse and
represent the disease: the former more frequently used
patient
about the nature of asthma or its treatment.
metaphors evoking on-going processes, the latter
A small number of semantic studies have shown
visualising the chest (in their use of metaphor) as a
how such potentially important differences in con
static container, emptying and filling throughout the
ceptualisation might be revealed by the vocabulary
course of the disease. Two commentaries from
used by patients” or by their use of metaphor.’*'1*
philosophical and anthropological literature are
Beate et al have argued” that such dissonance might
considered in order to offer theoretical accounts
relevant to this interpretation. The data suggest an
better be understood by considering theories of
i----- — r»:—
between
emerging duality in the approach to treatment plans, in knowledge,
citing Piaget.•’—
s distinction
the roles played by professionals and patients with
operational and figurative knowledge.18 ” The dual
asthma, and in the different types of knowledge used by taxonomy of knowledge alluded to by Piaget is ech----------------oecj ejsewhere> notably in Toulmin’s separation of
professionals and patients
to construct- their-----------------respective
working models of asthma.
universal (essentially scientific) and existential
(effectively personal, lived) knowledge?0 Michael
(J Med Ethics: Medical Humanities 2001 ;27:20-25)
Keywords: Asthma; metaphor; types of knowledge; episte Polanyi’s seminal text. Personal Knowledge, deals
centrally with the distinction, arguing for a
mology; language
reconciliation between the two epistemologies.2’
In a consultation about asthma or any other con
Introduction
dition, the way in which information is packaged,
Asthma is a common disease. Its prevalence is expressed and exchanged is central. Such infor
increasing, and despite a clearer understanding of mation, expressed in words, reflects the speaker’s
its pathogenesis, morbidity and mortality from the' thinking and, in turn, knowledge base. Analysing
disease remain high.1'5 Most patients with asthma consultations from a.;«.
—
this perspective, that is seeking
receive their care in the community, where a large to elicit the knowledge base which constitutes the
tmw. tncdicalhimianities. com
Supplied by The British Library - "The world's knowledge"
bedrock of the exchange might, therefore, shed
some light on how the participants think. Are the
knowledge bases of the participants in such a con
sultation the same? If they are different, how can we
elicit this, and what difference might that make?
This study set out to explore and compare how
doctors and nurses on the one hand, and individu
als with asthma on the other, expressed their
understanding of various aspects of asthma. We
attempted to compare the explanatory constructs
of the two groups, to see if these would cast any
light upon their respective understanding of how
the disease worked, and to see if they influenced the
way asthma was managed.
Methods
This paper reports on part of a large quantitative
study of asthma, whose aim is to identify predictors
of deterioration of asthma. In an introductory part
of this study, the initial aim was to seek out and
compare patient- and professional-based outcomes
of asthma. Both sets of outcomes were then to be
identified by analysing focus group data, with two
series of focus groups (patient and professional)
running in parallel. Focus groups were chosen as
the preferred qualitative method because they were
considered most.likely to give rich data, particularly
from patients, by allowing the group interaction to
encourage the formation of patient-sensitive out
comes.” After the first two groups, however, it was
obvious that little data about outcomes was emerg
ing from the patient group, while the professionals
very quickly rehearsed the well-known “medical”
outcomes for the disease. Preliminary field notes of
these meetings commented on the vocabulary used
by some patients to describe the experience of hav
ing asthma, and noted ho\tf this contrasted with the
conventional biomedical description of the profes
sionals.
The research question therefore changed, in
keeping with the heuristic nature of qualitative
research.” The inquiry then focused on the precise
way in which the two sets of participants described
various aspects of asthma: the analysis centred on
the use of language in their testimony. The research
questions became: “Do patients and professionals
describe asthma in different ways?”and “What can
be inferred from this data about the knowledge base
upon which these expressions are constructed?”
The study was convened in the South West of
England, with approval from the local ethical com
mittee. Two sets of focus groups were convened in
parallel, four comprising professionals (doctors and
nurses), and four drawn from patients with asthma.
The professional groups comprised one separate
group each of specialist doctors, secondary care
nurses, general practitioners and practice nurses.
Individuals with asthma were identified from
general practice disease registers. To obtain a suffi
cient spread of patients with the type of asthma
seen routinely in general practice, the sampling
frame was stratified by age, and use of inhaled ster
oids (which was used as a proxy indication of
asthma severity). The age bands were 16-44, and
Appendix
141
Siveeney, Edicards, Stead, Halpin
21
45-65. Inhaler use was dichotomised as regular
(defined as receiving a prescription for a steroid
inhaler monthly over a period of twelve months) or
infrequent (fewer received steroid inhaler prescrip
tions). Thus, one group each of young citizens who
had received regular or infrequent steroid inhalers
was convened, alongside two similar groups of older
patients. None of the participants had experienced
a hospital admission for asthma within the preced
ing year.
The groups, lasting 60-90 minutes, were con
vened between October 1997 and April 1998
according to published guidelines” ” and were
facilitated by a researcher with experience of quali
tative research.26 The following questions were
explored in each group: What is asthma? What are
the treatments of asthma and how do they
work/what do they do? What is it like to have
asthma, or what must it be like? What arc good
outcomes of asthma treatment?
Data in the focus groups were recorded by
audiotape with additional hand-written notes con
stituting a contact summary. A two-stage analysis
was used. Firstly, a content analysis elicited the fre
quency with which terms were used in the partici
pants’ descriptions. Then a second-stage analysis
identified the key conceptual themes in the data,
linking them together in categories where appropri
ate. Once these categories had been completed, the
researchers considered what type of knowledge
might have been used in their construction, and
also any power relationships suggested by their
comparison. The trustworthiness of the coding
frames was strengthened by review of an unmarked
transcript by two experienced qualitative research
ers and by presentation of the initial analysis to
participants, who offered their comments on the
analysis. Once the data analysis was completed, the
initial findings were reviewed by one author (KE),
an academic literary scholar, who commented spe
cifically on the use of metaphor.
Results
COMPOSITION OF THE FOCUS GROUPS
Two nurse groups were convened, each consisting
of six participants. All the general practice nurses
participated in the shared care of asthma. The two
doctor groups consisted of seven specialists,
ranging from registrar to consultant, and five
general practitioners, all of whom were trainers.
Between four and seven participants attended each
patient focus group. In total twenty-two partici
pants attended the patient groups, nine of whom
were men. The age range was 16-44, and 45-65
years. In all the groups, the participants were Cau
casian and English-speaking.
Data analysis
WHAT IS ASTHMA?
In the professional groups, the term “inflamma
tion” occurred more frequently than any other term
and incorporated the notions of “swelling”,
“oedema” and “obstruction”. Less frequently used
www. luedicalhtnnanities. com
Supplied by The British Library - "The world's knowledge"
I 42
Complexity in primary care
22 A comparison ofprofessionals’ and patients’ understanding of asthma: evidence of emerging dualities?
terms were “narrow”, “constrict”, “smaller” and brand names appear in the patients’ remarks
“tight”. The doctors’ groups described a detailed reported here. There was a clear consensus about
pathophysiological pathway for asthma which the effect of Ventolin: “The quickest relief. Less
included descriptions of “leukotrienes” and “cy than a minute. Straight away” (older female
tokines”. One hospital doctor said: “1 mean poten infrequent use). The role of Becotide was less clear.
tially it involves probably formal components of “It’s a medication which soothes” (young female,
your inflammatory pathway, so probably certain frequent use). “In the theory, I believe, I think it
parts can be switched off... some are more promi coats your lungs” (younger male, frequent use).
nent than others” (hospital doctor).
The patient groups confirmed the professionals’
For the patients, the terms “constrict”, “narrow” concerns about apprehension and understanding of
and “tight” predominate as descriptors of asthma. steroids. Fear of the use of steroid inhalers was
“Inflammation” was used less frequently and only repeatedly described by the patients, who linked the
once in the group of younger asthmatics using use of Becotide with suppression of disease activity
regular inhalers. The descriptions of asthma in and with a decreased need for regular Ventolin use.
these groups tended to be less conventionally The intermittent use of Becotide was a clear theme:
biomedical, but obviously made sense to the “I only take Becotide when I’ve got something hap
participants, none of whom were challenged or pening I can’t possibly miss” (younger female,
derided for having idiosyncratic visions of what regular inhaler use). Often, an individual’s decision
asthma was. One participant, for example said: “It to use Becotide was based upon a previous experiseems more tight here just above my. lungs,
„ , and ence of it, (and not, as the professionals might have
someone’s clenching them. This sounds ridiculous hoped, on its therapeutic rationale). “I take it if I get
hnr
Qn l0 my broWHS three Or four
but if
it feels as if
if thpr*»
there’’cs na load r»f
of /varpor
carpet in
in them”, bad or get a
(younger group, male, regular use of inhalers)
days then it works. All the bad stuff comes towards
The dominant metaphor in the professional data it and bounces off it or gets eaten possibly. Maybe
in this section was of processes, for example “show absorbed” (younger male, frequent inhaler use),
use).
pictures of swollen airway” (specialist nurse), or in This sometimes led individuals to act against medithe use of the phrase “path way ... so probably cer- cal advice to use Becotide continuously. “When you
tain parts can be switched off”. For the patients the haven’t got a tight chest you haven’t got asthma, I
dominant metaphor was of containers : “It’s like a just forget to take Becotide” (older group, male,
windsock” (younger, infrequent use) “Hubbly bub- infrequent inhaler use). Here, the participants seem
bly pipes,” (younger, frequent use).
to be constructing treatment plans retrospectively,
based upon their accumulated experience of
asthma.
WHAT ARE THE TREATMENTS FOR ASTHMA AND HOW
rKi— j-------- :_________ _ _______ .
t-
t_ _
-
- -
—- -
DO THEY WORK?
The professional groups identified beta stimulant THE EXPERIENCE OF ASTHMA: WHAT IS IT (OR MUST
drugs and corticosteroids as the essential drugs for IT BE) LIKE?
treating asthma. As one specialist nurse said: “one Both the .professionals and the ,patients agreed
„
that
opens up the airway, the other stops it closing”. The asthma was a stigma which could cause embarrassprofessional groups stressed the prospective ben- ment and restrictions in sporting, and some social,
efits of using corticosteroids: “one makes you better activities. This led some patients, in the view of the
at the time, the other keeps you better for 1professionals,. to resist the diagnosis:
„
“They
. ’re not
tomorrow” (general practitioner). Thus, treatment prepared to play the sick role, they deny the
strategies could or should be planned prospectively, diagnosis.
j:
” (general
'.2
' .practitioner)
''
' Teenag(
jers were
The professional groups saw steroids as central to more likely to do this, perhaps because, “they
tl
have
disease control, but recognised that patients were no sense of their own mortality do they?” (commusometimes apprehensive about using them: nity nurse). Asthma could be frightening, embar“They’re [patients]
„ ---- ’worried
------ J when
L
,you
“)u say this is a rassing and impair social activities, the professionsteroid inhaler and you give quite a little spiel about als suggested.
this Effincr
rhk
being quite safe” (general practitioner). SomeThe patients’ data produced some poignant
times patients did not fiilly “understand” their role, descriptions of the stigma and embarrassment
—:j__ .this
v:--exchange:
1-------attached (in their view) to having asthma. Some
Consider
patients felt they should use their inhalers
(Nurse speaker 3): “It’s quite interesting, they don’t times,
out of sight. “I’d sneak out on me (sic) own to use
seem to be aware of the significance of their my inhalers so nobody could see it.” (older male,
preventer, like you say.”
infrequent use of inhaler) “I go to the loo” [to use
(Nurse speaker 4) “It’s because they didn’t think inhalers] (younger male, frequent inhaler use)..One
they needed it.”
individual reported: “Asthmatics at school ■were
wimps. I don’t want to be thought of as an invalid,”
(Nurse speaker 3) “Yes, that’s the commonest (older male, infrequent inhaler use).
thing.”
a** M M aaK a* + *
aA * A
aJ ! aJ
— & £- - 1 l_ _
C4 __ _J
1 *9
. fl
*
fl
The patients’ groups identified salbutamol and
beclomethasonc as the two main treatments. The
patients referred to 0ese by their brand names
CVenfnlin
(Ventolin nnri
and RprntiHp
Becotide rAcnr'rf
respectively) and
and these
HOW DO YOU MANAGE (YOUR) ASTHMA?
Within the professionals’ data,fa clear responsibility
fnt- educating patients
for
'jSticnts about asthma emerged,
emerged. With
with
nurcAc
__ i
nurses seen as key players. Within this educational
titvu'. medicaUntmanities. com
Supplied by The British Library - "The world's knowledge"
Appendix
143
Sueeney, Edwards, Stead, Halpin
23
Table 1
What is asthma? Content analysis, main
descriptors
Dominant metaphor
Treatment perspective
Managing asthma
Locus of control
Taxonomy of knowledge used in
constructing the model of asthma
Professiona! group
Patiott group
“Inflammation”, “swelling”, “oedema",
“obstruction”
Process
Prospective
Prior professional responsibility assumed,
and then transferred
Operational (Piaget)
Universal (Toulmin)
“Constrict”, “narrow”, “tight”
Container
Retrospective
Patient as experts : “I know my asthma
better than anybody"
Figurative (Piaget)
Existential (Toulmin)
theme, the nurses described a relationship between with “difficult patients, where you have to allow
professionals and patients which most closely them to take responsibility”.
resembled a teacher-pupil relationship. For exam
ple, when one professional commented on some WHAT DO YOU THINK ARE GOOD OUTCOME
patients refusing “to go to classes”, (community MEASURES FOR THE TREATMENT OF ASTHMA?
nurse) the contact summary confirmed a group The specialist groups easily rehearsed the conven
consensus. Within this teacher-pupil relationship, tional outcome measures for asthma: absence of
the professional groups perceived the need to keep wheeze, absence of cough, absence of early
clinical messages “simple”: “I don’t mean to be morning waking, ability to perform exercise, and
rude, but I mean you need to make it as basic as not having to take time off work.
possible,” (specialist nurse).
The patient groups did not specifically identify
In one of the early patient groups an individual these conventional outcome measures in detail.
spontaneously reported: “I think you can generally Rather, they reported that they did not want to be
advise yourself... . I think I can advise myself embarrassed or fearful of their asthma, nor
better.”(younger male, frequent inhaler use). The inconvenienced by it. “I want a totally normal life”
facilitator tested this piece of testimony on that (younger female, infrequent inhaler use), said one
group and subsequent groups, and the contact group member. “Just not to have to use inhalers,”
summary confirms that there was a strong feeling (younger male, infrequent inhaler use) commented
among the individuals that they were experts in another. On the whole, individuals in these citizen
their own disease. “I know my asthma better than groups felt they could sense any deterioration in
anybody.” (older female, frequent inhaler use). This symptoms themselves. Little was added to that
data evokes a tension in roles: the professionals with impression by recording the peak flow rate. “I only
their desire to act as teachers, and the patients’ do peak flow rate to see the nurse or the doctor.”
sometimes distressing descriptions of their own (older participant, frequent use of inhaler).
expertise.
We summarise the main points from these results
A subsidiary theme within this category sug in table 1.
gested that patients may be the victims of their own
actions. “Some have really brought it upon
themselves because they smoke,” (practice nurse). Discussion
Referring to patients whose asthma was poorly There are a number of drawbacks to this study.
controlled one practice nurse commented: “you Firstly, focus groups, while appropriate initially,
have allowed it to take over.” Many of the when differences in outcome were being explored,
participants agreed that their asthma was not are probably not the best way to elicit differences in
always well controlled. Some blamed themselves the use of language. We accept this, but argue that
when they experienced exacerbations of asthma “I within these focus group, the patients were able to
feel guilty getting bad and having to go to the doc put forward ideas in their own words which the
tor quick—to get a nebuliser—you know. I hang on group could evaluate—for example by criticising a
and hang on.” (younger female, infrequent inhaler particularly peculiar vision of asthma. The field
use). The concept of guilt associated with exacerba notes bear this out: the focus groups actually were
tions was vivid, for example, when the deterioration “permission-giving” fora in which participants felt
occurred at a friend’s house thus causing inconven free to offer their private views when they saw that
ience). Some admitted not wanting to interrupt others were willing so to do. Secondly, the
their normal planned activities, or assumed that the participants in these citizen groups could all be said
deterioration they were experiencing would be to have accepted their diagnosis: Adams et al have
written about the important group of asthmatics
transitory.
By contrast, discussion of the locus of control in who deny their disease.27 One could speculate that
asthma by the professional group suggested prior different data might have emerged from groups of
ownership of responsibility for the disease by the such participants. Finally, no particular themes
to distinguish the understanding of
professional before returning or yielding it back to emerged
----- _
the patient. “We have to put control onto the person ' asthma as a function either of age or inhaler use, the
two strata upon which the groups «were mnvpnpd
convened.
themselves (sic) to manage it,” commented one -----------practice nurse, a proposition which could be tricky However, several key ideas which emerged from our
titctu titedicalhuinanities. coin
Supplied by The British Library - "The world's knowledge"
L
I 44
Complexity in primary care
24 A comparison of professionals’and patients’ understanding of asthma: evidence of emerging dualities?
data fit with previous published work in this area,
for example the notion of patients as experts, and
the stigma attached to having asthma.2’
In summary, this study has addressed two ques
tions: Do professionals and patients describe
asthma in different ways, and what can be inferred
from this data about the knowledge base upon
which these expressions are constructed? We
address each of these in turn.
DO PROFESSIONALS AND PATIENTS DESCRIBE ASTHMA
IN DIFFERENT WAYS?
The data show clear areas of congruence or shared
understanding between the professionals and the
patients: both groups broadly agreed on the
treatments for asthma, and agreed that there were
difficulties in the use of inhaled steroids. Within the
set of data describing the role of steroids, a paradox
begins to emerge, elicited particularly by a doctor’s
use of the phrase “keeps you better for tomorrow”.
In general, the data suggest that the professionals
stressed the prospective benefits from using ster
oids. Patient participants evaluated the efficacy of
the inhaled steroid in the context of their accumu
lated personal experience of it—ie retrospectively—
and judged its value accordingly. We interpret this
as evidence of a divergence or “duality” in respect
of the two perspectives upon which their respective
treatment strategies are based.
A second paradox emerges in the data describing
expertise in asthma, and locus of control. A clear
theme in the patients’ data suggests a developing
personal expertise in asthma management, allowing
the patient “to advise myself better”, or “know my
asthma better than anybody”. The health care pro
fessionals on the other hand, clearly feel a sense of
responsibility for managing asthma for patients:
they adopt the role of teacher and assume initial
responsibility for the disease before deciding, at
limes, to “put control onto the person”. We suggest
this might represent a further duality in relation to
disease management.
We postulate that these differences in approach
to treatment and management imply deeper differ
ences in the ways the two groups think about
asthma. That these differences are linked to
language is supported by the different metaphors—
process versus container noted in their descriptions
of the disease. Such a disparity between doctors
and patients in the use of metaphor when discuss
ing asthma has recently been highlighted," and
reflects an increasing interest in language-based
medical research.”
WHAT CAN BE INFERRED FROM THIS DATA ABOUT
THE KNOWLEDGE BASE UPON WHICH THESE
EXPRESSIONS ARE CONSTRUCTED?
This analysis supports the possibility that there is
an epistemological difference between doctors and
patients in this context. While some might consider
this a truism, that the two have different perspec
tives, we postulate that this may reflect a more pro
found distinction, namely that each group draws on
different types of knowledge to construct the
thoughts and words with which they describe
asthma. The possibility of such a distinction is rec
ognised in the philosophical literature. We can
begin to understand these differences by drawing
on Piaget’s distinction between figurative and
operational knowledge, referred to earlier." The
professionals, who in general do not have direct
experience of actually having asthma, use their
theoretically based knowledge to participate in the
dialogue. Piaget would call this type of knowledge
operational. The patients’ knowledge arises out of
direct lived experience, which Piaget classes as
figurative. Although these categories are not meant
to be mutually exclusive, it does introduces the
possibility, arising from the data, that the two
groups construe asthma in slightly different ways.
Using Toulmin’s taxonomy,20 the professionals
draw on universal knowledge, while their patient
counterparts use an “existential knowledge base”.
Again, Byron Good argues that medicine reconsti
tutes the familiar human body as “the medical
body”: doctors see the human body in medicine’s
own way, with medicine’s gaze, and use a particular
vocabulary to describe it.”’1 Within medicine,
Good argues, the human body is “newly consti
tuted as a medical body, quite distinct from the
bodies with which we interact in everyday life”.20
We argue that in this data set, the words, constructs,
and strategies described by the two groups reflect
the epistemological dualities postulated by these
commentators.
What are the implications of such a view? Firstly,
it underlines the relevance of the philosophical
literature to everyday clinical practice. Secondly, it
demands a “post-modern” view of illness in which
the biomedical component is inextricably inter
twined with the cultural, societal aspects: no one
component is more robust, more real or more
relevant. The experience of illness occurs at their
intersection.’2” Thirdly, it introduces a potential
impediment to implementation: if consultations
cannot be seen as literally shared dialogues, but
rather as a kind of fluid elusive exchange of extracts
from differing knowledge bases, the process of
implementation may not be a simple linear process
(that is just explaining and handing over infor
mation), but may be more complex.
Further research in this area would refine the
analytical approach in this study by examining
video recordings of live consultations, and inter
viewing the participants to find out precisely what
they were thinking when they made a contribution.
Language-based medical research should be en
couraged to explore the epistemological basis of
patients’ and health care professionals’ interactions
during consultations.
Authors’ note
Drs Halpin and Stead had the original idea for the
project, assisted in its planning and commented on
the article through each of the drafts. Dr Sweeney
carried out all the focus groups work and the first
level data analysis. Dr Edwards carried out the
WTffw. medicalhumanities.com
Supplied by The British Library - "The world's knowledge"
analysis of the use of metaphor, and commented on
the successive drafts of the paper.
Funding
This study arose out of a larger study funded by the
NHS National R and D Programme on Asthma
Management.
Acknowledgments
We wish to thank Mrs Helene Talbot for her help in
identifying the patients and administering the
collation of the focus groups, Mrs Annie Hills for
her secretarial support, and Drs Sindy Banga and
Maggie Cormack for their helpful analysis of a
sample of the transcripts.
K G Sweeney, MA, MPhil, MRCGIj is Lecturer,
Research and Development Support Unit, School of
Postgraduate Medicine and Health Sciences, University
of Exeter. Karen Edwards, PhD, is Lecturer, School of
English and American Studies, University of Exeter. Dr
Jonathan Stead, MPhil, MRCGP, is Consultant in
Clinical Effectiveness, North and East Devon Health
Authority. Dr David Halpin, DPhil, MRCP, is Senior
Lecturer in Respiratory Medicine, School of Postgradu
ate Medicine and Health Sciences, University of Exeter.
I
References
1 Hilton S, Anderson HR, Sibbald B, Freeling P. Controlled
evaluation of the effects of patient centred education on asthma
morbidity in general practice. Lancet 1986;i:26-9.
2 Turner-VC'anvick M. Nocturnal asthma: a study in general
practice. Journal of the Royal College of General Praaitioners
1989;39:239-43.
3 Horn CR, Cochraine GM. An audit of morbidity associated
with chronic asthma in general practice. Respiratory Medicine
1989;83:71-5.
4 Dudbridge SB, Stead J, Ward S, Amooie M, Halpin DMG. The
Devon primary care asthma initiative: effects on morbidity.
American Journal ofRespiratory Critical Care Medicine 1996;1S3:
A868.
5 Charlton IH, Charlton G, Broomfield J, Mullee MA. The effect
of nurse run asthma clinic on workload and patient morbidity
in general practice. British Journal of General Practice 1991;41:
227-31.
6 Jones KP, Bain DJG, Middleton M, Mullee MA. Correlates of
asthma morbidity in primary care. British Medical Journal
1992;304:361-4.
7 Jones KP, Charlton IH, Middleton hi, Preece WJ, Hill AP. Tar
geting asthma care in general practice using a morbidity index.
British Medical Journal 1992;304:1353-6.
8 O’Connor GT, Weiss ST. Clinical and symptom measures.
American Journal of Respiratory Critical Care Medicine 1994;149:
S21-8.
Appendix
145
Sweeney, Edwards, Stead, Halpin
25
9 Morris NV, Abrahamson MJ, Srrusscr RP. Adequacy of control
of asthma in general practice. Is maximum peak expiratory flow
rate a valid index of asthma severity? Medical Journal of
Auitralia 1994;160:68-71.
10 Apter AJ, Zu Wallack RL, Clive J. Common measures of
asthma severity lack association for describing its clinical
course. Journal of Allergy and Immunology 1994;94:732-7.
11 Atherton A, Tyrrell Z, White PT. Rrccption and reported use of
inhaled steroids in primary care [report to Department of
Health]. London: King’s College, 1998.
12 Adams 8, Pill R, Jones A. Medication chronic illness and
identity: the perspective of people with asthma. Social Science
and Medicine 1997;45:189-202.
13 Ostergaard MS. Childhood asthma: parents’ perspective, a
qualitative interview study Family Practice 1998;15:153-7.
14 Johnson M, backoff G. Metaphors tee live by. Chicago
University of Chicago Press, 1980.
15 Mabeck CE, Oleson E Metaphorically transmitted diseases.
How do patients embody medical explanations? Family Practice
1997;14:271-8.
16 Leary D, ed. Metaphors in the history of psychology. Cambridge
Univereity of Cambridge Press, 1994.
17 Beate D, Skorpen J, Malterud K. What did the doctor say-what
did the patient do: operational knowledge in clinical communi
cation. Family Practice 1997;4:376-81.
18 Piaget J. The language and thought of the child. London, Kegan
Paul, Trench Trubener, 1932.
19 Silverman H J, ed. Piaget, philosophy and the human sciences.
Evanston, Illinois: Northwestern University Press, 1980.
20 Toulmin S. Knowledge and art in the practice of medicine:
clinical judgment and historical reconstruction. Im
Delkeshamp-Hayes C, Gardell Cutter MAG, cds. Science, tech
nology and the art of medicine. European-American Dialogues.
Dordrecht: Kluwer, 1993.
21 Polanyi M. Personal knowledge. London: Routledge and Kegan
Paul, 1973.
22 Morgan DL. Focus groups in qualitative research. London: Sage
Publications, 1988.
23 Dcnzin NK, Lincoln YS, eds. Handbook of qualitative research.
Thousand Oaks, California: Sage, 1994.
24 Holloway I. Basic concepts in qualitative research. London:
Blackwell Press, 1997.
25 Miles MP, Hubermann AM. Qualitative analysis: an expanded
source book. London: Sage Publications, 1994.
26 Sweeney KG. The use offocus groups to assess consumers’vietes on
miscarriage and minor surgery [MPhil thesis]. Exeter: University
of Exeter, 1996.
27 Adams S, Pill R, Jones A. Medication, chronic illness and
identity: the perspective of people with asthma. Soaal Science
and Medicine 1997;45:189-210.
28 Goffman E. Stigma: notes on the management of spoiled identity.
Harmondsworth: Penguin, 1968.
29 Skelton JR, Hobbs FDR. Concordancing: use of language
based research in medical communication. Lancet 1999;353:
108-11.
30 Good B. Medicine, rationality and experience. Cambridge: Cam
bridge University Press, 1994.
31 Evans M The “medical body" as philosophy’s arena. Theoretical
A ledicine and Bioethics 2001 (in press).
32 Morris D. illness and Culture in the Postmodern Age. California,
University of California Press, 1998..
33 Muir Gray JA. Postmodern medicine. Lancet 1999;354:155O13.
www. tnedicalliuinanities. com
Supplied by The British Library - "The world’s knowledge"
L
1
References and further reading
References
Adorno T, Albert H, Dahrendors R, Habermas J, Piloth B, Popper K (1976) The Positivist
Dispute in German Sociology. Heinemann, London.
American Board of Internal Medicine (1985) Subcommittee of the Evaluation of Humanistic
Qualities in the Internist. A guide to awareness and evaluation. American Board of Internal
Medicine, Washington, DC.
Anon. (1995) Evidence-based medicine, in its place. Lancet. 346: 785.
Armenian HK, Lilienfield DE (1994) Applications of the case-control method: overview
and historical perspective. Epidemiol Rev. 16: 1-5.
Armstrong D, Reyburn H, Jones R (1996) A study of general practitioners' reasons for
changing their prescribing behaviour. BMJ. 312: 494-5.
Aron R (1970) Main Currents in Sociological Thought. Penguin, Harmondsworth.
Ashby R (1952) Design for a Brain. John Wiley & Sons, New York.
Bacon F (1858) The Works of Francis Bacon, Baron ofVerulam, Viscount St Alban, and Lord High
Chancellor of England. Longman & Co., London.
Bailey C (1947) Titus Lucretius Carus: de natura rerum. Oxford University Press, Oxford.
Balint M (1957) The Doctor, his Patient, and the Illness. International Universities Press, New
I
i
York.
Battram A (1998) Navigating Complexity. The Industrial Society, London.
Baxt WG (1994) Complexity chaos and human physiology: the justification for non-linear
neural computational analysis. Cancer Lett. TT: 85-93.
Becker MH (1974) The health belief model and sick role of behaviour. Health Educ Monogr.
2:409-19.
Berkeley G (1967) The Principles of Human Knowledge. Fontana, London.
Bloor M (1976) Bishop Berkeley and the adeno-tonsillectomy enigma: an exploration in
the social construction on medical disposals. Sociology. 10: 43-51.
Boston Area Anticoagulation Trial for Atrial Fibrillation Investigators (1990) The effects of
low-dose warfarin on the risk of stroke in patients with non-rheumatic atrial fibrillation.
NEJM. 323: 1505-11.
Bradley F, Field I (1995) Evidence-based medicine (letter). Lancet. 346: 838-9.
Briggs J, Peat FD (1989) Turbulent Mirror. Harper and Row, New York.
British Medical Association (2003) New GMS Contract. Investing in general practice. NHS
Confederation, London.
Brock W (1985) From Prototype to Proton: William Prout and the nature of matter, 1785-1985.
Hilger, Bristol.
Brock WH (1993) The biochemical tradition. In: Bynum WF, Porter R (eds) Companion
Encyclopaedia of the History of Medicine. Routledge, London.
Brook R, Kosecoff JB, Park RE et al. (1988) Diagnosis and treatment of coronary disease: a
comparison of doctors' attitudes in the UK and USA. Lancet. 1: 750—3.
Brown JR (2001) Who Rules in Science? Harvard University Press, Cambridge, MA.
Bunker J (1995) Medicine matters after all. J R Coll Physicians. 29: 105-12.
Burke P (1981) Montaigne. Oxford University Press, Oxford.
Burnand B, Kernan WN, Feinstein AR (1990) Indexes and boundaries for 'quantitative
significance' in statistical decisions. J Clin Epidemiol. 43: 1273-84.
L
1
148
■■
i'
Complexity in primary care
Bynum WF, Porter R (eds) (1993) Companion Encyclopaedia to the History of Medicine.
Routledge, London.
Byron A, Boyd C (1991) Rudolf Virchow: the scientist as citizen. Garland, New York.
Cannon WB (1939) The Wisdom of the Body. Norton, New York.
Capra F (1983) The Turning Point: science, society and the rising culture. Flamingo, New York.
Capra F (1996) The Web of Life. Doubleday, New York.
Cartwright A, Anderson R (1981) General Practice Revisited: a second study of patients and their
doctors. Tavistock, London.
Casti JL (1995) Complexification. Harper Perennial, New York.
Centre foi Disease Control (2000) Centre for Disease Control Daily News (accessed 22
September 2000); www.ama-assn.org/special/hiv/newsline/cdd/091800g3.htmI .
Chalmers AF (1982) What is This Thing Called Science? Open University Press, Buckingham.
Charlton BG (1995a) Megatrials are subordinate to medical science (letter). BMJ. 311:257.
Charlton R (1995b) Balancing science and art in primary care research: past and present. Br
J Gen Tract. 45: 639-40.
Cilliers P (1998) Complexity and Postmodernism. Routledge, London.
Cochrane AL (1971) Effectiveness and Efficiency: random reflections on health services. Nuffield
Provincial Hospital Trust (Rock Carling Fellowship), London.
Cohen J, Stewart I (1994) The Collapse of Chaos. Penguin, Harmondsworth.
Collingwood RG (1946) The Idea of History. Clarendon Prhss, Oxford.
Collins R, Peto R, Gray R et al. (1996) In: Weatherall DJ, Ledingham JGG, Warrell DA (eds)
The Oxford Textbook of Medicine. Oxford University Press, Oxford.
Comte A (1875) (trans. Martineau H) (1974) The Positive Philosophy. Ams Press, New York.
Comte A (1993) In: Microsoft Encarta Encyclopaedia. Funk and Wagnails, Seattle.
Connolly SJ, Laupacis A, Gent M et al. (1991) Canadian Atrial Fibrillation Anticoagulation
(CAFA) Study. J Am Cardiol. 18: 349-55.
Creswell JW (1988) Qualitative Enquiry and Research Design. Sage Publications, London.
Cromarty I (1996) What do patients think about during their consultations? A qualitative
study. Br J Gen Tract. 46: 525-8.
Crosby A (1997) The Measure of Reality: quantification in Western society 1250-1600. Cambridge
University Press, New York.
Cunningham A (1989) Thomas Sydenham: epidemics, experiments and the good old
cause. In: French R, Wear A (eds) The Medical Revolution in the Seventeenth Century.
Cambridge University Press, Cambridge.
Darlington S (2000) Brazil Becomes Model in AIDS Fight (accessed 7 November 2000);
www.aegis.org/news/re/2000/re001107.html.
Davidoff F, Haynes B, Sackett D et al. (1995) Evidence-based medicine (editorial)
BMJ. 310: 1085-6.
Dawes MG (1996) On the need for evidence-based general and family practice. EvidenceBased Med. 1: 68-9.
Delblanco A (2001) Night vision. Review of The Moral Obligation to be Intelligent: selected
essays by L Trilling. New York Rev Books. 48.
Denzm NK, Lincoln YS (1994) A Handbook of Qualitative Research. Sage Publications,
Thousand Oaks, CA.
Department of Health (1962) A Hospital Plan for England and Wales. HMSO, London.
Department of Health (1972) Report of the Working Party on Medical Administrators. The Grey
Book. HMSO, London.
Department of Health and Social Security (1983) NHS Management Enquiry: Griffiths Report.
Department of Health and Social Security, London.
Descartes R (trans. Veitch A) (1912) A Discourse on Method: meditations and principles. Dent &
Sons Ltd, London.
Devor E J (1993) Genetic disease. In: Kiple K (ed.) The Cambridge History of Human Disease.
Cambridge University Press, Cambridge.
1
References and further reading
149
Oilman I (1973) Induction and Deduction: a study in Wittgenstein. Basil Blackwell, Oxford.
Dixon M, Sweeney KG (2000) The Human Effect in Medicine: theory, research and practice.
Radcliffe Medical Press, Oxford.
Doll R, Hill AB (1950) Smoking and carcinoma of the lung: preliminary report. BMJ. ii:
1225-36.
Dubos R (1960) The Mirage of Health. George Allen 8- Unwin, London.
Dunn J, Urmson JO, Ayer AJ (1992) The British Empiricists. Locke, Berkeley, Hume. Oxford
University Press, Oxford.
Durie R, Wyatt K, Fox M, Sweeney K (2004) Receptive Context in the Pursuing Perfection
Programme. Report for the Modernisation Agency of the Department of Health. Health Complex
ity Group, Exeter.
Ellis J, Mulligan I, Rowe J et al. (1995) In-patient general medicine is evidence based.
Lancet. 346: 407-10.
Elstein AS, Kagan N, Hulman D et al. (1972) Methods and theory in the study of medical
inquiry. J Med Educ. 47: 85-92.
Engel GL (1977) The need for a new medical model: a challenge for biomedicine. Science.
196: 129-36.
European Atrial Fibrillation Study Group (1993) Secondary prevention in non-rheumatic
atrial fibrillation after transient ischaemic attack or minor stroke. Lancet. 342: 1255-62.
Evans M, Sweeney KG (1998) The Human Side of Medicine. Occasional Paper No. 72. Royal
College of General Practitioners, London.
Evidence-Based Medicine Working Group (1992) Evidence-based medicine: a new
approach to teaching the practice of medicine. JAMA. 268: 2420-5.
Ezekowitz MD, Bridgets SL, James KE et al. for the Veterans Affairs Stroke Prevention in
Non-Rheumatic Atrial Fibrillation Investigators (1992) Warfarin in the prevention of
stroke associated with non-rheumatic atrial fibrillation. NEJM. 327. 1406—12.
Fahey T, Schroder K (2004) Recent advances in primary care: cardiology. BrJGen Pract. 54:
696-703.
Feinstein AR (1992) Insidious comparisons and unmet clinical challenges. Am J Med. 92.
Feinstein AR (2002) Will clinicians' challenges be solved by another theoretical model?
Commentary on Sweeney and Kernick (2002) Clinical evaluation: constructing a new
model for post-normal medicine. J Clin Evaluation. 8: 139-41.
Fishbein M, Azjen I (1975) Belief, Attitude, Intention and Behavior. John Wiley 8 Sons, New
York.
.
Fletcher R (1971) The Making of Sociology: a study of sociological theory. Volume 1. Michael
Joseph, London.
Foucault M (1963) The Birth of the Clinic. Tavistock, London.
Fowler PBS (1995) Evidence-based medicine (letter). Lancet. 346: 838.
Fraser S, Greenhalgh T (2001) Coping with complexity: educating for capability. BMJ. 323:
799-802.
Fraser SW, Conner M, Yarrow D (eds) (2003) Thriving in Unpredictable Times: a reader on
agility in health care. Kingsham Press, Chichester.
Freeman A, Sweeney K (2001) Why general practitioners do not implement evidence: a
qualitative study. BMJ. 323: 1100-14.
Freund J (1968) The Sociology of Max Weber. Allen Lane, London.
Fukuyama F (1993) The End of History and the Last Man. Penguin, Harmondsworth.
Garfinkel A, Spano ML, Ditto WL, Weiss JN (1992) Controlling cardiac chaos. Science. 257:
1230-5.
Garratini S, Garratini L (1993) Pharmaceutical prescribing in four European countries.
Lancet. 342: 1 191-2.
Geyer R (2001) Beyond the Third Way: the science of complexity and the politics of choice. Paper
presented to the Joint Session of the ECPR, Grenoble, April 2001.
r
150
i
Complexity in primary care
Giddens A (1974) Positivism and Sociology. Heinemann Educational, London
Gill P, Dowell AC, Neal RD et al. (1996) Evidence-based general practice: a retrospective
study of interventions in one training practice. BMJ. 312: 819-21.
Gillies J (2005) Getting it Right in the Consultation. Hippocrates' problem. Aristotle's
answer.
Occasional Paper No. 86. Royal College of General Practitioners, London.
Gillon R, Wesley S (1998) Medicalisation of distress. R Soc Arts J. 4: 80-6.
Gleick J (1998) Chaos. Vintage, London.
Gloubermann S, Zimmerman B (2002) Complicated and Complex Systems: what would successful
reform of Medicare look like? Discussion Paper No. 8. Commission on the Future of Health
Care in Canada, Montreal.
Goldberger AL, West BJ (1987) Applications of non-linear dynamics to clinical cardiology
Ann N Y Acad Sci. 504: 195-213.
Grbich C (1999) Qualitative Research in Health. Sage Publications, Thousand Oaks, CA.
Greaves D (1996) Mystery in Western Medicine. Ashgate Publishing Ltd, Aidershot.
Greenhalgh T (1996) Is my practice evidence-based? BMJ. 313: 957-8.
GreenhalghT(2002) Intuition and evidence: uneasy bedfellows? Br J Gen Pract. 52: 395-400.
Gribbin J (2004) Deep Simplicity. Allen Lane, London.
Griffiths F (2002) Complexity and primary healthcare research. In: Sweeney K, Griffiths F
(eds) Complexity and Healthcare: an introduction. Radcliffe Medical Press Oxford
Grimshaw JM, Russell IT (1993) Effect of clinical guidelines on medical practice: a
systematic review of rigorous evaluations. Lancet. 342: 1317-22.
Gl'lIbert J-’ (1987) Educational objectives. In: Educational Handbook for Health Care Personnel
World Health Organization, Geneva.
Guyatt GH, Rennie D (1993) Users' guides to the medical literature. JAMA 270- 2096-7
Guyatt GH, Sackett DL, Cook DJ (1994) Users' guides to the medical literature. II. How to
use an article about therapy or prevention. JAMA. 271: 59-63.
Haines A, Jones R (1994) Implementing findings in research. BMJ. 308: 1488-92.
Hammer M (1995) The Re-Engineering Revolution. Harper Business, New York
Haraway DJ (1976) Crystals. Fabrics and Fields: metaphors of organicism in twentieth-century
developmental biology. Yale University Press, New Haven, CT.
Harrison S (1988) Managing the NHS: shifting the frontier? Chapman and Hall, London.
Harrison T (1947) The future of sociology. Pilot Papers. 2: 10-21.
Hassey A (2002) Complexity and the clinical encounter. In: Sweeney K, Griffiths F (eds)
(2002) Complexity and Healthcare: an introduction. Radcliffe Medical Press, Oxford
H pTe^Cambiidg^
arid DeSpair: a history °f so^logy. Cambridge University
Health Complexity Group (2004) Creating an Enabling Context for Change: linking research to
practical examples of quality improvement. Workshop presented at the European Qualitv
Forum, Copenhagen, June 2004.
Hcaib I (1995) The Mystery of General Practice. Nuffield Provincial Hospitals Trust. London,
eisenberg W (1971) Physics and Beyond. Harper and Row, New York.
Holland JH (1998) From Chaos to Order. Oxford University Press, Oxford.
Holloway I (1997) Basic Concepts for Qualitative Research. Blackwell Science, Oxford.
Holt T (2002a) A tight model for diabetes control. Diabet Med. 19: 274-8.
Holt T (2002b) Clinical knowledge, chaos and complexity. In: Sweeney K, Griffiths F (eds)
Complexity and Healthcare: an introduction. Radcliffe Medical Press, Oxford.
Hondericb T (ed.) (1995) The Oxford Companion to Philosophy. Oxford University Press,
x 1 o ru.
Howie JGR (1972) Diagnosis: the Achilles' heel? J R Coll Gen Pract. 22: 310-15.
Howie JGR, Bigg AR (1980) Family trends in psychotropic and antibiotic prescribing in
general practice. BMJ. 280: 836-8.
HUISoh7nS *195.9? Consciousness and Society- ^e real orientation of European social thought 18901930. MacGibbon and Kee, London.
*
References and further reading
I5I
Hume D. Enquiries Concerning Human Understanding and Concerning the Principles of Morals
(3e) (1975). Oxford University Press, Oxford.
Hume D (1739) A Treatise on Human Nature (2e) (1978). Clarendon Press, Oxford.
Iggo N (1995) Evidence-based medicine (letter). Lancet. 346: 839.
Illich I (1975) Medical Nemesis: the expropriation of health. Calder and Boyars, London.
Johnson M (1995) Health-related behaviour change. In: Preventing Coronary Heart Disease in
Primary Care. King's Fund, London.
Kaplan SH, Greenfield S, Ware JE (1989) Assessing the effects of physician-patient
interactions on the outcomes of chronic disease. Med Care. 27 (suppl. 3): SI 10-27.
Kauffman S (1993) The Origins of Order. Oxford University Press, New York.
Kember T, MacPherson G (1994) The NHS: a kaleidoscope of care - conflict ofservice and business
values. Nuffield Hospitals Provincial Trust, London.
Kernick D (2002) Complexity and healthcare organisation. In: Sweeney K, Griffiths F (eds)
Complexity and Healthcare: an introduction. Radcliffe Medical Press, Oxford.
Kernick D, Sweeney K (2001) Post normal medicine (editorial). Fam Tract. 18: 356-8.
King LS (1982) Medical Thinking: a historical preface. Princeton University Press, Princeton, NJ.
Kirkpatrick B (ed.) (1994) Concise English Dictionary. Cassell, London.
Klein R (1989) The Politics of the NHS (2e). Longman, London.
Kleinman A (1988) The Illness Narrative. Basic Books, New York.
Koch R (1891) Ueber Bakteriologische Forschung. In: Verhandlungen des X Intemationalen
Medicinischen Congresses, Berlin, 4-9 August 1890. Hirschwald, Berlin.
Kuhn TS (1970) The Structure of Scientific Revolutions. University of Chicago Press, Chicago.
Landau RL, Gustaffson JM (1984) Death is not the enemy. JAMA. 252: 2458.
Laupacis A (1993) Anticoagulants for atrial fibrillation (editorial). Lancet. 343: 1251.
Laupacis A, Wells G, Richardson SW, Tugwell P (1994) Users' guides to the medical
literature. V. How to use an article about prognosis. JAMA. 272: 234-7.
Le Fanu J (1999) The Rise and Fall of Modern Medicine. Abacus, London.
Lilford RJ, Braunholtz D (1996) The statistical basis of public policy: a paradigm shift is
overdue. BMJ. 313: 603-7.
Lilienfield AM, Lilienfield DE (1979) A century of case-controlled studies: progress? J Chron
Dis. 32:5-13.
Lilienfield DE, Stolley PD (1994) Foundations of Epidemiology (3e). Oxford University Press,
New York.
Liska-Hackzell JJ (1999) Predictions of blood glucose levels in diabetic patients using a
hybrid Al technique. Comput Biomed Res. 32: 132-44.
Lorenz EN (1963) Deterministic non-periodic flow. J Atmosph Sci. 20: 130-41.
Lowe GDO (1993) Antithrombotic treatment and atrial fibrillation (editorial). BMJ. 305:
1445-6.
McColl A, Smith H, White P, Field J (1998) General practitioners' perceptions of the route
to evidence-based medicine: a questionnaire survey. BMJ. 316: 361-5.
McWhinney IR (1983) Changing models: the impact of Kuhn's theory on medical practice.
Fam Pract. 1: 3-8.
Malinowski B (1922) Argonauts of the Western Pacific. Routledge, London.
Marinker M (1970) Balint seminars and vocational training in general practice .JR Coll Gen
Pract. 19: 79.
Marinker M (1981) Clinical method. In: Cormack J, Marinker M, Morrell D (eds) Teaching
Clinical Method. Kluwer Medical, London.
Marinker M (1994) The End of General Practice (Bayliss Lecture). Royal College of Physicians,
London.
Mass Observation (1938) First Year's Work. Lindsay Drummond, London.
Mass Observation (1949) Meet Yourself at the Doctor's. Naldrett Press, London.
Mears R, Sweeney K (2000) A preliminary study of the decision-making process within
general practice. Fam Pract. 17: 428-9.
<r
pH G -iOO
10'179
P°4
A
''/ ■ r <■
lz (
. .. ;
v
1 ''' I
I 52
tv
I
Complexity in primary care
Midgley M (1992) Science as Salvation: a modern myth and its meaning. Routledge, London.
Mill J (1974) A System of Logic: being a connected view of the principles of evidence and methods of
scientific investigation. University of Toronto Press, Toronto.
Mitelton-Kelly E (2003) Ten principles of complexity and enabling infrastructures. In:
Mitelton-Kelly E (ed.) Complex Systems and Evolutionary Perspectives of Organisations:
applications of complexity theory to organisations. Elsevier, London.
Mitelton-Kelly E (2004) The Principles of Complex Evolving Systems. Presentation to the
Fourth Exeter Complexity Conference, September 2004, Exeter.
Mitelton-Kelly E, Papefthimiou B (2000) The Bank Case Study. International Workshop on
Feedback and Evolution in Software and Business Processes (FEAST), July 2000, London.
Morris DB (1998) Illness and Culture in a Postmodern Age. University of California Press,
Berkeley, CA.
Murphy E, Dingwall R, Greatbatch D, Parker S, Watson P (1998) Qualitative Research
Methods in Health Technology Assessment: a review of the literature. Health Technology
Assessment, NHS R&D HTA Programme, London.
NHS Management Executive (1992) First Steps Towards the New NHS. Department of Health,
London.
Neighbour R (1987) The Inner Consultation. MTP Press, Lancaster.
Oxman AD, Sackett DL, Guyatt GH (1993) Users' guides to the medical literature. I. How to
get started. JAMA. 270: 2093-5.
Pederson S (2005) Anti-condescensionism. London Rev Books. 1 September: 7-8.
Peppiatt R (1992) Eliciting patients' views of the cause of their problems: a practical
strategy for GPs. Fam Pract. 9: 295-8.
Petersen P, Boysen G, Godtfredsen J, Andersen ED, Andersen B (1989) Placebo-controlled
randomised trial of warfarin and aspirin for prevention of thromboembolic complications
in chronic atrial fibrillation: the Copenhagen AFASAK study. Lancet. 1: 175-9.
Phillips ED (1973) Greek Medicine. Thames and Hudson, London.
Piaget J (1932) The Language and Thought of the Child. Kegan Paul, London.
Pinchot G, Pinchot E (1996) The Intelligent Organisation: engaging the talent and initiative of
everyone in the workplace. Berrett-Koehler, San Francisco, CA.
Plato (trans. ) (1979) The Sophist. Garland, New York.
Pisek P (2000) Crossing the Quality Chasm: a new health system for the twenty-first century.
National Academy Press, Washington, DC.
Pisek P (2002) Foreword. In: Sweeney K, Griffiths F (eds) Complexity and Healthcare: an
introduction. Radcliffe Medical Press, Oxford.
Pisek P, Wilson T (2001) Complexity, leadership and management in healthcare organ
isations. BMJ. 323: 746-9.
Poincare H (1952) Science and Hypothesis. Dover, New York.
Poincare H (1958) Science and Value. Dover, New York.
Pokmyi M^( 1958) Personal Knowledge: towards a post critical philosophy. University of Chicago
Popper KR (1963) The Open Society and its Enemies. Princeton University Press, Princeton
NJ.
Porter R (1987) A Social History of Madness. Weidenfeld and Nicholson, London.
Porter R (1997) The Greatest Benefit to Mankind. Fontana, London.
Pradilla R (1992) Qualitative and quantitative models ofsocial situations: the case for triangulation
of paradigms. International Symposium on Qualitative Research Process and Computing
Bremen, Germany.
Prein G (1992) Traps of triangulation: what can be done by combining qualitative and quantitative
methods? International Symposium on Qualitative Research Process and Computing
Bremen, Germany.
Pngogine I (1998) The End of Certainty: time chaos and the new laws of nature. The Free Press,
New York.
References and further reading
153
Puustinen M (2000) What is medicine? In search for (sic) a theory of practice. In: Louhala P,
Stenman S (eds) Philosophy Meets Medicine. Helsinki University Press, Helsinki.
Rassam SM (1993) Anticoagulation in patients with atrial fibrillation (letter). BMJ. 307:
1492.
Reiser SJ (1991) Medicine and the Reign of Technology. Cambridge University Press, Cam
bridge.
Robinson R, LeGrand J (1994) Evaluating the NHS Reforms. King's Fund, London.
Rogers EM (1983) Diffusion of Innovations. The Free Press, New York.
Rosenberg W, Donald A (1995) Evidence-based medicine: an approach to clinical problem
solving. BMJ. 310: 1122-6.
Ross D (trans.) (1988, revised) Aristotle: the Nicomachean ethics. Oxford University Press,
Oxford.
Rossler OE, Rossler R (1994) Chaos in physiology. Integr Physiol Behav Sci 29: 328-33.
Rossman GWB (1985) Numbers and words: combining qualitative and quantitative
methods in a single large-scale evaluation study. Evaluation Rev. 9: 627-43.
Royal College of General Practitioners (1972) The Future General Practitioner. Learning and
teaching. Royal College of General Practitioners, London.
Rudebeck CE (1992) Humanism in medicine: benevolence or realism? Scand J Prim Health
Care. 10: 161-2.
Ruse M (L995) In: Honderich T (ed.) (1995) The Oxford Companion to Philosophy. Oxford
University Press, Oxford.
Russell B (1961) The History of Western Philosophy. Unwin Paperbacks, London.
Sackett DL, Cook RJ (1994) Understanding clinical trials (editorial). BMJ. 309: 755-6
Sackett DL, Rosenberg WM (1995) The need for evidence-based medicine. J R SocMed. 88:
620-4.
Sackett DL, Haines RB, Tugwell P (1985) Clinical Epidemiology: a Basic Science for Clinical
Medicine. Little & Brown, Boston, MA.
Sackett DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS (1996) Evidence-based
medicine: what it is and what it isn't. BMJ. 312: 71-2.
Schaffer WM (1985) Can non-linear dynamics elucidate mechanisms in ecology and
epidemiology? IMA J Math Appl Med Biol. 2: 221-52.
Schneider L (1967) The Scottish Moralists on Human Nature and Society. University of Chicago
Press, Chicago.
Scruton R (1982) Kant. Oxford University Press, Oxford.
Second International Study of Infarct Survival Collaborative Group (1988) Randomised
trial of intravenous streptokinase, oral aspirin, both or neither among 17 187 cases of
suspected acute myocardial infarction. Lancet, ii: 349-60.
Silagy C, Jewell D (1994) Review of 39 years of randomised controlled trials in the British
Journal of General Practice. Br J Gen Pract. 44: 359-63.
Singer C, Underwood EA (1962) A Short History of Medicine (2e). Clarendon Press, Oxford.
Skrabanek P, McCormick J (1989) Follies and Fallacies in Medicine. Tarragon Press,
Chippenham.
Smeeth L, Haines A, Ebrahim S (1999) Numbers needed to treat derived from meta
analyses - sometimes informative, usually misleading. BMJ. 318: 1548-51.
Smith A (1776) An Enquiry into the Nature and Causes of the Wealth of Nations (1976).
University of Chicago Press, Chicago.
Smith A (1993) In: Microsoft Encarta Encyclopaedia. Funk and Wengells, Seattle.
Smith BH (1995) Quality cannot always be quantified (letter). BMJ. 311: 258.
Smith J, Heshusius L (1986) Closing down the conversation: the end of the quantitativequalitative debate among educational enquirers. Educ Researcher. 15: 4-12.
Snowden D (2002) Story Telling, Narrative Analysis and Transformational Change in MultiNational Corporations. Presentation at the Second Exeter Complexity Conference, Exeter,
September 2002.
<r
154
Complexity in primary care
Southgate L, Jolly B (1994) Determining the content of re-certification procedures. In:
Newbie D, Wakeford R, Jolly B (eds) The Certification and Re-Certification of Doctors
Cambridge University Press, Cambridge.
StRnai^
Stra^3lC Mana9ement anci Organisational Dynamics: the challenge of complexity.
Stacey R (2001) Complex Responsive Processes in Organisations. Routledge, London.
Stalking G (1995) After Taylor: British socio-anthropology, 1888-1951. University of Wisconsin
Press, Madison, WI.
'
Starfield B (2001) New paradigms for quality in primary care. Br J Gen Pract. 51: 303-9.
Stephenson RC (2004) Using a complexity model of human behaviour to help inter
professional clinical reasoning. Int J Ther Rehabil. 11: 168-75.
Stewart I (1989) Does God Play Dice? Blackwell, Cambridge, MA.
Stroke Prevention in Atrial Fibrillation Investigators (1991) Stroke prevention in atrial
fibrillation study: final results. Circulation. 84: 527-39.
Sweeney BJ (1992) Is there an expertise in human relationship? (MPhil thesis). University of
Glasgow, Glasgow.
7
Sweeney KG (1996) Evidence and uncertainty. In: Marinker M (ed.) Sense and Sensibility in
Healthcare. BMJ Publishing, London.
Sweeney K (2003a) Progressing clinical governance through complexity: from managing
nx c°-creatmg- In: Kermck D (ed.) Complexity and Healthcare Organisations. Radcliffe
Medical Press, Oxford.
Sweeney K (2003b) Final Report to the Nuffield Trust for the Health Complexity Group. Nuffield
Trust, London. Unpublished.
Sweeney KG. Griffiths FE (eds) (2002) Complexity and Healthcare: an introduction. Radcliffe
Medical Press, Oxford.
Sweeney KG, Kernick D (2002) Clinical evaluation: a new model for post normal
medicine. J Eval Clin Pract. 8: 131-8.
Sweeney K Mannion R (2002) Complexity and clinical governance: using the insights to
develop the strategy. Br J Gen Pract. 52 (Suppl.) S4-9.
Sweeney KG, Pereira Gray DJ, Steele RJ, Evans PH (1995) The use of warfarin in nonrheumatic atrial fibrillation: a commentary from general practice. Br J Gen Pract. 45:
153—8.
Sweeney KG. MacAuley D, Pereira Gray DJ (1998) Personal significance: the third
dimension. Lancet. 351: 134-6.
Taylor FW (1911) Principles of Scientific Management. Harper & Brothers, New York
Ten Have HAM (1990) Knowledge and practice in European medicine. The case of
infectious diseases. In: Ten Have HAM, Kimsa GK, Spicker SF (eds) The Growth of Medical
Knowledge. Academic Publishers, Dordrecht.
Tennison B (2002) Complexity and epidemiology in public health. In: Sweeney K, Griffiths
(eds) Complexity and Healthcare: an introduction. Radcliffe Medical Press, Oxford
Times (2002) Leading article. The Times. 27 June: 1.
Toon P (1994) What is Good General Practice? Occasional
Paper No. 65. Royal College of
General Practitioners, London.
^bridge*1"3’ Phil0S°Phy and Government 1572-1651. Cambridge University Press, Cam-
U Acadcrn9/*
S°Cial Anthropo,o^: essays on ,he history of British anthropology. Harwood
Vai032-4VleUten CP' NeWbIC DI <1995) H0W Can We teSl CliniCal reasonin8?
345:
Virchow R (1895) (trans. Rather LJ) (1958) One hundred years of general pathology. In:
Diseases, Life and Man. Stanford University Press, Stanford, CA.
Volberda HW, Lewin AY (2003) Co-evolutionary dynamics within and between firms:
from evolution to co-evolution. J Manag Stud. 40: 2111-36.
References and further reading
I 55
von Bertalanffy L (1968) General Systems Theory. Braziller, New York.
von Foerster H, Zoff GW (eds) (1962) Principles of Self-Organisation. Pergamon, New York.
Weber M (1963) 'Objectivity' in Social Science and Social Policy. Random House, New York.
Wheatley M (2000) Leadership and the New Science. Berrett Koehler, San Francisco, CA.
Whitehead AN (1929) Process and Reality. Macmillan, New York.
Willett C (1999) Knowledge sharing shifts the power paradigm. In: Maybury M, Morey D,
Thuraisingham B (eds) Knowledge Management: classic and contemporary works. Massachu
setts Institute of Technology Press, Cambridge, MA.
Williamson JW, Goldschmidt PG, Jilson IA (1979) Information Demonstration Project: final
report. Policy Research, Baltimore, MD.
World Bank (1997) Confronting AIDS: public priorities in a global epidemic; www.worldbank.
org/aids-econ/confrontfull/summary.html.
World Health Organization (2002) Archives: the World Health Report 2000. World Health
Organization, Geneva.
Wright P, Treacher A (eds) (1982) The Problem of Medical Knowledge. Edinburgh University
Press, Edinburgh.
Wulff HR (1990) Function and value of medical knowledge in modern diseases. In: Ten
Have HAM, Kimsa GK, Spicker SF (eds) The Growth of Medical Knowledge. Academic
Publishers, Dordrecht.
Yarrow D, Fraser S, Tennison B (2003) Planning and scheduling: maintaining flow in
adaptive systems. In: Fraser SW, Conner M, Yarrow D (eds) Thriving in Unpredictable
Times: a reader on agility in health care. Kingsham Press, Chichester.
Zimmerman B, Pisek P (1998) Edgeware: insights from complexity science for healthcare leaders.
VHA, Irving.
Zukav G (1979) The Dancing Wu Li Masters. Bantam, New York.
Further reading
Abramsom JS, Mizrahi T (1996) When social workers and physicians collaborate: positive
and negative interdisciplinary experiences. Soc Work. 41: 270-81.
Adams S, Pill R, Jones A (1997) Medication, chronic illness and identity: the perspective of
people with asthma. Soc Sci Med. 145: 189-201.
Anon. (1992) Systematical overview of controlled trials (meta-analysis) helps us to clarify
treatment effects. Drug Ther Bull. 30.
Apter AJ, Wallack RL, Clive J (1994) Common measures of asthma severity lack associ
ation for describing its clinical course. J Allergy Immunol. 94: 732-7.
Atherton A, Tyrrell A, White PT (1998) Perception and Reported Use of Inhaled Steroids in
Primary Care. Report to Department of Health. King's College, London.
Barry CA, Bradley CP, Britten N, Stevenson FA, Barber N (2000) Patients' unvoiced
agendas in general practice consultations: qualitative study. BMJ. 320: 1246-50.
Barton S (2001) Using clinical evidence. BMJ. 322: 503-4.
Bass MJ, Dempsey B, McWhinney IR (1983) The natural history of headaches in family
practice. In: Proceedings of the Tenth World Conference of the World Organisation of National
Colleges and Academies of General Practitioners, Singapore.
Beate D, Skorpen J, Malterud K (1997) What did the doctor say and what did the patient
do? Operational knowledge in clinical communication. Fam Pract. 4: 376-81.
Benson J, Britten N (2002) Patients' decisions about whether or not to take antihyperten
sive drugs: a qualitative study. BMJ. 325: 873-6.
Benson J, Britten N (2003) Patients' views about taking antihypertensive drugs: a
questionnaire study. BMJ. 326: 1314-15.
Borzak S, Ridker PM (1995) Discordance between meta-analysis and large-scale random
ised controlled trials. Ann Intern Med. 123: 873-7.
A"
I 56
Complexity in primary care
Bradley NCA, Sweeney KG, Waterfield M (1999) The health of their nation: how would
citizens develop England's health strategy? Br J Gen Pract. 49: 801-6.
Bryman A (1984) The debate about qualitative and quantitative methods: a question of
method of epistemology? Br J Social. 35: 75-92.
Budd J, Dawson S (1994) Influencing Clinical Practice: implementation of research and
development results. Report to North Thames Regional Health Authority. Management
School, Imperial College of Science, Technology and Medicine, London
Byles JE, Hanrahan PF, Schofield MJ (1997) It would be good to know you're not alone:
the health care needs of women with menstrual symptoms. Fam Pract. 14: 249-54.
Cassell E (1991) The Nature of Suffering and the Goal of Medicine. Open University Press
Buckingham.
Chalmers I, Hetherington J, Newdick M et al. (1986) The Oxford Database of Perinatal
Trials: developing a register of published reports of controlled trials. Control Clin Trials 1306-24.
Charlton IH, Charlton G, Broomfield J, Mulke MA (1991) Audit of the effect of a nurse-run
227 "il Cl'niC On workload and Patient morbidity in a general practice. Br J Gen Pract. 41:
i
Clarke N (2000) Improving the Performance of Social Services (PhD thesis). University of
Exeter, Exeter.
Corney R (1995) Mental health services. In: Owens P, Carier J, Border J (eds) Interprofessional Issues in Community and Primary Health Care. Macmillan, London.
Dailey B (1991) Beliefs and behaviour: professionals and the policy process. J Ageing Stud.
5: 163—80.
Department of Health/Social Services Inspectorate (1989) Working with Child Sexual Abuseguidance for training social services staff. HMSO, London.
Department of Health/Social Services Inspectorate (1991) Care Management and Assessment
managers' guide. HMSO, London.
Dixon AS (1986) There is a lot of it about: clinical strategies in family practice J R Coll Gen
Pract. 36: 468-71.
Dudbndge SB, Stead J, Ward S, Antony M, Halpin DMG (1996) The Devon primary care
asthma initiative: effects on morbidity. Am J Respir Crit Care Med. 53: A868.
Early Breast Cancer Trialists' Collaborative Group (1992) Systematic treatment of early
breast cancer by hormonal, cytotoxic or immune therapy: 133 randomised trials, involv
ing 31 000 recurrences and 24 000 deaths among 75 000 women. Lancet. 339: 71-85.
Egger M, Davey Smith G (1995) Misleading meta-analysis. BMJ. 310- 752-4
Egger M, Davey Smith G (1998) Meta-analysis: bias in location and selection of studies
BMJ. 316: 61-6.
Egger M, Davey Smith G, Schneider M, Minder CE (1997) Bias in meta-analysis detected
by a simple graphical test. BMJ. 315: 629-34.
Evans M (2001) The medical body as philosophy's arena. Theor Med Bioethics. 1: 17-23
™eta’analysis: statistical alchemy for the twenty-first century. J Clin
Fibrinolytic Therapy Trialists' Group (1994) Indications for fibrinolytic therapy in sus
pected acute myocardial infarction: collaborative overview of early mortality and major
reSU'tS fr°m 311 rand0mised trials of more than 1000 Patients. Lancet. 343:
Fritsche L, Greenhalgh T, Falck-Ytter Y, Neumayer HH, Kunz R (2002) Do short courses in
evidence-based medicine improve knowledge and skills? Validation of Berlin ques
tionnaire before and after study of courses in evidence-based medicine BMJ 3251338-41.
Gambetta D (1998) Trust: making and breaking co-operative relations. Blackwell Oxford
Glaser B (1992) Emergence Versus Forcing Basics of Grounded Theory Analysis. Sociology Press,
1
References and further reading
I 57
Glaser B, Strauss A (1967) The Discovery of Grounded Theory. Aldine, Chicago.
Goffman E (1968) Stigma: notes on the management of spoiled identity. Penguin, Harmondsworth.
Good B (1994) Medicine, Rationality and Experience. Cambridge University Press, Cambridge.
Green J, Britten N (1998) Qualitative research and evidence-based medicine. BMJ. 316:
1230-2.
Greenhalgh T, Hughes J, Humphrey C, Rogers S, Swinglehurst D, Martin P (2002) A
comparative case study of two models of a clinical informaticist service. BMJ. 324: 524—9.
Habermann-Little M (1991) Qualitative research methodologies: an overview. J Neurosci
Nurs. 23: 188-91.
Haynes RB (1993) Some problems in applying clinical evidence. In: Warren KS, Mosteller F
(eds) Doing More Good Than Harm: the evaluation of health care interventions. Annals of New
York Academy of Science, New York.
Haynes RB, Sackett D, Guyatt G, Cook D (1997) Transferring evidence from research to
practice: overcoming barriers to application. Evidence-Based Med. 2: 68-9.
Haynes RB, Devereaux PJ, Guyatt GH (2002) Physicians' and patients' choices in evidence
based medicine (editorial). BMJ. 324: 1350.
Hilton S, Anderson HR, Sibbald B, Freeling P (1986) Controlled evaluation of the effects of
patient-centred education on asthma morbidity in general practice. Lancet, i: 26-9.
Howitt A, Armstrong D (1999) Implementing evidence-based medicine in general practice:
audit and qualitative study of antithrombotic treatment for atrial fibrillation. BMJ. 318:
1324-7.
Hunter D (1993) It's not the knowing, it's the doing. Health Serv J. 7 October: 21.
Hurwitz B (1996) Clinical guidelines and the law (editorial). BMJ. 311: 1517-18.
Jacobson LD, Edwards ACK, Granier SK, Butler CC (1997) Evidence-based medicine and
general practice. Br J Gen Tract. 47: 449-52.
Jaeschke R, Guyatt G, Sackett DL for the Evidence-Based Medicine Working Group (1994)
How to use an article about a diagnostic test. JAMA. 271: 389-91.
Johnson M, Lakoff G (1980) Metaphors We Live By. University of Chicago Press, Chicago.
Jones I, Charlton IH, Middleton M, Preece WJ, Hill AP (1992) Targeting asthma care in
general practice using a morbidity index. BMJ. 304: 1353-6.
Jones K, Bain DJG, Middleton M, Mullee MA (1992) Correlates of asthma morbidity in
primary care. BMJ. 304: 361-5.
Jones R (1995) Why do qualitative research? (editorial). BMJ. 311: 2.
Kellner R (1985) Functional somatic symptoms and hypochondriasis. Arch Gen Psychiatry.
42: 821-33.
Kernick DP (2000) Muddling through in a parallel-track universe (letter). Br J Gen Tract.
50: 65.
Kerridge I, LoweM, Henry D (1998) Ethics and evidence-based medicine. BMJ. 316:1151-3.
Leary D (1994) Metaphors in the History of Psychology. Cambridge University Press, Cam
bridge.
Leedham I, Wistow G (1992) Community Care and General Practitioners. Nuffield Institute,
Leeds.
Lehti A, Mattson B (2001) Health, attitude to care and pattern of attendance among gypsy
women - a general practice perspective. Fam Pract. 18: 445-8.
Lincoln YS, Cuba EG (1985) Naturalistic Enquiry. Sage, Beverley Hills, CA.
Loxeley A (1997) Collaboration in Health and Welfare: working with difference. Jessica Kingsley
Publishers, London.
Mabeck CE, Oleson F (1997) Metaphorically transmitted diseases. How do patients embody
medical explanations? Fam Pract. 14: 271-8.
McCormack J (1996) The death of the personal doctor. Lancet. 348: 667-8.
McWhinney IR (1981) Problem solving and decision making. In: An Introduction to Family
Medicine. Oxford University Press, Oxford.
I 58
Complexity in primary care
McWhinney IR (1991) Primary care research in the next twenty years.
years. In
In Norton
Norton P,
P,
Stewart M, Tudiver F, Bass M, Dunn E (eds) Primary Care Research: traditional and
innovative approaches. Sage, Newbury Park, CA.
Maxwell JA (1992) Understanding and validity in qualitative research. Harvard Educ Rev
62: 279-300.
Mayer J, Piterman L (1999) The attitudes of Australian GPs to evidence-based medicine: a
focus group study. Fam Pract. 16: 627-32.
Melia KM (1996) Rediscovering Glaser. Qual Health Res. 6: 368-78.
Miles HB, Huberman AM (1994) Qualitative Data Analysis: an expanded sourcebook. Sage
Publications, Thousand Oaks, CA.
Mitchell D (2001) Implementing evidence in general practice. BMJ Rapid Responsesbmj.com/cgi/eletters/323/7231 / HOP.
Morgan DL (1988) Focus Groups as Qualitative Research. Sage Publications, London
Morris NV, Abrahamson MJ, Strusser IT (1994) Adequacy of control of asthma in general
practice. Is maximum peak expiratory flow rate a valid index of asthma severity7 Med J
Aust. 160: 68-71.
Muir Gray JA (1999) Postmodern medicine. Lancet. 354: 1550-3.
Murphy E, Mattison D (1992) Qualitative research and family practice: a marriage made in
heaven? Fam Pract. 9: 85-91.
O Connor GT, Weiss ST (1994) Clinical and symptom measures. Am J Respir Crit Care Med
149: S21-8.
Ostergaard MS (1998) Childhood asthma: parents' perspective, a qualitative interview
study. Fam Pract. 15: 153-7.
Oxman A, Thomson MA, Davis DA, Haynes RB (1995) No magic bullets: a systematic
153le?423
trialS °f interVentiOnS t0 imProve professional practice. J Can Med Assoc.
Pereira Gray DJ (1995) Primary care and
health. Health
Health Hygiene.
Hygiene. 16:
and public
public health.
16: 49-62.
49-62.
Perrm K (2001) Better communication helps. BMJ Rapid Responses: bmj.com/cgi/eletters/
■^23/7231/1100.
Potter J (1996) Discourse analysis and constructionist approaches: theoretical background.
In: Richardson JT (ed.) Handbook of Qualitative Research Methods in Psychology and Social
Sciences. BPS Books, Leicester.
Pringle M, Bradley CP, Carmichael CM, Wallace H, Moore A (1995) Significant Event Audit.
Occasional Paper No. 70. Royal College of General Practitioners, London
Rollnick S, Kinnersley P, Stott N (1993) Methods of helping patients with behaviour
change. BMJ. 307: 188-90.
Rudebeck CE (1992) General practice and a dialogue of clinical practice on symptoms
symptom presentations and bodily empathy. Scand J Prim Health Care. Suppl. 1.
Ryle G (1949) The Concept of Mind. Hutchinson, London.
Sackett DL (1992) A primer on the precision and accuracy of the clinical examination
JAMA. 267: 2638-44.
Sackett DL, Richardson WS, Rosenberg W, Haynes RB (1997) Evidence-Based Medicine: how
to practise and teach EBM. Churchill Livingstone, London.
Salinsky J (2001) Psychoanalysis and general practice: what did the Romans do for us7 Br J
Gen Pract. 51: 506.
Schon DA (1995) The Reflective Practitioner: how professionals think in action (3e). Basic Books
King's Lynn.
Scott AJ (1977) Diagnostic accuracy would be improved by developing more categories of
non-disease. Med Hypotheses. 3: 135-7.
Secretary of State for Health (1996) Promoting Clinical Effectiveness. A framework for action in
and through the NHS. NHS Executive, London.
Sheldon B (1994) Biological and social factors in mental disorders: implications for services
Int J Psychiatry. 40: 87-105.
References and further reading
| 59
Silverman HA (ed.) (1980) Piaget, Philosophy and the Human Sciences. Northwestern
University Press, Evanston, IL.
Silverman WA, Altman DG (1996) Patients' preferences and randomised trials. Lancet. 347:
171-4.
Skelton JR, Hobbs EDR (1999) Concordancing: use of language-based research in medical
communication. Lancet. 353: 108-1 1.
Skelton J, Hobbs FDR (1999) Descriptive study of cooperative language in primary care
consultations by male and female doctors. BMJ. 318: 576-9.
Stagnaro S (2001) The complex relations between GPs and evidence. BMJ Rapid Responses;
bmj.com/cgi/eletters/323/7231/1100.
Stern PN (1994) Eroding grounded theory. In: Morse JM (ed.) Critical Issues in Qualitative
Research Methods. Sage Publications, Thousand Oaks, CA.
Strauss A (1978) Negotiations: varieties, contexts, processes and social order. Jossey-Bass, San
Francisco, CA.
Strauss A, Corbin J (1990) The Basics of Qualitative Research-Grounded Theory Procedures and
Techniques. Sage, Newbury Park, CA.
Strauss SE, Sackett DL (1998) Using research findings in clinical practice. BMJ. 317:
339-42.
Sweeney G, Sweeney K, Greco M, Stead J (2002) Softly softly - the way forward. A
qualitative study of the first year of implementing clinical governance in primary care.
Prim Health Care Res Dev. 3: 53-64.
Sweeney K (1996) The use offocus in groups to assess consumers' views on miscarriage and minor
surgery (MPhil thesis). University of Exeter, Exeter.
Teo KK, Yusuf S, Colling R et al. (1991) Effects of intravenous magnesium in suspected
acute myocardial infarction: overview of randomised trials. BMJ. 303: 1499-503.
Thomas KB (1974) Temporarily dependent patients in general practice. BMJ. i: 59-65.
Thompson SG, Pocock SJ (1991) Can meta-analysis be trusted? Lancet. 338: 1127-30.
Tomlin Z, Humphrey C, Rogers S (1999) General practitioners' perceptions of effective
health care. BMJ. 318: 1532-5.
Toulmin S (1993) Knowledge in the practice of medicine: clinical judgment and historical
reconstruction. In: Delkeshamp-Hayes C, Gardell Cutter MAG (eds) Science, Technology
and the Art of Medicine. European-American dialogues. Kluwer, Dordrecht.
Turner-Warwick M (1989) Nocturnal asthma: a study in general practice. J R Coll Gen Tract.
39: 239-43.
Veldhuis M, Wigersma L, Okkes I (1998) Deliberate departures from good general practice:
a study of motives among Dutch general practitioners. Br J Gen Tract. 48: 1833-6.
Wakefield AJ, Murch SH, Anthony A et al. (1998) Ileal lymphoid nodular hyperplasia, non
specific colitis, and pervasive developmental disorder in children. Lancet. 351: 637-41.
Wells S (2001) Knowledge management, a balanced approach. BMJ Rapid Responses;
bmj.com/cgi/eletters/323/7231/1 100.
West RR (1993) A look at the statistical overview (or meta-analysis). J R Coll Physicians
Lond. IT'. 111-15.
Westcott R, Sweeney G, Stead J (2000) Significant event audit in practice: a preliminary
study. Pam Tract. 17: 173-9.
Willis J (1995) The Paradox of Progress. Radcliffe Medical Press, Oxford.
Workman S (2001) It appears evidence-based practice is a skill which must be taught. BMJ
Rapid Responses; bmj.com.
Index
adaptation
complex systems 73-4
organisational development 85-6
adjacent possibles, complex systems 122-3
AIDS strategy, Brazilian, complexity
principles 90-1
ancient Greece
medicine 16
structure/pattern 61-2
anticoagulation, EBM 34
appendix 126-45
applications
commercial world 76-8
complexity 87-91
EBM 28-9
non-linearity 76-9
practical 87-91
asthma
research 107-8
understanding 107-8
atrial fibrillation/warfarin
compliance 10-11
EBM 8-14
implications 11-14
monitoring 10-11
randomised controlled trials 8-9, 11-14
reviewing evidence 8-9
study populations 9-10
Bacon, Francis 17, 46, 47
bank case study, European, complexity
principles 87-8
Berkeley, Bishop George 48-50
von Bertalanffy 63-4
Bichat 18-19
biomedical tradition 15-26
complementarity 83-5
linearity in the NHS 24-5
origins, contemporary medicine 16-18
science in medicine 18-21
Stacey diagram 84
twentieth century medicine 22-4
birds flocking, self-organisation 73
Boerhaave, Hermann 16-18
Brazilian AIDS strategy, complexity
principles 90-1
butterfly, Lorenz 69-70
cardiology, implications 92-3
Cartesian doubt, qualitative research 47-8
causality, qualitative research 50-1
cell differentiation, non-linear tradition
62-3
chaos
characteristics 68-9
understanding 83-96
chaotic systems, cf. complex systems 79-80
characteristics
chaos 68-9
complex systems 71-5
EBM 6
science 5-7
chronic obstructive airways disease, EBM
38
clinical care applications, non-linearity
78-9
clinical medicine, complexity 79
clinical significance, EBM 42
co-evolution
complex systems 73-4
organisational development 85-6
commercial world
applications 76-8
non-linearity 76-8
comorbidity, EBM 33-4
complementarity
implications 85-91
world-views 83-5
complex responsive processes
complex systems 72-3
organisational development 85-6
complex systems
adaptation 73-4
adjacent possibles 122-3
cf. chaotic systems 79-80
characteristics 71-5
co-evolution 73-4
complex responsive processes 72-3
describing 75-6
emergence 74-5
recognising 75-6
self-organisation 73
I!"
162
Index
complexity
applications 87-91
clinical medicine 79
data analysis 97-115
historical development 61-81
implications 85-6
non-linear tradition 61-81
and organisations 121-3
practical applications 87-91
principles 85-6, 87-91, 97-115
and science 120-1
understanding 83-96
Comte, Auguste 51
constant conjunctions, qualitative research
50-1
consultations
EBM 42-4
inner 43
patient-centred 3-4
patient's contribution 3-4, 44
'the questions' 1-2, 117-20
contemporary medicine, origins 16-18
Dark Ages, medicine 16
data analysis
complexity principles 97-11 5
research 97-115
second-level analysis 97-115
decision-making process, research 101-4
dementia, EBM 36
depression, EBM 38
Descartes, Rene 46, 47-8
diabetes, implications 93-4
diagnosis, evidence-based diagnosis 31-2
disease, cf. illness 38
doubt, qualitative research 47-8
duality, qualitative research 47-8
EBM see evidence-based medicine
emergence, complex systems 74-5
empiricism, medicine 16
Engel, GL 23
Enlightenment, medicine 16-18
Epicurus, structure/pattern 62
epidemiology, implications 91-4
epistemological origins, qualitative
research 45-59
European bank case study, complexity
principles 87-8
evidence-based diagnosis, EBM 31-2
evidence-based medicine (EBM) 27-44
anticoagulation 34
applications 28-9
atrial fibrillation/warfarin 8-14
characteristics 6
chronic obstructive airways disease 38
clinical significance 42
comorbidity 33-4
consultations 42-4
criticisms 32-6
dementia 36
depression 38
evidence-based diagnosis 31-2
evidence in 32-5
general practice, nature of 37-40
human relationships 37-40
inner consultations 43
measuring the measurable 35-6
paradigm claim 30
personal significance 40-4
prescribing behaviour 39-40
principles 28
problems 32-4
qualitative research 98-101
questioning 7-8, 11—14
randomised controlled trials 8-9, 11-14,
19, 32-5
research 98-101
reservations 29-30
statistical significance 41-2
steps 27-8
evidence implementation, research 104-6
explanatory models
EBM 27-44
types of knowledge 4-7
understanding 3-14
world-views 4-7
feedback, self-reinforcing, non-linear
tradition 64, 65-7
Feigenbaum, Mitchell 68
Feigenbaum's constant, non-linear
biological systems 68-9
flocking, birds, self-organisation 73
fractal patterning, non-linear biological
systems 69
further reading 155-9
future
prospects 117-23
research 113-15
general practice, nature of 37-40
generalists, general practitioners as 1
Greece, ancient
medicine 16
structure/pattern 61-2
Index
Harvey, William 17
health technology
implications 94-5
research 94-5
healthcare policy
applications, non-linearity 78-9
implications 95-6
Heisenberg's uncertainty principle 57-8,
63
hierarchy, sciences, qualitative research
51-5
historical development
complexity 61-81
medicine 16-26
qualitative research 45-59
human relationships 37-40
Hume, David 50-1
hyperbolic doubt, qualitative research
47-8
I
I
I
I
I
illness, cf. disease 38
immaterialism
qualitative research 48-50
social constructionism 48-50
implications
atrial fibrillation/warfarin 11-14
cardiology 92-3
complementarity 85-91
complexity 85-6
diabetes 93-4
epidemiology 91-4
practical applications 87-91
public health 91-4
research 94-5
theoretical 85-6
induction, qualitative research 48, 54-5
initial conditions sensitivity 69-70, 71-2
inner consultations 43
Kant, Immanuel 51-3
knowledge types 4-7
Koch 20-1
linearity
National Health Service 24-5
science 6-7
Lorenz butterfly, non-linear biological
systems 69-70
Lorenz, Edward 69
Louis, Pierre 19
Ludwig, Karl 20
Malinowski, B 55-6
163
mass observation, qualitative research
55-6
mathematics
non-linear biological systems 67-70
non-linear equations 65-7
non-linear tradition 65-70
self-reinforcing feedback 65-7
milestones, twentieth century medicine 22
Mill, John Stuart 51-2, 54-5
National Health Service, linearity 24-5
naturalism, medicine 16
naturalistic tradition
complementarity 83-5
qualitative research 45-59
Stacey diagram 84
Newton, Isaac 17
nineteenth century, science in medicine
18-21
non-linear biological systems
Feigenbaum's constant 68-9
fractal patterning 69
Lorenz butterfly 69-70
mathematics 67-70
period doubling 68-9
non-linear equations 65-7
non-linear tradition 61-81
cell differentiation 62-3
complementarity 83-5
Heisenberg's uncertainty principle 63
mathematics 65-70
non-linear biological systems 67-70
non-linear equations 65-7
origins 61-2
pattern 61-2
self-organisation 64-5
self-reinforcing feedback 64, 65-7
Stacey diagram 84
structure 61-2
systems theory 63-4
non-linearity
applications 76-9
clinical care applications 78-9
commercial world 76-8
healthcare policy applications 78-9
observation, qualitative research 47
organisational development
adaptation 85-6
co-evolution 85-6
complex responsive processes 85-6
relational dynamics 85-6
self-organisation 85-6
164
Index
organisations, complexity and 121-3
origins
contemporary medicine 16-18
non-linear tradition 61-2
qualitative research, epistemological
origins 45-59
randomised controlled trials 19
social constructionism 48-50
overview, twentieth century medicine
22-4
r
patient-centred consultations 3-4
patient's contribution, consultations 3-4,
44
pattern, non-linear tradition 61-2
Perfection Programme, Pursuing,
complexity principles 88-9
period doubling, non-linear biological
systems 68-9
personal relationships 37
personal significance, EBM 40-4
Poincare, Henri 67
policy, healthcare see healthcare policy
Porter, R 24
positive feedback
non-linear tradition 64, 65-7
self-organisation 64
positivism
medicine 16
qualitative research 51-5
practical applications
complexity 87-91
principles 87-91
prescribing behaviour, EBM 39-40
Prigogine, I 65
principles
complexity 85-6, 87-91, 97-115
EBM 28
Heisenberg's uncertainty principle 57-8,
63
practical applications 87-91
research, complexity principles in 97-115
prospects, future 117-23
Prout, William 19-20
public health, implications 91-4
Pursuing Perfection Programme,
complexity principles 88-9
qualitative-quantitative tensions,
twentieth century 56-8
qualitative research
Cartesian doubt 47-8
causality 50-1
II
constant conjunctions 50-1
doubt 47-8
duality 47-8
EBM 98-101
epistemological origins 45-59
evidence implementation 104-6
hierarchy, sciences 51-5
historical development 45-59
hyperbolic doubt 47-8
immaterialism 48-50
induction 48, 54-5
mass observation 55-6
naturalistic tradition 45-59
observation 47
positivism 51-5
Reformation 47
Renaissance 47
seventeenth century 46-51
twentieth century 55-8
quantitative-qualitative tensions,
twentieth century 56-8
questioning
asking questions 1-2, 117-20
EBM 7-8, 11-14
'the questions' 1-2, 117-20
randomised controlled trials
see also evidence-based medicine
atrial fibrillation/warfarin 8-9, 11-14
evidence in 32-5
origins 19
rationality, medicine 16
reductionism, science 6-7
references 147-55
Reformation, qualitative research 47
relational dynamics, organisational
development 85-6
relationships
human 37-40
personal 37
Renaissance
medicine 16—18
qualitative research 47
research
asthma 107-8
complexity principles 97-115
data analysis 97-115
decision-making process 101-4
EBM 98-101
evidence implementation 104-6
future 113-15
health technology 94-5
implications 94-5
Index
qualitative see qualitative research
second-level analysis 97-115
science
characteristics 5-7
common-sense view 5
complexity and 120-1
defining 5
hegemony of 7
hierarchy 51-5
linearity 6-7
in medicine, nineteenth century 18-21
reductionism 6-7
social 51-3
terminology 5-7
second-level analysis, data analysis 97-115
self-organisation
complex systems 73
non-linear tradition 64-5
organisational development 85-6
positive feedback 64
self-reinforcing feedback, non-linear
tradition 64, 65-7
sensitivity, initial conditions 69-70, 71-2
seventeenth century, qualitative research
46-51
significance
clinical 42
EBM 41-2
personal 40-4
statistical 41-2
social constructionism
immaterialism 48-50
origins 48-50
social science 51-3
sociology, development 53-4
Stacey diagram 77-8, 84
1
165
biomedical tradition 84
naturalistic tradition 84
non-linear tradition 84
Stacey, Ralph 77
statistical significance, EBM 41-2
structure, non-linear tradition 61-2
Sydenham, Thomas 18
systems theory, non-linear tradition 63-4
technology, health, research 94-5
terminology, science 5-7
thinking, like doctors 15-26
trials, randomised controlled see
randomised controlled trials
twentieth century
medicine, eclectic overview 22-4
qualitative-quantitative tensions 56-8
qualitative research 55-8
types of knowledge
explanatory models 4-7
world-views 4-7
uncertainty principle, Heisenberg's 57-8,
63
verificationism, medicine 16
von Bertalanffy 63-4
warfarin/atrial fibrillation see atrial
fibrillation/warfarin
weather systems, Lorenz butterfly 69-70
Weatherall, Sir David 24
Weber, Max 51, 53-4
world-views
complementarity 83-5
explanatory models 4-7
types of knowledge 4-7
- Media
10479.pdf
Position: 448 (13 views)