Content uploaded by Mariana Barosa
Author content
All content in this area was uploaded by Mariana Barosa on Nov 13, 2023
Content may be subject to copyright.
What multimorbidity show us about
guideline-driven evidence-based medicine
MSc History and Philosophy of Science
Department: Science and Technology Studies, University College London
Student number: 22087649
Supervisor: Prof. Phyllis Illari
HPSC0097 September 2023
Word count: 10992
i
Abstract
This dissertation investigates the tension between evidence-based medicine (EBM) in
its fullest sense and guideline-driven EBM. The rise of EBM fuelled the dissemination of
guidelines for almost all individual clinical conditions as practical tools. Over time, a
guideline-driven version of EBM became dominant, where following guidelines became
synonymous with delivering optimal evidence-based care. Notably, this belittled the
epistemic role of clinical expertise in individualising care, a vital part of EBM. Because most
guidelines are designed for single conditions, several challenges arise for physicians dealing
with multimorbid patients, i.e. patients with multiple conditions, inherently complex and
highly heterogeneous. Proposed solutions for these challenges have hitherto focused on
improving guidelines for multimorbidity. In this dissertation, I argue that guideline-focused
approaches to multimorbidity are bound to be a fiasco because they are premised on
guideline-driven EBM, an epistemologically limited interpretation of EBM. Finally, I argue
that clinical expertise becomes even more important in multimorbidity care because only
clinical expertise can effectively capture the sheer idiosyncratic nature of multimorbid
patients. To my knowledge, this dissertation is the first formal exposition of the limitations of
guideline-focused approaches to multimorbidity. Understanding the limitations of guideline-
driven EBM in the context of multimorbidity has important implications for medical
education, primary care practice, and broader healthcare strategies. This emphasises the need
for an urgent revival of EBM in its fullest sense.
ii
Where is the wisdom we have lost in knowledge?
T.S. Eliot “The Rock” 1934
iii
Acknowledgements
I would like to give special thanks to my supervisor Prof. Phyllis Illari, for her
unwavering support, invaluable guidance and knowledge, which have greatly enriched this
dissertation.
I also wish to express my deep gratitude to my parents. This dissertation would not
have been possible without their continuous support.
I would also like to thank Pedro not only for his meticulous review of this dissertation
and insightful comments as a physician, but also, and most importantly, for accompanying
me in this journey and in my pursuit of truth.
I am also grateful to Prof. José Alves, the director of the department where I was
training in Internal Medicine, an internationally distinguished medical department, for
wholeheartedly supporting my academic pursuit. Also, for sparking my interest in philosophy
of science since medical school by constantly promoting critical thinking in medicine.
I would like to express my sincere gratitude to Dr. McCormack, a distinguished
Canadian primary care physician and researcher, for generously dedicating his time to meet
with me via online to discuss, early on, some of the topics of this dissertation. His insights
were truly enlightening. Moreover, his work has been a longstanding source of inspiration in
my medical practice.
iv
Table of Contents
Introduction ................................................................................................................................ 1
1. A guideline-driven version of EBM is dominant ............................................................... 3
1.1. What is EBM? ................................................................................................................. 3
1.2. Guideline-driven EBM is dominant and belittles clinical expertise ............................... 4
1.3. Clinical expertise: a vital part of the EBM story ............................................................ 8
2. A guideline-driven EBM amplifies the multimorbidity challenge .................................. 10
2.1. Multimorbidity (in a nutshell) ....................................................................................... 10
2.2. The challenges of multimorbidity to a guideline-driven EBM ..................................... 11
2.3. Decision-making in multimorbidity through the lens of the EBM-Ex model .............. 15
3. Putative solutions for the multimorbidity challenge are limited guideline-focused
approaches ................................................................................................................................ 19
3.1. Approach A: better crafting single-condition guidelines for multimorbid patients ...... 19
3.1.1. The inability to account for the sheer scale of individual variation ........... 19
3.1.2. The increasingly impractical size of guidelines .......................................... 20
3.1.3. Two points of conflict with EBM ................................................................. 21
3.2. Approach B: developing specific guidelines for multimorbidity ................................. 22
3.2.1. How useful is generic guidance? ................................................................ 23
3.2.2. Two same two points of conflict with EBM ................................................. 23
3.3. The root of the problem: guideline-driven EBM .......................................................... 23
4. More clinical expertise is needed for multimorbidity care .............................................. 26
4.1. What is clinical expertise? ............................................................................................ 26
4.2. The augmented role of clinical expertise in multimorbidity ......................................... 27
4.2.1. The prioritisation of health problems ......................................................... 28
4.2.2. The translation of research evidence to the patient of interest ................... 30
v
4.2.3. Final considerations ................................................................................... 33
Conclusion ............................................................................................................................... 34
References ................................................................................................................................ 36
1
Introduction
Primary care physicians are increasingly voicing concerns against a guideline-driven
paradigm of clinical practice, specifically a push for evidence-based guidelines. At the root of
these concerns lies a tension between what these physicians think is the ideal way of
practicing medicine and how that ideal has been implemented: a tension between evidence-
based medicine (EBM) in its fullest sense and guideline-driven EBM, which I will lay out
this dissertation. While EBM is typically conceived as a model for clinical decision-making
for individual patients, guideline-driven EBM focuses on developing population-level
recommendations meant to represent the best decisions for care. It is worth noting from the
outset that the former concedes a vital role to clinical expertise, but the latter does not. I will
address these issues in Chapter 1.
Notably, this tension between EBM and guideline-driven EBM gains full expression in
primary care, as opposed to specialist care, because primary care practice is dominated by
multimorbidity. Up to 80% of general practitioners’ consultations, the bedrock of the English
primary care system, involve the management of multimorbid patients (Salisbury et al.,
2011). In short, patients with multimorbidity have several concomitant conditions (e.g. old
age, diabetes1, kidney disease, depression, and high risk of falls) and are much more complex
than the guideline-driven EBM acknowledges, as I will explain in Chapter 2.
This conflict between multimorbidity care and guideline-driven EBM is a topic of
timely relevance for several reasons. Firstly, multimorbidity is on the rise and is associated
with multiple negative consequences, such as reduced quality of life, higher mortality, and
significant treatment burden (Skou et al., 2022). Secondly, multimorbidity care concerns a
substantial number of physicians in a crucial sector of the healthcare system. Most care for
multimorbid patients takes place in primary care (ibid.), and primary care is widely
recognised as a key component of all high-performing health systems as a model of care that
supports comprehensive and coordinated person-focused care (Hanson et al., 2022). Thirdly,
the practical and philosophical challenges posed by multimorbidity to EBM – which
collectively constitute what I call the multimorbidity challenge – are the subject of active
debate (Skou et al., 2022, Fuller, 2016).
1 A state of severely dysregulated blood sugar levels.
2
Thus far, proposed solutions to the multimorbidity challenge focus on creating better
guidelines for multimorbid patients. So, in this dissertation I wish to address the following
research questions:
(i) What is the role of evidence-based guidelines for multimorbid patients in
primary care practice?
(ii) What is the role of clinical expertise in multimorbidity care?
For my purposes, I will address both research questions through the lens of the most
popular version of the EBM model for clinical decision-making. I will introduce this model
in the first chapter. The reason for adopting this approach is that this particular model
captures key aspects of the multimorbidity challenge that will allow me to lay out the tension
between EBM and guideline-driven EBM.
In this dissertation, I contend that guideline-focused approaches to multimorbidity are
ultimately limited because they are premised on a guideline-driven interpretation of EBM,
which is itself an epistemologically limited version of EBM. As I will argue, guideline-driven
EBM is limited because it belittles clinical expertise. Furthermore, I will argue clinical
expertise plays an augmented role in multimorbidity care. I adopt an analytical approach that
appeals primarily to epistemological and practical insights from philosophy of medicine, but
which also draws on my personal experience as a physician caring for multimorbid patients.
While this work is not nearly a comprehensive treatment of the multimorbidity challenge, I
hope to provide a modest, yet novel, contribution to the wider multimorbidity debate.
This dissertation is divided into four chapters that follow my line of argument. In
Chapter 1, I introduce EBM and argue that a guideline-driven version of EBM dominates and
belittles clinical expertise. In Chapter 2, I explain the multimorbidity challenge and argue that
a guideline-driven EBM amplifies it. In Chapter 3, I discuss the limitations of two distinct
guideline-focused approaches addressing multimorbidity and argue that neither looks
promising because both stem from a guideline-driven EBM. In Chapter 4, I close the
argument explaining why clinical expertise plays an augmented role in multimorbidity care.
3
1. A guideline-driven version of EBM is dominant
1.1. What is EBM?
EBM was born in the early nineties as a new paradigm for clinical decision-making
(Guyatt et al., 1992). Rapidly it became one of the most influential frameworks for clinical
practice (Djulbegovic et al., 2009). EBM emerged to deemphasise the authority of clinical
“experts” and stress the importance of examining evidence from clinical research in a
systematic way. Despite its success, establishing what EBM really is has been challenging
(ibid.). For the purposes of this dissertation, I focus on the conceptualisation of EBM as a
decision-making model because it provides a convenient framework for examining the
challenge of multimorbidity.
The most cited definition states that EBM is “the conscientious, explicit, and judicious
use of current best evidence in making decisions about the care of individual patients” and
that “[t]he practice of evidence-based medicine means integrating individual clinical
expertise with the best available external clinical evidence from systematic research” (Sackett
et al., 1996, p.71; emphasis added). Clinical expertise was there defined as “the proficiency
and judgment that individual clinicians acquire through clinical experience and clinical
practice” (ibid., p.71). In the same paper, the authors concede that clinical expertise is
important for practising EBM, but do not allow it any evidential role in establishing that an
intervention works.
EBM has undergone a series of transformations, but its central message – that
comparative clinical studies, in general, provide better evidence than mechanistic reasoning
and expert judgement – has been left untouched (Howick, 2011, p.6). Although not the focus
here, another key message behind this hierarchical view of evidence is that well-controlled
experimental studies, particularly randomised controlled trials (RCTs), tend to be
epistemically superior to observational studies with respect to medical interventions. It is
worth acknowledging that this is one of the most contentious epistemological issues of EBM.
However, the premise is that evidence is good insofar as it provides knowledge of causation,
hence supporting a causal relation between an intervention and a particular outcome (Bird,
2011a). This is often just called “efficacy”. Because experimental studies are typically more
reliable than observational ones in eliminating confounding factors to the target hypothesis,
they tend to generate evidence of higher quality for efficacy claims (ibid.). Widely
4
established therapies, initially supported by “lower” forms of evidence, were often found
ineffective or harmful when tested in RCTs (Howick, 2011).
In their seminal paper, Sackett et al. (1996) offered an early version of the EBM model
for clinical decision-making. Six years later, Haynes et al. (2002) proposed a reformulation of
the model to expand the nature and scope of clinical expertise. According to this new
account, individual clinical decisions should be based on three components – “patient’s
clinical state and circumstances”, “patient’s preferences and actions” and “research evidence”
– and clinical expertise plays a central role in their integrating these components. For the
sake of clarity, I will call it “EBM-Ex model”.
Even if underappreciated, the EBM-Ex model is widely endorsed as representing EBM
in the fullest sense. Notably, this recognition includes the most extensive philosophical
treatment of EBM to date (Howick, 2011). As I noted in the Introduction, I will use this
particular model to build my argument in this dissertation because it captures key aspects of
the multimorbidity challenge. The effectiveness of my argument will also rely on the visual
representation of this model, which I shall present in the next section.
1.2. Guideline-driven EBM is dominant and belittles clinical expertise
While there are different interpretations of EBM, I will now demonstrate that a limited
guideline-driven version of EBM has become dominant. As I will explain, this guideline-
driven EBM results from a narrow interpretation of EBM because it neglects vital
components of the EBM-Ex model, most notably clinical expertise. This is important
because, as I will argue in the next chapter, a guideline-driven EBM amplifies the
multimorbidity challenge.
The EBM movement, since its early days, has focused on the importance of using the
best available research evidence in every individual clinical decision. As it became clear that
individual physicians lacked the time and skills to critically assess evidence themselves
(Guyatt et al., 2000), and that the volume of evidence generated was becoming unmanageable
(Greenhalgh et al., 2014), EBM advocates started to encourage the use of tools. In this
regard, evidence-based clinical guidelines, hereafter abbreviated to “guidelines”, attained
widespread popularity (Upshur, 2005).
5
In theory, guidelines are recommendations developed by expert panels and
organisations, based on the best available evidence, typically involving a systematic review
of research literature, with the intention of assisting decision-making about appropriate care
for specific clinical situations (Hayward et al., 1995, Guyatt et al., 2008, Muth et al., 2019;
emphasis added). On a broader level, the motivations behind guidelines are twofold: to
reduce variation in practice (Eddy, 2005) and improve the quality of care (Garber, 2005). In
practice, these motivations led to a large-scale push for guidelines as rules to be followed
rather than as mere tools (Copeland, 2020), giving rise to what I call “guideline-driven
EBM”. Numerous organisations, professional societies and regulatory agencies, such as the
Agency for Healthcare Research and Quality in the United States and the National Institute
for Health and Care Excellence in the United Kingdom, have increasingly embraced
guideline-driven EBM.
Now, whether guidelines actually improve the quality of care remains uncertain. In
principle, the claim that guidelines can improve quality of care hinges on the assumption that
guidelines can “handle” the complexity of medicine and be effective knowledge translation
tools, i.e. effectively translate, or individualise, research knowledge to the patient-at-hand.
Guidelines curate vast amounts of research evidence and simplify medical decisions. True,
busy primary care physicians do find some value and utility in guidelines (Farquhar et al.,
2002). Yet, what are the epistemic implications of this simplification? The following
considerations partially explore this issue.
Silva and Wyer (2009) have pointed out that EBM has two distinct “dimensions”, at
least from the perspective of how EBM is actually practised. This means that, in practice,
EBM admits two distinct approaches to using evidence to solve clinical problems: evidence-
based guidelines and evidence-based individualised care. Despite descriptively accurate, this
bidimensional view of EBM is not how the early EBM proponents idealised EBM. Only the
latter dimension really corresponds to EBM, as I will explain next. Nevertheless, this
bidimensional view sheds light into the limitations of guideline-driven EBM and will help the
reader later understand why I argue that multimorbidity is not being properly addressed.
The evidence-based guidelines dimension stems from the appraisal of research
evidence on a particular topic. Our best research methodologies produce population-level
estimates that allow us to determine average effects of medical interventions. Such
information is then used to produce generic guidelines catered to the needs of groups of
6
people, which affect individual patients only indirectly by influencing the decisions of
physicians (Eddy, 2005). Of note, most guidelines focus on the management of single
conditions. This is because studies are usually designed to isolate the effect of a single
exposure (typically an intervention) on a single condition outcome (Boyd and Kent, 2014).
We will return to this issue in Chapter 2.
The other dimension, evidence-based individualised care, is actually what EBM is all
about. Let me present the diagram of the EBM-Ex model to illustrate this – see Figure 1. I
shall note beforehand that I made a small change to the original diagram published by Haynes
et al. (2002) for the purposes of my argument, as I will immediately explain; because of that,
Figure 1 features both the standard version by Haynes et al.’s and my own. Still, both
versions display three intersecting circles (representing the three components of decision-
making) with “clinical expertise” overlaid.
Now, in the standard version clinical expertise is depicted as a horizontal ellipse,
missing an area of overlap between the two bottom circles (there is no mention in the original
publication of this being deliberate). In my version, clinical expertise becomes more circular
because I believe it must cover all the overlaps to truly represent its all-encompassing
integrative role in decision-making. Hence, I will use this diagram hereafter, still calling it
Figure 1. The EBM-Ex diagram.
7
EBM-Ex. This will be an important visual aspect in Chapter 4 when I argue that the role of
clinical expertise is augmented in multimorbidity care. Yet, my point here is that evidence-
based individualised care is about integrating research evidence with individual patient
variables.
In light of the above, it appears that guideline-driven EBM results from the reduction
of EBM to the guidelines dimension. This is problematic because this is an epistemologically
limited interpretation of EBM. To explain why I think this is the case, let me first introduce a
well-known philosophical problem of evidence-based individualised care: the knowledge-to-
practice gap. While EBM is envisioned as a way “to close the gulf between good clinical
research and clinical practice” (Rosenberg and Donald, 1995, p.1112) in order to provide
individualised care, this metaphoric gulf is actually an intrinsic epistemological gap (Tonelli,
1998). This is the knowledge-to-practice gap. To be more precise, the gap is twofold (Bird,
2011a). The first gap is a difficulty in providing an account of inferences from statistical data
to predictions concerning an individual. If a research study shows that an intervention
reduces death by 50%, the average results of this trial do not tell us what will happen to an
individual patient. The second concerns the integration of patient circumstances and values in
decision-making.
The crux of the matter is that EBM claims to bridge the knowledge-to-practice gap
with clinical expertise (Tonelli, 2006). So, evidence-based individualised care is possible
because of clinical expertise. However, the guideline-driven version of EBM focuses only on
research evidence and neglects all the other components of the EBM-Ex model, most notably
clinical expertise. Thus, a guideline-driven EBM cannot overcome the knowledge-to-practice
gap. This aspect of guideline-driven EBM limits its ability to provide individualised care and
becomes particularly problematic in multimorbidity, as I will demonstrate in the following
chapters. There is one caveat: in the absence of research evidence for a given clinical
situation, guideline-driven EBM recognise the need for clinical expertise but takes it from
physicians and places it in guideline panels, as I will show in Chapter 3. Despite this, I will
maintain that guideline-driven EBM belittles clinical expertise by not recognising it at the
level of the practitioner where it belongs.
8
1.3. Clinical expertise: a vital part of the EBM story
Let me now briefly explain why clinical expertise has been given an important
epistemic role in EBM. Not only will this explanation support the argument of the previous
section, but it will also lay the groundwork for upcoming chapters, particularly Chapter 4.
There I will consider clinical expertise in greater depth and argue that it has an augmented
role in multimorbidity.
While EBM proponents acknowledge the need to integrate different sorts of
knowledge, they say very little about how such integration should take place. But they do say
that clinical expertise is responsible for that integration. What sort of thing is clinical
expertise to be able to overcome an epistemological gap? Although the position of EBM on
clinical expertise is somewhat ambiguous, EBM supporters tend to situate it in the realm of
the physician’s knowledge-how (Howick, 2011, Ch.11), commonly known as “know-how”.
To clarify, knowledge-how is knowledge of how to perform a particular action or task. This
kind of knowledge is typically distinguished from knowledge-that, which is knowledge of
facts. There are various distinguishable kinds of knowledge, but I do not wish to go into such
complex epistemological issues in this dissertation.
However, the distinction between knowledge-how and knowledge-that matters for us
because it contributes to our understanding of EBM and clinical expertise. In EBM only
knowledge-that is typically given an evidentiary role. For instance, knowledge generated by
high-quality clinical research is the cornerstone of knowledge that interventions work. But a
physician can also know that a certain patient has financial problems, or a preference for
treatment A instead of B. This is knowledge about individual patient circumstances and
preferences, respectively, both of which are circles in the EBM-Ex model (see Figure 1).
Clinical expertise is a different sort of knowledge, one that is often associated with elements
of judgment. In Chapter 4, I will use the case of multimorbidity to explain in more detail
what clinical expertise is and how it operates.
Indeed, it is not without reason that clinical expertise has been recognised as an
essential element of the art of medicine, integrating the understanding of particulars with the
understanding of universals, and helping bridge the gap between theory and practice in
medicine (Malterud, 1995). To practise EBM is to meet an individual person with unique
9
“preferences and actions” and “clinical state and circumstances” and build something out of
that clinical encounter in order to arrive at the right clinical decision for that person.
Clinical expertise has always been perceived as a vital part of the EBM story. The
EBM-Ex model recognises tacit elements of judgement are necessary for the care of
individual patients (Haynes et al., 2002). Medicine is, for many, not only about the care of
persons, but also contingent on human relationships (Malterud, 1995). That is, knowing how
to integrate an individual person’s needs and wishes with research evidence hinges on tacit
elements that underpin basic human experiences and judgements.
Moreover, the role of clinical expertise in handling multiple and complex sources of
information has only been trivially acknowledged in the literature (Bird, 2011b). For
example, the management of patients with multiple conditions requires sorting out trade-offs
and prioritising problems, as I explain in the next chapter. Furthermore, the conditions under
which EBM is practiced vary greatly between locations, even when those locations are in
proximity: resources and patients often differ, so the same research evidence cannot be
applied exactly the same way. To illustrate, if research shows that physiotherapy helps with a
particular issue but there are no nearby clinics or no one to take the patient there, then the
right course of action is to pursue an alternative approach. This is where clinical expertise
plays a part.
However, as I argued earlier in this chapter, the trend has been to push for guidelines
and belittle clinical expertise. Indeed, EBM proponents have been accused of transforming
the practice of medicine into “cookbook” medicine, where guidelines are akin to recipes that
practitioners should follow to the letter (Martini, 2021). While I do think that guidelines have
a role in assisting decision-making in relatively simple cases, their usefulness falls short with
increasing complexity. I develop this point throughout the following chapters, arguing that
this is partly the reason why a guideline-driven EBM is bound to be a fiasco for
multimorbidity.
10
2. A guideline-driven EBM amplifies the multimorbidity challenge
2.1. Multimorbidity (in a nutshell)
Multimorbidity is usually defined as “the co-occurrence of at least two chronic
conditions in the same individual” (Skou et al., 2022). However, this definition is simplistic
and some have suggested more sophisticated definitions (ibid.). For my purposes here, I
define multimorbidity as the coexistence of two or more chronic conditions, where (i) one
condition is not necessarily more central than the others (Boyd and Fortin, 2010), and (ii)
each condition may influence optimal clinical management of other condition(s) (Uhlig et al.,
2014).
Conditions are generic terms that include recognised diseases (e.g. diabetes) and
conditions deviate from the traditional disease model (e.g. risk of falls, frailty2, etc.) (Boyd
and Fortin, 2010). It is worth noting that there is intense debate around the concept of disease,
but it is ultimately tangential to my argument. In short, the traditional disease model is
usually seen as a model that conceptualises disease as a perturbation from normal biological
functioning resulting from biological causes (Fuller, 2016). For example, while diabetes is
generally explained in purely biological terms, a “high risk of falls” often results from the
interaction of different kinds of conditions, such as previous stroke, vision problems and
insufficient vigilance from caregivers.
Multimorbidity is associated with the idea of complexity. Without going further into
this difficult concept, the point is that multimorbid patients are more complex than non-
multimorbid ones. To illustrate this, Figure 2 presents a conceptual diagram of
multimorbidity within a particular individual patient context. Grey ellipses represent
conditions that correspond to recognised disease states. Coloured ones, though, represent
conditions that often arise and become relevant in multimorbid patients but not in non-
multimorbid ones (Skou et al., 2022). Furthermore, all conditions typically interact, as
indicated by the intersections between the ellipses.
2 A medical term for a state of increased vulnerability typical of older adults, that is associated with several
adverse health outcomes and has complex multifactorial aetiologies.
11
2.2. The challenges of multimorbidity to a guideline-driven EBM
The challenges posed by multimorbidity to the practice of EBM, particularly in
primary care, are being increasingly addressed in the literature (Fortin et al., 2007, Ross and
Shawn, 2008, Skou et al., 2022, Fuller, 2016). Here, though, I will focus on showing that
multimorbidity is particularly challenging to a guideline-driven EBM. For this purpose, I
consider three main epistemological challenges and one major practical issue. Most are
familiar in the wider EBM literature but gain full expression when applied to multimorbidity.
The first epistemological challenge is defining and measuring multimorbidity (Weiss
et al., 2014). There is currently no standard way of doing it, let alone doing it comparably
across studies. Multimorbidity measures may rely on the presence of diseases alone, or also
include – together with a severity assessment – other non-disease conditions (frailty,
loneliness, etc.), measures of function (hearing, vision, gait, etc.) and treatment burden
Figure 2. Multimorbidity within a particular individual patient’s context.
12
indicators . As Weiss (2014) notes, all of these are likely to be relevant for a given research
question and may be desired as part of a multimorbidity measure. For instance, the simple
“two or more chronic conditions” definition has received criticism for (i) including people
with combinations of conditions that do not significantly affect the individual nor the
management of each condition (e.g. well-controlled hypertension, tobacco use and mild
asthma), and (ii) not assessing condition burden. One may argue that overcoming these
limitations is merely a matter of doing more research, breaking down multimorbidity into
recognised condition clusters, and developing better diagnostic and measuring tools. But
guidelines wishing to account for multimorbidity will find it hard to do it because
multimorbidity is not a disease or condition in the usual sense and affects every person
differently3. Multimorbidity requires an approach to care that is more than simply optimising
care for every individual condition present and which must also take into consideration each
specific condition. The core of its diagnosis comprises the physician actively engaging with
the patient (and/or carer) to understand the patient’s experience (Skou et al., 2022, p.9). Thus,
the definition of multimorbidity seems bound to be a generic one (as the one I gave in section
2.1).
Another epistemological problem is that guidelines are built on an evidence base
mostly concerned with single conditions. In other words, evidence is largely based on trials
of interventions for single conditions, from which multimorbid patients are generally
excluded (Skou et al., 2022). Thus, the research evidence base for multimorbid patients is
weak. This challenge is well captured in the vexing trade-off between internal and external
validity, which has been well described by philosophers of science (Blunt, 2015, Cartwright,
2007). Suffice it to say for my purposes here that (1) internal validity ensures the accuracy
and reliability of findings within a specific study population, and that the more “sanitised” the
trial population – i.e. carefully selected to eliminate contaminating influences – the greater
the internal validity of the trial; (2) external validity concerns the extent to which findings can
be applied to broader populations and real-world settings that include the patient of interest.
Generally, studies with high internal validity have low external validity, and vice versa.
RCTs tend to have high internal validity – hence being assigned a high level of evidence –
precisely because they normally focus on sanitised populations, yet have low external
3 I am not here attempting a metaphysical claim about multimorbidity (that goes well beyond the scope of
this work); for one of the few discussions in the literature about that see (Fuller, 2016). What I mean is
that, in practice, multimorbidity has implications different from the typical single-condition state (be they
diseases or not).
13
validity. In the case of multimorbidity, the problem of external validity is amplified.
Although single-condition guidelines are the norm, we live in a world where single problems
and single therapies are increasingly rare. For example, there are dozens of heart failure
guidelines with hundreds of recommendations. But when we delve into the evidence, we
realise the typical primary care patient is much more complex than the type of patient
included in the evidence base (typically patients from Cardiology specialty departments).
Primary care patients are older and typically multimorbid (Skou et al., 2022). Extrapolating
recommendations from these guidelines into primary care practice is epistemically
problematic (Fuller, 2021).
This generates two problems: uncertainty of benefits and higher likelihood of harms
(Tinetti et al., 2012). The first is not only related to the limited generalisability of research
findings, but also to the fact that medicine is increasingly a science of marginal benefits
(Greenhalgh et al., 2014). Further, what additional benefit does the eighth or ninth medication
offer over the second or third? The long-term benefits associated with the combination of all
treatments that are taken in adherence to single-condition guidelines for patients with
multimorbidity is unclear. The second problem is multifaceted: (i) harms are systematically
underestimated in clinical research (Stegenga, 2016), which can be exacerbated in
multimorbid patients given their complexity; (ii) there is a higher risk of condition-condition,
condition-drug and drug-drug interactions (e.g. worsening one condition by treating a co-
existing one according to single-condition guidelines) (Boyd et al., 2005); (iii) the mention of
harm in guidelines is limited to a few well-known effects of individual medications (e.g.
renal abnormalities) (Tinetti et al., 2004); and (iv) treatment burden arising from adhering to
multiple condition-specific guidelines is a potential harm in itself and is rarely accounted for.
The last epistemological challenge concerns the prioritisation of health problems,
which is essential for the management of multimorbidity. Although multimorbidity is
generally viewed as requiring a kind of holistic or “patient-centred” care (care that integrates
context, conceives multimorbidity as something more than a collection of physiological
abnormalities and addresses the broader understanding of the patient’s state), it is also the
case that each individual condition needs a specific management (Skou et al., 2022).
However, conditions are hardly ever addressed all at the same time because: (i) at any given
moment, patients and physicians focus on the most pressing problems that the patient is
experiencing (ibid.); (ii) the most pressing conditions – either because they are the patient’s
14
priority, the source of symptoms or the primary determinants of prognosis – vary over time
(Shippee et al., 2012); and (iii) following condition-specific guidelines for all conditions at
once often results in conflicting management plans and massive treatment burden (Boyd et al.
(2005) deftly show this by applying guidelines to a hypothetical 79-year-old multimorbid
patient).
For instance, the decision to initiate a blood thinner in a young, otherwise healthy
patient who had a heart attack in order to prevent another event would follow seamlessly
from guidelines. But if the patient was older (though still active) and had concomitant
anaemia4 caused by chronic blood losses, kidney disease and gait imbalance, not only would
the decision to start aspirin be much less clear, due to a higher risk of bleeding complications,
but the anaemia and risk of falls would likely become priorities in the follow-up and
management of this patient. The point here is that it is highly unlikely that any guideline will
address all possible individual variables in enough detail to effectively help setting priorities.
I say this because guidelines are averaged out recommendations, each multimorbid patient is
a unique set of conditions, and the prioritisation of problems directly depends on the
circumstances and preferences of that individual patient.
Let us now consider a major practical issue faced by primary care physicians: the
unmanageable volume of condition-specific guidelines (Greenhalgh et al., 2014). Although
guidelines, in general, have issues – such as not being nearly as “evidence-based” as they
purport to be (Primiano et al., 2016), making implicit value judgements (Guthrie and Boyd,
2018) and frequently having conflicts of interest (Neuman et al., 2011) – my concern here is
their unrestrained growth in size and number. For illustrative purposes, the recent growing
size of three influential guidelines for diabetes is presented in Table 1. The Canadian and
American ones are now over 200 pages – who would ever consider that useful for routine
decision-making?
4 A condition occurring when the body lacks enough healthy red blood cells to transport oxygen
effectively.
15
In 2005, the average primary care physician would need 18h/day to follow
recommendations for chronic disease and preventive care (Østbye et al., 2005). The last two
decades have witnessed a dramatic increase in the number of guidelines (Upshur, 2014) and
the number of recommendations within guidelines continues to grow (Tricoci et al., 2009).
This seriously undermines guidelines’ ability to translate research evidence into clinical
practice.
2.3. Decision-making in multimorbidity through the lens of the EBM-Ex model
This section builds on preceding ones and aims to reveal, through the lens of the
EBM-Ex model, another relevant aspect of the multimorbidity challenge: the prominence of
patient preferences and patient circumstances in decision-making. While the standard version
of the EBM-Ex model with three equally-sized intersecting circles fares relatively well for
most simple clinical situations, this is arguably not the case in multimorbidity, as I explain
next. The specific implications for clinical expertise will be addressed in Chapter 4. Terms
Table 1. Selected examples of influential diabetes guidelines.
Most recent versions are highlighted in bold.
16
such as “relatively”, “more”, “simple” or “complex” are here used with the sole purpose of
establishing gross relative comparisons.
Earlier in this chapter I showed that a patient has multimorbidity when: (i) one
(chronic) condition is not necessarily more central than the others, and (ii) each condition
may influence the optimal management of the others. However, to fully grasp the complexity
of the multimorbidity challenge, it is helpful to get a deeper sense of multimorbidity. That is,
that multimorbidity is strongly associated with particular things that significantly influence
the way multimorbidity is (and should be) approached, both conceptually and practically. I do
not intend to provide a comprehensive list of all such things, but rather indicate those that
seem most relevant for my analysis.
Thus, in addition to (i) and (ii), multimorbid patients tend to: (iii) have numerous
potential combinations of conditions (i.e. multimorbidity affects every patient differently),
(iv) have advanced age (conditions accrue as a person ages); (v) have less capacity for self-
management and increased social needs; (vi) experience polypharmacy (the intake of
multiple medications), which increases the likelihood of adverse events, (vii) have significant
treatment burden due to complex drug regimens and follow-up plans, (viii) be frail and thus
more susceptible to harm from medical interventions, and (ix) not be enrolled in clinical trials
(for the epistemological reasons given in section 2.2, but also because, practically speaking,
many of these patients are house-bound or have difficulty in accessing research units), thus
leading to uncertainty regarding benefits and harms of interventions (Skou et al., 2022).
The multimorbidity features (i)-(viii) explain why the individual clinical state and
circumstances and the patient’s preferences become more salient when managing
multimorbid patients than non-multimorbid ones. Each case is a unique complex set of
conditions. Moreover, old, multimorbid patients generally have a lower life expectancy.
Focusing on improving quality of life according to what the patient values (minimising
symptoms, reducing disability, etc.) is recognised as a primary goal of care in these patients
(Grumbach, 2003). Setting priorities and making trade-offs according to an individual
patient’s preferences are a major part of the management of multimorbidity.
17
In Figure 3 I appeal to the EBM-Ex diagram to illustrate the difference between
decision-making in typical single-condition and multimorbidity situations. In single-condition
situations, the three circles are well balanced in size. In multimorbidity this is not the case.
First, as the point (ix) above explains, the evidence base is typically weak – hence the smaller
circle. Secondly, the greater complexity of multimorbid patients justifies larger patient
circumstances and patient preferences circles because they tend to dominate management
decisions.
To further exemplify this, consider the following scenarios. A healthy 40-year-old
patient has a hip fracture requiring surgery. The course of action is relatively straightforward:
surgery, intensive rehabilitation, and pain management. These therapeutic interventions are
supported by research (Morrison and Siu, 2023), and individual patient preferences and
circumstances are not expected to have a significant impact in decision-making. Young
patients typically recover well from these care measures with minimal risk of harm (ibid.).
Conversely, if an 80-year-old multimorbid patient with heart disease, obesity, chronic
anaemia and respiratory problems sustains a similar hip fracture, the situation becomes more
complex. While surgery may fix the fracture, the associated risk of complications and
Figure 3. Clinical decision-making differs between single-condition and
multimorbidity situations.
18
difficult recovery make the final outcome uncertain (ibid.). Patient preferences would likely
be determinant here. If this patient showed preference for a non-surgical approach focused on
pain control and rehabilitation, even if that meant bearing walking sequelae, that would be
the right clinical decision to make. Hence the bigger circles for patient preferences and
circumstances in multimorbidity situations, as depicted in Figure 3.
In view of the above, translating research evidence to clinical practice in a
multimorbidity scenario, considering the weaker research evidence and greater influence of
individual patient variables in decision-making, is harder than in a non-multimorbidity one.
In other words, the knowledge-to-practice gap is wider. Further, as I argued in this chapter, a
condition-focused guideline-driven EBM framework amplifies the multimorbidity challenge.
All things considered, we are now in a good position to explore potential solutions to this
challenge.
19
3. Putative solutions for the multimorbidity challenge are limited guideline-focused
approaches
So far I have argued that managing multimorbidity within a guideline-driven EBM
system is problematic. In the face of this growing challenge, several independent researchers
have joined efforts to address it. Based on a non-systematic literature review, I identified two
distinct approaches, here called “approach A” and “approach B”. I will first briefly introduce
and discuss important limitations of each approach, and then argue that neither looks
promising because both stem from the guideline-focused dimension of EBM.
3.1. Approach A: better crafting single-condition guidelines for multimorbid patients
Approach A is best represented by the solution offered by a leading international
research collaboration called “Improving Guidelines for Multimorbid Patients”. Interestingly,
the group claims that their work is embedded within the EBM framework so that they
leverage EBM to create guidelines that better fit the needs of multimorbid patients (Boyd and
Kent, 2014). The approach consists in “crafting guidelines” [sic] to address multimorbidity-
related issues at all major steps of guideline development, including topic scoping, assessing
quality of evidence and trading off benefits and harms (Uhlig et al., 2014). Furthermore,
guideline developers are allowed to consider “typical” patient preferences to guide their
recommendations, provided they are explicitly stated. Ultimately, these guidelines remain
interested in single conditions.
I shall concentrate on three limitations of this approach: (i) the inability to account for
the sheer scale of individual variation (ii) the increasingly impractical size of guidelines; and
(iii) two points of conflict with EBM.
3.1.1. The inability to account for the sheer scale of individual variation
Consider the following scenario. Mr. Robertson is a 78-year-old patient with diabetes,
kidney disease, heart disease and gout5, who lives alone and usually does not drink much
water. It is a hot summer day and the patient develops a gout attack. Steroid drugs, taken
5 A common disease characterized by recurrent attacks of inflamed joints.
20
once a day, are excellent to treat gout attacks but are generally not advised in diabetic patients
because they increase blood sugar levels (Gaffo, 2023). Yet, physicians know that alternative
drugs, which are taken twice-daily, might pose even greater risks for patients with pre-
existing kidney or heart disease (ibid.). In Mr. Robertson’s context, this is particularly
worrisome. Not only does he have established kidney and heart disease, but also lacks an
adequate water intake, which further predisposes him to kidney injury, especially in the
summer. Let’s say that Mr. Robertson struggles with polypharmacy and is keen on simple
regimens. A once-daily steroid drug regimen is likely to suit him better. His preferences
could rightly be the tipping point in this clinical decision trade-off. But it could also be the
case that other individual circumstances weighed more in the decision-making process and
the decision would be different.
Mr. Robertson’s case is merely a single instance of a complex multimorbidity
situation. However, the number of possible combinations of conditions and individual
variables is colossal. There is no way to put into a guideline all the unique problems and
concerns of multimorbid patients. One could even consider the changing nature of
multimorbidity (conditions typically accrue over time), the individual dynamics of each
condition, and the evolving profile of patient circumstances and preferences. Yet, for lack of
space, I will not address these additional points.
3.1.2. The increasingly impractical size of guidelines
One practical concern immediately arises for any guideline user. As noted in section
2.2, guideline documents are increasingly unwieldy resources. Extending them in length to
include numerous multimorbidity considerations will likely render them useless.
Recall the present-day situation of diabetes guidelines (Table 1). Indeed, the latest
guideline versions seem to have already partially adopted approach A, making
recommendations for some selected, albeit vague, situations of multimorbidity – e.g. “the
elderly”. However, if guidelines are to account for several potentially relevant situations of
multimorbidity, instead of 100-300 pages they can easily reach 1000 pages! One might argue
that the purpose of this approach is that guideline developers select the most relevant
considerations only, and that the moderate increase in size of guidelines would not
compromise their use because of their higher applicability to multimorbid patients. However,
this leads to other problems, as I explain in the next section.
21
3.1.3. Two points of conflict with EBM
Recall from Chapter 1 that EBM first appeared as a response to an excessive reliance
on expert judgement as evidence that an intervention has a given effect. Epistemologically
speaking, EBM gives primacy to high-quality research. Evidence-based guidelines, in
principle, are based on a similar rationale and seek to ground recommendations in research
evidence as much as possible. However, Uhlig et al. (2014) grant that guideline workgroups
must make complex judgements in the absence of high-quality or direct evidence. As I have
shown in the previous chapters, the absence of such evidence is the norm for multimorbidity.
My point is that, if guideline recommendations become primarily based on the “expert
judgement” of guideline developers, guidelines are not as “evidence-based” as they purport to
be.
One might counter this point by arguing that, when based on a systematic review of
the literature, expert consensus is actually the “best available evidence” for managing
multimorbidity, in consonance with our definition of guidelines. However, to reiterate, EBM
is about integrating the best available research evidence, not expert judgement, with other
elements in order to make informed decisions in clinical practice. Guidelines are also meant
to stem from the appraisal of research evidence.
The other way that approach A conflicts with EBM has to do with the location of
expertise. As I alluded to in the first chapter, EBM considers clinical expertise the key locus
of judgement for decision-making and locates it at the level of the practitioner (I will return
to this point in the next chapter). In contrast, Uhlig et al. (2014) place clinical expertise at the
level of guideline developers. This becomes clear when the authors mention that guideline
panels must make complex judgements 1) in the absence of high-quality or direct evidence,
as I explained above, 2) when selecting the “relevant” clinical conditions and circumstances
to address, and 3) when considering “typical” values and preferences. So, the problem is that
approach A locates clinical expertise in the wrong place. As I shall argue in the next chapter,
the sort of judgements just mentioned are to be done by physicians, in the clinic, when they
make clinical decisions.
22
3.2. Approach B: developing specific guidelines for multimorbidity
Approach B consists in developing multimorbidity-specific evidence-based
guidelines. A few such guidelines have been developed internationally (Farmer et al., 2016,
Boyd et al., 2019, Onder et al., 2022). Unlike in approach A, these guidelines typically focus
on shared features of multimorbid patients (frailty, polypharmacy, etc.).
A recent systematic multimorbidity and polypharmacy guideline review (Muth et al.,
2019) identified key themes and recommendations in those guidelines. As Figure 4 shows,
these arguably look more like general principles of care for multimorbidity, rather than
recommendations for specific clinical situations, as per the definition of guidelines (see
section 1.2). These multimorbidity guidelines do make some “specific recommendations”, as
they call them, such as undertaking a yearly medication review. However, others, despite
being termed “specific”, veer off towards generic guidance too – for example, “establish
disease and treatment burden” or “explore patient’s expectations and objectives about
treatments before prescribing” (Muth et al., 2019, p.280). I will comment on this shortly.
Figure 4. Key themes and recommendations in multimorbidity guidelines.
Adapted from Box 2 in (Skou et al., 2022).
23
I will build my case here around two points: the uncertainty about the usefulness of
this generic guidance, and, much like in approach A, the overreliance on the expert
judgement of guideline developers and the mislocation of clinical expertise (the same two
points of conflict with EBM).
3.2.1. How useful is generic guidance?
I contend that these guidelines are futile because, as they are, they cannot provide any
meaningful support for decision-making. Such generic guidance looks more “primary care
101”, or what primary care physicians know already from their basic training. As I shall
explain better in the next chapter, primary care physicians, through education and regular
encounters with multimorbid patients in their practice, naturally develop some intuitive
general principles of care for such patients that contribute to their clinical expertise in
providing evidence-based individualised care.
Nevertheless, those principles in Figure 4 could be formally integrated into medical
education and training programs. That way junior physicians could be better prepared to
handle multimorbidity in clinical practice from an early stage. Returning to Mr. Robertson’s
case, the consideration of risk factors or treatment burden from polypharmacy, for example,
is part of the provision of primary care.
3.2.2. The same two points of conflict with EBM
Not surprisingly, approach B also suffers from the dearth of research evidence on
multimorbid patients. Developing guidelines under these conditions leads to an overreliance
on the guideline developer’s speculatory judgement and the promotion of “consensus-based
guidelines”, which, as I argued for approach A, is problematic. Finally, approach B also
crucially mislocates clinical expertise (see section 3.1.3).
3.3. The root of the problem: guideline-driven EBM
A major constraint on both approaches is the limited available research evidence on
multimorbidity. However, is the success of either approach contingent on the research
24
context and state of knowledge? Or is there a deeper root of the problem, and both
approaches are limited from a conceptual standpoint, specifically that both are premised on
the guideline-focused dimension of EBM? I will seek to answer these questions.
Guidelines can be said to have a limited role in multimorbidity in a “weak sense” if
their success is contingent on the research context and state of knowledge. Indeed, I have
argued that a major challenge in translating research evidence into the care of multimorbid
patients is the fact that these patients are seldomly enrolled in conventional clinical trials,
notably for two reasons: (i) medical research tends to be single-condition focused, and (ii)
multimorbid patients face practical difficulties in accessing research units. Moreover, it is
also worth noting that trials of multimorbidity interventions (e.g. interventions targeting
polypharmacy or frailty) face their own set of challenges (Skou et al. 2022).
However, if efforts were made not only to overcome the hurdles in involving
multimorbid patients in trials, but also to address the challenges in existing trials of
multimorbidity interventions, both arguably achievable in the future, the multimorbidity
evidence base would improve. In that case physicians would indeed be in a more
advantageous position to manage multimorbidity. For example, more research evidence on
the effects of polypharmacy and interventions to reduce it (when appropriate) could
effectively inform clinical decisions.
Nevertheless, even if the robustness of research evidence for multimorbidity was not a
concern, both approaches A and B would still be limited in addressing the multimorbidity
challenge because individual patient variables always exert a dominant influence in decision-
making. Guidelines cannot account for each individual patient’s context, set priorities, or sort
out trade-offs. So, we can see that the success of either approach is not contingent on the
research context and state of knowledge, and that there must be a deeper root of the problem.
Thus, these guideline-focused approaches for multimorbidity seem bound to be limited not in
a “weak” sense, but in a “strong” sense.
In theory, both approaches position themselves within the EBM framework. In reality,
they are rooted in guideline-driven EBM, i.e. they reduce EBM to guidelines. Despite both
approaches ultimately recognising the need for clinical expertise, they locate it in the wrong
place; they still belittle clinical expertise properly located at the level of the practitioner. As I
have argued in the previous chapters, guideline-driven EBM is epistemologically limited in
25
managing individual patients because it belittles clinical expertise (properly located). This is
the limitation that thwarts both approaches, and likely thwarts any guideline-focused
approach to the multimorbidity challenge.
26
4. More clinical expertise is needed for multimorbidity care
Proposed solutions to the multimorbidity challenge, despite being prima facie EBM
approaches, have hitherto insisted on a limited version of EBM that downplays clinical
expertise. In this final chapter, I will argue that clinical expertise has an augmented role in
multimorbidity care.
4.1. What is clinical expertise?
Defining clinical expertise remains surprisingly difficult. I do not think, though, that
clinical expertise needs to be a very precise concept to be practically useful. My aim here is
simply to provide a concise notion of it, and then expand it to some degree in section 4.2
using the case of multimorbidity.
As I alluded to in Chapter 1, clinical expertise can be regarded as a kind of knowledge
that enables the integration of research evidence with individual patient circumstances and
preferences. As I build my argument in the next section, I will appeal to other notions that
offer valuable insights into what clinical expertise is, namely tacit knowledge, contingency on
clinical encounters, and clinical experience. Yet, as a preliminary sketch, clinical expertise
typically involves tacit elements and increases with experience; clinical expertise can only be
acquired and used by persons – typically physicians, but other practitioners, like nurses and
physiotherapists, can too – and its performance is contingent on clinical encounters. For our
purposes, a clinical encounter encompasses any direct human interaction between a patient
and a physician. This is not to say that clinical teams do not play a role, particularly in the
acquisition of clinical expertise. I will acknowledge this later in section 4.2.1.
This is clinical expertise as I understand it, an interpretation that draws on Sackett’s
original definitions of clinical expertise and EBM, as well as on the EBM-Ex model. Because
clinical expertise performs certain tasks to enable decision-making, I conceptualise clinical
expertise as knowledge-how. Furthermore, as I will show, this is the kind of knowledge that is
needed to a greater extent when dealing with multimorbidity. References to “judgement”,
“experience” and “skill”, among others, are also relevant to our discussion, even if they
emphasise different aspects of this sort of knowledge (Wieten, 2018).
27
4.2. The augmented role of clinical expertise in multimorbidity
Back in Chapter 2 I saved the discussion about the implications of multimorbidity to
clinical expertise for this chapter. There I argued that patient circumstances and preferences
exert a dominant influence in decision-making for multimorbid patients compared to single-
condition situations. I visually represented this by enlarging their respective circles in the
EBM-Ex diagram. Similarly, the research evidence circle was scaled down to illustrate the
weaker evidence base of multimorbidity and its limited support for clinical decisions. The
result was a disharmonious arrangement of the three circles, reflecting the wider knowledge-
to-practice gap in multimorbidity.
In simple visual terms, Figure 5 (an updated version of Figure 3) depicts what
happens to clinical expertise in multimorbidity: to cover the overlaps and unify the three
circles, clinical expertise is needed to a greater extent. Its role is not different in essence, only
augmented.
Figure 5. Clinical expertise is needed to a greater extent in multimorbidity
compared to single-condition situations.
28
To further explain this augmentation of the role of clinical expertise, I will use a
clinical case, concentrating on two key tasks of clinical expertise that, I shall argue, are
required to a higher degree when caring for multimorbid patients: (1) the prioritisation of
health problems and (2) the translation of research evidence to the individual patient. Recall
Mr. Robertson from Chapter 2, an elderly patient with diabetes, kidney and heart disease,
gout, who lives alone and, I now add, has no family support. Now suppose Mr. Robertson
presents for a routine appointment with his primary care physician and his blood tests reveal
a significant worsening of diabetes and a decline in kidney function, despite him feeling “as
usual”. We will see how vital these two taks of clinical expertise are.
4.2.1. The prioritisation of health problems
As I argued in Chapter 2, the prioritisation of health problems is an essential part of
multimorbidity care. It might be because some problem is the source of symptoms or affects
prognosis, or that addressing all conditions at once leads to conflicting or cumbersome
therapeutic plans.
Given Mr. Robertson’s exam results, his physician is likely to prioritise the
management of diabetes and kidney disease over the other conditions. Mr. Robertson’s
physician somehow knows that these conditions affect his short-term prognosis more than the
others, thus demanding certain interventions to halt clinical deterioration. Heart disease and
gout still need their own management, but these conditions are currently clinically stable, and
can be left temporarily in “stand-by”. However, these conditions may still influence
management decisions. Suppose that now Mr. Robertson requires more frequent diabetes
consultations and kidney function testing, particularly if initiating insulin therapy for
diabetes. His painful gouty joints make it difficult for him to walk to the health centre for
such visits, though. Let’s say that the physician, in a process involving the patient as possible,
prioritises the management of diabetes over the worsening kidney disease. The reasons for
that depend on how the physician judges individual patient factors, such as the degree and
speed at which each condition worsened. It is possible that the physician perceives the change
in the kidney laboratory value as a normal variation rather than a true, prognostically
relevant, decline in kidney function. This differs from a situation where a single condition is
the sole focus of clinical decisions and no other condition needs to be side-lined. Clinical
expertise in the clinic plays an important role in judging which interactions are most relevant
29
to prioritise problems. Taking into consideration these individual circumstances and Mr.
Robertson’s preference for less frequent follow-up, the physician might suggest a prudent
approach and prescribe a low insulin dose. Indeed insulin therapy may be dangerous without
close monitoring and with worsening kidney function (Wexler, 2023).
Let’s see how the notions of tacit knowledge, contingency on clinical encounters and
clinical expertise expand our understanding of problem prioritisation. First, tacit knowledge
plays a dominant role in the sort of judgements I just described (Henry, 2010), which allow
the physician to know how to prioritise. Even if Mr. Robertson’s physician tried to list all the
individual features she thought influenced the decision to prioritise diabetes, many other
signals from their clinical encounter and pieces of knowledge that influenced her judgments
would likely remain unnoticed (ibid). Tacit knowledge is a sort of knowledge-how that is
grounded in practice, which does not rely on explicit formulation to be effective (e.g. riding a
bicycle or driving a car) (Gascoingne and Thornton, 2013). Moreover, tacit knowledge
accrues from experience (Polanyi, 1966).
Second, the prioritisation of problems is contingent on clinical encounters; it happens
in the clinic. Within the context of multimorbidity, to prioritise health problems is to make
decisions about which conditions should be addressed first. I endorse Pellegrino (1979)’s
view that clinical decisions are made within the clinical encounter. This is because clinical
decisions arise from the interplay between persons – a point further developed in the next
section. Moreover, the clinical encounter is the key locus of tacit knowledge in clinical
practice (Henry, 2006).
Third, clinical experience is required to know how to prioritise problems in practice.
Clinical experience encompasses the knowledge derived from the direct care of patients,
either based on personal experience or the personal experience of others (Tonelli, 2011).
Recall that EBM depreciates the role of accumulated clinical experience – so-called “expert
judgement” – as evidence that an intervention works. Indeed, for those purposes, clinical
experience is often misleading. Personal experience is characterised by small numbers of
patients and is prone to several cognitive biases (Elstein and Schwarz, 2002). Although
clinical experience fell into disrepute with the advent of EBM, its importance in the process
of acquiring clinical expertise was upheld by the EBM founders (see section 1.1). Until this
day, personal clinical experience remains highly valued for the physicians’ performance and
its importance widely recognised in medical education. In Europe, for instance, most
30
physicians undergo long periods of training after graduation (more than five years) (Union
Européene Des Médecins Spécialistes, 2023). Additionally, it is worth noting that physicians
learn from the experience of others. Physicians learn from ongoing and vibrant dialogue with
peers, and increasingly in virtual environments where physicians form communities and
share personal experiences, contextualised cases, and critical appraisals of research evidence
(Elwyn et al., 2016). Notably, this does not diminish the responsibility of the physician to
continuously search for the best available research evidence.
In sum, experience helps physicians make sense of the idiosyncrasies of each clinical
encounter. Not only in determining which ones are relevant for the outcome of interest and
which ones to discard, but also in making estimates for patients for whom there is little, if
any, research data, as in multimorbidity (Flores, 2016, p.37). In Mr. Robertson’s case, a
novice physician could feel unnecessarily worried about the patient’s altered renal function
blood tests, when research evidence supports that “normal” variations in renal function tests
occur (Inker and Perrone, 2023). Having seen many similar instances in practice, experienced
physicians typically can better, and more confidently, judge the clinical significance of such
variations, even if that experience and confidence can originate other cognitive biases (e.g.
overconfidence bias) (O'Sullivan and Schofield, 2018). Generating sound research evidence
supporting the role of clinical experience in decision-making is challenging due to the variety
of reasons behind clinical decisions, but it remains a worthwhile endeavour.
4.2.2. The translation of research evidence to the patient of interest
Let me delve deeper into Mr. Robertson’s physician’s decision-making process of
intensifying diabetes therapy to show that clinical expertise has an augmented role in
translating research evidence to individual multimorbid patients. As I shall argue, some form
of skilled judgment that considers patients in all their individuality operates to a greater
extent.
Research evidence suggests that intensifying diabetes therapy with insulin therapy is
the fastest and most effective way to control high blood sugar levels. Yet, this is done at the
expense of considerable treatment burden (cumbersome injections, finger-pricking and
frequent health centre visits, among others), possible severe drug adverse events particularly
with concomitant kidney disease, and no guarantee that this treatment reduces the chances of
31
dying or having a serious event, like a stroke (Wexler, 2023). Presumably Mr. Robertson’s
physician knows about all this and has a face-to-face discussion with the patient about insulin
therapy. Mr. Robertson, who is feeling well, manifests reluctance towards such an onerous
treatment.
On the other hand, if Mr. Robertson were a young man and only had diabetes, he
would be more willing to consider insulin therapy. When single conditions are the focus of
care, which typically happens with young patients, both physicians and patients tend to focus
on strict condition control, believing that benefits will accrue over time. Diabetes is
paradigmatic in this respect, and treatment choices differ between older and younger patients
in various ways (Munshi, 2023). For example, young patients usually cope better with insulin
without assistance due to their greater physical and cognitive capabilities. Additionally,
research evidence on diabetes therapy is much weaker for older, typically multimorbid,
patients (ibid.).
Let’s look at this decision-making process through the EBM-Ex diagram – see Figure
6. We can see that dealing with Mr. Robertson's multimorbidity requires a higher level of
clinical expertise compared to a “diabetes only” situation. There is a larger number of factors
to consider and integrate, including the paucity of research evidence. The absence of family
support is notable here because insulin therapy could be manageable with family members
Figure 6. The decision-making process of diabetes therapy intensification in
Mr. Robertson’s case versus a “diabetes only” situation.
32
providing assistance. To complicate things further, patients often feel uncertain about their
preferences (Haynes et al., 2002). Indeed, Mr. Robertson showed reluctance towards insulin
therapy. The physician may have captured that as a preference for a more convenient
therapeutic option. This preference, together with other individual factors, needs to be
integrated with existing research evidence to make an individualised evidence-based
decision.
In view of the above, insulin therapy seems far from being the ideal choice for Mr.
Robertson. Alternative options could be improving diet habits and trying other medications
(Munshi, 2023). While less effective in rapidly controlling blood sugar levels, these
alternatives are much easier to manage and carry a lower risk of side effects.
Here too I want to emphasise the importance of tacit knowledge, the clinical encounter
and clinical experience in the translation of research evidence to the individual multimorbid
patient. The reasons are multiple and interrelated, which make them difficult to explain
briefly, but I will articulate a few.
Firstly, much of what permits clinical expertise to unify the EBM-Ex’s tripartite
decision-making model is tacit knowledge. There is an ineliminable role for tacit judgement
in clinical expertise that enables integrative tasks and the application of general knowledge to
particular cases (Thornton, 2006). In multimorbidity, because the patient’s context
(individual circumstances and preferences) exerts a dominant influence in decision-making,
more individual factors need to be integrated with research evidence. A tacit dimension of
clinical expertise is often key to capture all that individuality and judge what matters for a
given decision (ibid.).
Secondly, the clinical encounter is essential for clinical expertise to operate because (i)
patient preferences and circumstances emerge within it, and (ii) the individualisation of
research evidence to multimorbid individuals involves a holistic complexity appreciation that
is best performed in the clinical encounter, not in guidelines’ “consensus panels”. For
example, patients develop their preferences for a certain decision as they learn about
competing benefits and harms. When there are multiple competing factors, as in
multimorbidity, emerging preferences assume increased importance in sorting out a larger
number of trade-offs. This learning of competing factors takes place within the clinical
encounter. It is there that the patient and the physician both learn, through personal
communication, what is at play for decision-making. So, tacit elements of communication are
33
also relevant because they precede explicit understanding – for instance, clinical decisions
based on telephone consultations are often harder to make because tacit elements of
communication are missed (e.g. face expressions conveying personal preferences) (Henry,
2006).
Regarding holistic complexity appreciations, the clinical encounter seems to be where
clinical expertise better performs such appreciations precisely because that is where the
patient’s context effectively emerges. Although clinical expertise has an important tacit
dimension, it also involves a good amount of explicit knowledge about the patient’s context
brought forth during the clinical encounter – for example, Mr. Robertson mentioning that he
lives alone.
Lastly, clinical expertise is a practical skill acquired through experience and training.
An analogy with craftsmanship is helpful to understand the role of experience in effectively
translating research to multimorbid patients. Clinical expertise, much like woodworking or
any other craft skill, is a professional ability to perform complex tasks requiring individual
decisions and judgements (McWhinney, 1978). As I have been arguing, multimorbid patients
are complex patients to manage. The more experienced the physician, the more refined
clinical judgement gets in translating research evidence in tailored decisions for complex
patients (Tonelli, 2011).
4.2.3. Final considerations
One might say that what I am doing here is showing that clinical decision-making can
be made explicit. Though I did make explicit that opting for second-line treatments for Mr.
Robertson would be an appropriate clinical decision, I must clarify that this was based on
certain assumptions made for the sake of this discussion – for instance, the supposition that
Mr. Robertson wants to avoid insulin injection therapy. However, such knowledge about Mr.
Robertson’s preferences can only be acquired in real clinical encounters, and often only
tacitly. Acquiring such knowledge and factoring it into a decision-making process is a
responsibility of clinical expertise (Haynes et al., 2002). Even if clinical expertise performs
mostly in a tacit way, that does not mean that physicians cannot, at times, make explicit some
parts of a decision-making process, as I did here, for the purposes of medical records and
discussions.
34
Conclusion
The multimorbidity challenge is not new and keeps growing. Professed EBM attempts
to address it have so far insisted on improving guidelines for multimorbid patients. However,
as I argued in Chapter 3, these guideline-focused approaches are fighting a losing battle. Such
putative solutions are grounded in a guideline-driven version of EBM that, I argued, cannot
effectively translate research knowledge to individual patients, whereas EBM in its fullest
sense can. The nub of the issue is that a guideline-driven EBM pushes the eyes away from
physicians’ clinical expertise, which is the ultimate EBM element that allows the provision of
evidence-based individualised care.
As I explained throughout the dissertation, a major challenge in delivering evidence-
based care to multimorbid patients is the lack of high-quality research evidence. While I
strongly support appeals for more and better research on multimorbidity, individual patient
circumstances and preferences will always tend to hold more sway in decision-making for
multimorbid patients than for their counterparts. Thus, one-size-fits-all approaches like
guidelines will never be able to account for the idiosyncratic nature of multimorbid patients,
let alone negotiate multiple conflicting warrants for action. However, clinical expertise can.
As I argued in Chapter 5, clinical expertise is the epistemic element that operates, in the
clinic, the incorporation of clinical nuances and other contextual factors into clinical
decisions. Regrettably, guideline-driven EBM belittles its role in clinical practice.
Understanding the limitations of guidelines within the context of multimorbidity has
important implications. For medical education, the recognition that clinical expertise plays a
vital epistemic role in EBM, especially when dealing with complex patients, means that
primary care training, and likely all medical training, must continue to be structured around
practical experience. It is worth noting that this does not diminish the importance of teaching
other key EBM skills to students and graduated physicians, specifically the ability to find and
critically appraise research evidence at a basic level. Implications for primary care practice
are at least twofold: first, physicians must be allowed to practice in a discretionary way
without being accused of deviating from guidelines and EBM itself; second, perhaps the time
and energy that primary care organisations and physicians put into implementing guidelines
could be more effectively used elsewhere. Finally, the consequences of a dominant guideline-
driven EBM extend way outside of the clinic. Such a view of EBM influences how
researchers, medical bodies and governments address the multimorbidity challenge. Once we
35
acknowledge the key role of clinical expertise in multimorbidity care, it will be important to
learn which strategies allow physicians to better use their clinical expertise. For example,
increasing primary care consultation times is a strategy that aligns with the recognition of the
practical import of clinical expertise, emphasising the importance of the clinical encounter,
and which can be subject to research.
This dissertation is a defence of EBM as envisaged by its early proponents. By calling
for more clinical expertise in multimorbidity care I am not advocating for a blind trust in
some kind of “expert judgment”. Rather I am urging a rejection of guideline-driven EBM and
revival of EBM in its fullest sense. A misalignment between the EBM ideal and actual
practice does not mean that we have to reject EBM, or even guidelines, altogether. We may
recognise that it has been “hijacked” (Ioannidis, 2016), use guidelines when appropriate and
only as mere tools (a few can indeed be useful, if concise and grounded in robust research
evidence), and bring clinical expertise to the fore again.
To my knowledge, this dissertation is the first formal exposition of the limitations of
guideline-focused approaches to multimorbidity. I arrived at this point primarily driven by a
personal motivation. As someone who has personally faced the multimorbidity challenge and
is a strong supporter of EBM, I felt compelled to put into a formal argument what I intuited
from reflecting on my own practice and that of others in primary care: that guideline-driven
EBM might lead to suboptimal or even harmful care. Also, complex multimorbid patients are
typically in a state of both physiological and social vulnerability that prevents them from
advocating for themselves for better care. This reinforces the need to properly address the
multimorbidity challenge. As a primary care physician and researcher puts it, “We need to
stop giving guidelines such reverence” (McCormack, 2016).
36
References
American Diabetes Association (2007) 'Standards of Medical Care in Diabetes – 2007', Diabetes
Care, 30(Suppl 1), S4-S41.
American Diabetes Association (2020) 'Standards of Medical Care in Diabetes - 2021', Diabetes
Care, 44(Suppl 1), S1–S232.
Bird, A. (2011a) 'The epistemological function of Hill's criteria', Preventive Medicine, 53(4), pp. 242-
245.
Bird, A. (2011b) 'What can philosophy tell us about evidence-based medicine? An assessment of
Jeremy Howick’s The philosophy of evidence-based medicine', International Journal of
Person Centered Medicine, 1(4), pp. 642-648.
Blunt, C. (2015) Hierarchies of evidence in evidence-based medicine. PhD thesis, London School of
Economics and Political Science, London.
Boyd, C., Smith, C. D., Masoudi, F. A., Blaum, C. S., Dodson, J. A., Green, A. R., Kelley, A.,
Matlock, D., Ouellet, J., Rich, M. W., Schoenborn, N. L. and Tinetti, M. E. (2019) 'Decision
Making for Older Adults With Multiple Chronic Conditions: Executive Summary for the
American Geriatrics Society Guiding Principles on the Care of Older Adults With
Multimorbidity', Journal of the American Geriatrics Society, 67(4), pp. 665-673.
Boyd, C. M., Darer, J., Boult, C., Fried, L. P., Boult, L. and Wu, A. W. (2005) 'Clinical Practice
Guidelines and Quality of Care for Older Patients With Multiple Comorbid Diseases:
Implications for Pay for Performance', JAMA, 294(6), pp. 716-724.
Boyd, C. M. and Fortin, M. (2010) 'Future of Multimorbidity Research: How Should Understanding
of Multimorbidity Inform Health System Design?', Public Health Reviews, 32(2), pp. 451-
474.
Boyd, C. M. and Kent, D. M. (2014) 'Evidence-Based Medicine and the Hard Problem of
Multimorbidity', Journal of General Internal Medicine, 29(4), pp. 552-553.
Canadian Diabetes Association (2003) 'Clinical Practice Guidelines Expert Committee. Canadian
Diabetes Association 2003 clinical practice guidelines for the prevention and management of
diabetes in Canada', Can J Diabetes, 27(Suppl 2), S1-S163.
Cartwright, N. (2007) 'Are RCTs the gold standard?', BioSocieties, 2(1), pp. 11-20.
Copeland, S. (2020) 'The Guidelines Challenge', in Anjum, R.L., Copeland, S. & Rocca, E. (eds.)
Rethinking Causality, Complexity and Evidence for the Unique Patient: A CauseHealth
Resource for Healthcare Professionals and the Clinical Encounter. Cham: Springer
International Publishing, pp. 95-110.
Cosentino, F., Grant, P. J., Aboyans, V., Bailey, C. J., Ceriello, A., Delgado, V., Federici, M.,
Filippatos, G., Grobbee, D. E., Hansen, T. B., Huikuri, H. V., Johansson, I., Jüni, P., Lettino,
M., Marx, N., Mellbin, L. G., Östgren, C. J., Rocca, B., Roffi, M., Sattar, N., Seferović, P.
M., Sousa-Uva, M., Valensi, P., Wheeler, D. C. and Group, E. S. D. (2019) '2019 ESC
Guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration
with the EASD: The Task Force for diabetes, pre-diabetes, and cardiovascular diseases of the
European Society of Cardiology (ESC) and the European Association for the Study of
Diabetes (EASD)', European Heart Journal, 41(2), pp. 255-323.
37
Diabetes Canada Clinical Practice Guidelines Expert Committee (2018) 'Diabetes Canada 2018
Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada', Can
J Diabetes, 42(Suppl 1), A1-A18, S1-S326.
Djulbegovic, B., Guyatt, G. H. and Ashcroft, R. E. (2009) 'Epistemologic Inquiries in Evidence-Based
Medicine', Cancer Control, 16(2), pp. 158-168.
Eddy, D. M. (2005) 'Evidence-based medicine: a unified approach', Health affairs, 24(1), pp. 9-17.
Elstein, A. S. and Schwarz, A. (2002) 'Clinical problem solving and diagnostic decision making:
selective review of the cognitive literature', BMJ, 324(7339), pp. 729-732.
Farmer, C., Fenu, E., O’Flynn, N. and Guthrie, B. (2016) 'Clinical assessment and management of
multimorbidity: summary of NICE guidance', BMJ, 354, pp. i4843.
Farquhar, C. M., Kofa, E. W. and Slutsky, J. R. (2002) 'Clinicians' attitudes to clinical practice
guidelines: a systematic review', Medical Journal of Australia, 177(9), pp. 502-506.
Flores, L. (2016) Clinical judgment in the era of Evidence Based Medicine. Doctor of Philosophy,
King’s College London.
Fortin, M., Soubhi, H., Hudon, C., Bayliss, E. A. and Van den Akker, M. 2007. Multimorbidity's
many challenges. BMJ.
Fuller, J. (2016) The new medical model: Chronic disease and evidence-based medicine. Doctor of
Philosophy, University of Toronto.
Fuller, J. (2021) 'The myth and fallacy of simple extrapolation in medicine', Synthese, 198(4), pp.
2919-2939.
Gaffo, A. L. (2023) 'Treatment of gout flares', in: UpToDate, Post, TW (Ed), UpToDate, Waltham,
MA (Accessed on August 18, 2023).
Garber, A. (2005) 'Evidence-Based Guidelines As a Foundation For Performance Incentives', Health
Affairs, 24(1), pp. 174-179.
Gascoingne, N. and Thornton, T. (2013) Tacit Knowledge. Durham: Acumen.
Greenhalgh, T., Howick, J. and Maskrey, N. (2014) 'Evidence based medicine: a movement in crisis?',
BMJ, 348, pp. g3725.
Grumbach, K. (2003) 'Chronic Illness, Comorbidities, and the Need for Medical Generalism', The
Annals of Family Medicine, 1(1), pp. 4-7.
Guthrie, B. and Boyd, C. M. (2018) 'Clinical Guidelines in the Context of Aging and Multimorbidity',
Public Policy & Aging Report, 28(4), pp. 143-149.
Guyatt, G., Cairns, J., Churchill, D., Cook, D., Haynes, B., Hirsh, J., Irvine, J., Levine, M., Levine,
M., Nishikawa, J., Sackett, D., Brill-Edwards, P., Gerstein, H., Gibson, J., Jaeschke, R.,
Kerigan, A., Neville, A., Panju, A., Detsky, A., Enkin, M., Frid, P., Gerrity, M., Laupacis, A.,
Lawrence, V., Menard, J., Moyer, V., Mulrow, C., Links, P., Oxman, A., Sinclair, J. and
Tugwell, P. (1992) 'Evidence-Based Medicine: A New Approach to Teaching the Practice of
Medicine', JAMA, 268(17), pp. 2420-2425.
Guyatt, G. H., Meade, M. O., Jaeschke, R. Z., Cook, D. J. and Haynes, R. B. (2000) 'Practitioners of
evidence based care: Not all clinicians need to appraise evidence from scratch but all need
some skills', BMJ, 320(7240), pp. 954-955.
38
Guyatt, G. H., Rennie, D., Meade, M. O. and Cook, D. J. (2008) Users' guides to the medical
literature. Essentials of evidence-based clinical practice. 2nd edn.: McGraw-Hill
Professional.
Hanson, K., Brikci, N., Erlangga, D., Alebachew, A., De Allegri, M., Balabanova, D., Blecher, M.,
Cashin, C., Esperato, A., Hipgrave, D., Kalisa, I., Kurowski, C., Meng, Q., Morgan, D., Mtei,
G., Nolte, E., Onoka, C., Powell-Jackson, T., Roland, M., Sadanandan, R., Stenberg, K., Vega
Morales, J., Wang, H. and Wurie, H. (2022) 'The Lancet Global Health Commission on
financing primary health care: putting people at the centre', The Lancet Global Health, 10(5),
pp. e715-e772.
Haynes, R. B., Devereaux, P. J. and Guyatt, G. H. (2002) 'Clinical expertise in the era of evidence-
based medicine and patient choice', BMJ Evidence-Based Medicine, 7(2), pp. 36-38.
Hayward, R. S., Wilson, M. C., Tunis, S. R., Bass, E. B. and Guyatt, G. (1995) 'Users' Guides to the
Medical Literature: VIII. How to Use Clinical Practice Guidelines A. Are the
Recommendations Valid?', JAMA, 274(7), pp. 570-574.
Henry, S. G. (2006) 'Recognizing Racit Knowledge in Medical Epistemology', Theoretical Medicine
and Bioethics, 27(3), pp. 187-213.
Henry, S. G. (2010) 'Polanyi's tacit knowing and the relevance of epistemology to clinical medicine',
Journal of Evaluation in Clinical Practice, 16(2), pp. 292-297.
Howick, J. H. (2011) The Philosophy of Evidence-based Medicine. Chichester, West Sussex, UK:
Wiley.
Inker, L. A. and Perrone, R. D. (2023) 'Assessment of kidney function', in: UpToDate, Post, TW (Ed),
UpToDate, Waltham, MA (Accessed on August 20, 2023).
Ioannidis, J. P. A. (2016) 'Evidence-based medicine has been hijacked: a report to David Sackett',
Journal of Clinical Epidemiology, 73, pp. 82-86.
Malterud, K. (1995) 'The legitimacy of clinical knowledge: Towards a medical epistemology
embracing the art of medicine', Theoretical Medicine, 16(2), pp. 183-198.
Martini, C. (2021) 'What “Evidence” in Evidence-Based Medicine?', Topoi, 40(2), pp. 299-305.
McCormack, J. P (2016). End of Guidelines - a parody of End of the Line, Youtube. Available at:
https://www.youtube.com/watch?v=DHDnqQ_mCBA (Accessed 21 July, 2023).
McWhinney, I. R. (1978) 'Medical Knowledge and the Rise of Technology', The Journal of Medicine
and Philosophy: A Forum for Bioethics and Philosophy of Medicine, 3(4), pp. 293-304.
Morrison, R. S. and Siu, A. L. (2023) 'Hip fracture in adults: Epidemiology and medical
management', in: UpToDate, Post, TW (Ed), UpToDate, Waltham, MA (Accessed on August
7, 2023).
Munshi, M. (2023) 'Treatment of type 2 diabetes mellitus in the older patient', in: UpToDate, Post,
TW (Ed), UpToDate, Waltham, MA (Accessed on August 7, 2023).
Muth, C., Blom, J. W., Smith, S. M., Johnell, K., Gonzalez-Gonzalez, A. I., Nguyen, T. S., Brueckle,
M. S., Cesari, M., Tinetti, M. E. and Valderas, J. M. (2019) 'Evidence supporting the best
clinical management of patients with multimorbidity and polypharmacy: a systematic
guideline review and expert consensus', Journal of Internal Medicine, 285(3), pp. 272-288.
39
Neuman, J., Korenstein, D., Ross, J. S. and Keyhani, S. (2011) 'Prevalence of financial conflicts of
interest among panel members producing clinical practice guidelines in Canada and United
States: cross sectional study', BMJ, 343.
O'Sullivan, E. D. and Schofield, S. J. (2018) 'Cognitive Bias in Clinical Medicine', Journal of the
Royal College of Physicians of Edinburgh, 48(3), pp. 225-232.
Onder, G., Vetrano, D. L., Palmer, K., Trevisan, C., Amato, L., Berti, F., Campomori, A., Catalano,
L., Corsonello, A., Kruger, P., Medea, G., Nobili, A., Trifirò, G., Vecchi, S., Veronese, N.
and Marengoni, A. (2022) 'Italian guidelines on management of persons with multimorbidity
and polypharmacy', Aging Clinical and Experimental Research, 34(5), pp. 989-996.
Pellegrino, E. D. (1979) 'The Anatomy of Clinical Judgments: Some Notes on Right Reason and
Right Action', in H. T. Engelhardt, J., S. F. Spicker, and B. Towers (ed.) Clinical Judgment: A
Critical Appraisal Philosophy and Medicine. Boston: D. Reidel Publishing Company, pp.
169–194.
Polanyi, M. (1966) The Tacit Dimension. Garden City, New York: Doubleday & Company.
Primiano, I., Nicola, M., Monica, M., James, D., Paolo, C. and Antonino, C. (2016) 'Wrong
guidelines: why and how often they occur', Evidence Based Medicine, pp. ebmed-2016-
110606.
Rosenberg, W. and Donald, A. (1995) 'Evidence based medicine: an approach to clinical problem-
solving', BMJ, 310(6987), pp. 1122-1126.
Ross, E. G. U. and Shawn, T. (2008) 'Chronicity and complexity', Canadian Family Physician,
54(12), pp. 1655.
Rydén, L., Standl, E., Bartnik, M., Van den Berghe, G., Betteridge, J., de Boer, M.-J., Cosentino, F.,
Jönsson, B., Laakso, M., Malmberg, K., Priori, S., Östergren, J., Tuomilehto, J., Thrainsdottir,
I., Vanhorebeek, I., Stramba-Badiale, M., Lindgren, P., Qiao, Q., Priori, S. G., Blanc, J.-J.,
Budaj, A., Camm, J., Dean, V., Deckers, J., Dickstein, K., Lekakis, J., McGregor, K., Metra,
M., Morais, J., Osterspey, A., Tamargo, J., Zamorano, J. L., Deckers, J. W., Bertrand, M.,
Charbonnel, B., Erdmann, E., Ferrannini, E., Flyvbjerg, A., Gohlke, H., Juanatey, J. R. G.,
Graham, I., Monteiro, P. F., Parhofer, K., Pyörälä, K., Raz, I., Schernthaner, G., Volpe, M.,
Wood, D. and Authors/Task Force Members, O. C., ESC Committee for Practice Guidelines ,
Document Reviewers (2007) 'Guidelines on diabetes, pre-diabetes, and cardiovascular
diseases: executive summary: The Task Force on Diabetes and Cardiovascular Diseases of the
European Society of Cardiology (ESC) and of the European Association for the Study of
Diabetes (EASD)', European Heart Journal, 28(1), pp. 88-136.
Sackett, D. L., Rosenberg, W., Gray, M., Haynes, B. and Richardson, S. (1996) 'Evidence based
medicine: what it is and what it isn't', BMJ, 312(7023), pp. 71-72.
Salisbury, C., Johnson, L., Purdy, S., Valderas, J. M. and Montgomery, A. A. (2011) 'Epidemiology
and impact of multimorbidity in primary care: a retrospective cohort study', British Journal of
General Practice, 61(582), pp. e12-e21.
Shippee, N. D., Shah, N. D., May, C. R., Mair, F. S. and Montori, V. M. (2012) 'Cumulative
complexity: a functional, patient-centered model of patient complexity can improve research
and practice', Journal of Clinical Epidemiology, 65(10), pp. 1041-1051.
Silva, S. A. and Wyer, P. C. (2009) 'Where is the wisdom? II – Evidence-based medicine and the
epistemological crisis in clinical medicine. Exposition and commentary on Djulbegovic, B.,
40
Guyatt, G. H. & Ashcroft, R. E. (2009) Cancer Control, 16, 158–168', Journal of Evaluation
in Clinical Practice, 15(6), pp. 899-906.
Skou, S. T., Mair, F. S., Fortin, M., Guthrie, B., Nunes, B. P., Miranda, J. J., Boyd, C. M., Pati, S.,
Mtenga, S. and Smith, S. M. (2022) 'Multimorbidity', Nature Reviews Disease Primers, 8(1),
pp. 48.
Stegenga, J. (2016) 'Hollow Hunt for Harms', Perspectives on Science, 24(5), pp. 481-504.
Thornton, T. (2006) 'Tacit knowledge as the unifying factor in evidence based medicine and clinical
judgement', Philosophy, Ethics, and Humanities in Medicine, 1(1), pp. 2.
Tinetti, M. E., Bogardus, S. T. and Agostini, J. V. (2004) 'Potential Pitfalls of Disease-Specific
Guidelines for Patients with Multiple Conditions', New England Journal of Medicine,
351(27), pp. 2870-2874.
Tinetti, M. E., Fried, T. R. and Boyd, C. M. (2012) 'Designing Health Care for the Most Common
Chronic Condition—Multimorbidity', JAMA, 307(23), pp. 2493-2494.
Tonelli, M. (1998) 'The Philosophical Limits of Evidence-based Medicine', 73(12), pp. 1234-1240.
Tonelli, M. R. (2006) 'Integrating evidence into clinical practice: an alternative to evidence-based
approaches', Journal of Evaluation in Clinical Practice, 12(3), pp. 248-256.
Tonelli, M. R. (2011) 'Integrating clinical research into clinical decision making', Annali dell'Istituto
superiore di sanita, 47, pp. 26-30.
Tricoci, P., Allen, J. M., Kramer, J. M., Califf, R. M. and Smith, S. C. (2009) 'Scientific Evidence
Underlying the ACC/AHA Clinical Practice Guidelines', JAMA, 301(8), pp. 831-841.
Union Européene Des Médecins Spécialistes (2023), European Standards in Medical Training, Union
Européene Des Médecins Spécialistes, Brussels. Available at https://www.uems.eu/areas-of-
expertise/postgraduate-training/european-standards-in-medical-training (Accessed: 30 July
2023).
Uhlig, K., Leff, B., Kent, D., Dy, S., Brunnhuber, K., Burgers, J. S., Greenfield, S., Guyatt, G., High,
K., Leipzig, R., Mulrow, C., Schmader, K., Schunemann, H., Walter, L. C., Woodcock, J. and
Boyd, C. M. (2014) 'A Framework for Crafting Clinical Practice Guidelines that are Relevant
to the Care and Management of People with Multimorbidity', Journal of General Internal
Medicine, 29(4), pp. 670-679.
Upshur, R. (2005) 'Looking for rules in a world of exceptions: reflections on evidence-based practice',
Perspectives in biology and medicine, 48(4), pp. 477-489.
Upshur, R. E. (2014) 'Do clinical guidelines still make sense? No', Ann Fam Med, 12(3), pp. 202-3.
Weiss, C. O., Varadhan, R., Puhan, M. A., Vickers, A., Bandeen-Roche, K., Boyd, C. M. and Kent,
D. M. (2014) 'Multimorbidity and Evidence Generation', Journal of General Internal
Medicine, 29(4), pp. 653-660.
Wexler, D. J. (2023) 'Insulin therapy in type 2 diabetes mellitus', in: UpToDate, Post, TW (Ed),
UpToDate, Waltham, MA (Accessed on August 7, 2023).
Wieten, S. (2018) 'Expertise in evidence-based medicine: a tale of three models', Philos Ethics
Humanit Med, 13(1), pp. 2.
41
Østbye, T., Yarnall, K. S., Krause, K. M., Pollak, K. I., Gradison, M. and Michener, J. L. (2005) 'Is
there time for management of patients with chronic diseases in primary care?', Ann Fam Med,
3(3), pp. 209-14.
