UNEASY BEDFELLOWS? - reconciling intuition and evidence based practice
Trisha Greehhalgh says it is high time to call a halt to the false dichotomy between clinical intuition and the demands of an evidence based approach, by raising the status of intuition as a component of expert decision making
During a busy GP surgery recently, I saw a 31-year-old man who wanted a note for the social security department to confirm his address. He had a somewhat abrupt and surly manner, and initially refused to sit down, but he was not overtly aggressive. He had seen a doctor once in the last five years, for sinusitis. His previous contact before that was at university. I asked what he was trained in. Computer programming. Why no job, I asked? Fed up with working. Any partner? Not any more.
I encouraged him to sit down. He was a good-looking man, who appeared in a hurry and didn't make much eye contact. He struck me as cynical and somewhat preoccupied. He told me that his partner had told him to get the hell out, find a job and sort himself out, and then she would think about taking him back. They had an 8-year-old daughter. I asked about depressive symptoms and he admitted to low mood, tearfulness and disturbed sleep. He admitted occasional suicidal thoughts but denied specific plans.
He didn't want to fill in a questionnaire; he didn't want to see our counsellor; he didn't want tablets; and he didn't want to talk about it any more now. I invited him to come back next week and talk some more. I wrote in the notes: 'See next week and chase if fails to attend. I am concerned about this man.'
The next week, at exactly the time he had booked to see me, he was found hanging from a beam in a disused warehouse.
This story raises a number of questions about the appropriate clinical management of the patient, but I include it here to show that intuitive insights are commonplace in general practice, and they may or may not save lives. They are rarely as impressive as the one I first heard quoted by Professor Nigel Stott (and which I subsequently analysed in detail1) from a GP in Cardiff: 'I got a call from a lady saying her three year old daughter had had diarrhoea and was behaving strangely. I knew the family well, and was sufficiently concerned to break off my morning surgery and visit immediately.'
FEATURES OF INTUITION |
|
This GP's hunch led him to diagnose correctly, and treat successfully, a case of meningococcal meningitis on the basis of two non-specific symptoms reported over the phone - an estimated 'hit rate' for that particular GP of one in 96,000 consultations, and a veritable tour de force for clinical intuition. The intuitive judgements we make on a daily basis in clinical practice are generally less dramatic but no easier to explain on a rational level. So, what is intuition?
Few doctors dispute that intuition plays a part in their practice, but there has been relatively little formal research into how (and to what extent) intuition contributes to decision-making in the clinical setting. The term evades a precise definition, but it is probably useful to describe its key features (see box).
Mystery writer Sir Arthur Conan Doyle, a medical doctor himself, likened the work of a detective solving a crime to that of a doctor diagnosing a mystery illness from subtle physical signs. Indeed, Conan Doyle is said to have based the character of Holmes on one of his own medical teachers at Edinburgh. In The Memoirs of Sherlock Holmes, Holmes states: 'It is of the highest importance in the art of detection to be able to recognise out of a number of facts which are incidental and which are vital...'2
In one of the first Sherlock Holmes novels, the detective is asked to explain a particularly impressive and obscure feat of reasoning. He responds as follows: 'From long habit the train of thoughts ran so swiftly through my mind that I arrived at the conclusion without being conscious of intermediate steps' (my italics).
Educationalists Hubert and Stuart Dreyfus, writing about intuition in industrial engineering, concluded that, 'Experienced intuitive [practitioners] do not attempt to understand familiar problems and opportunities using calculative rationality...When things are proceeding normally, experts don't solve problems and don't make decisions: they do what normally works' (my italics).3 Like Conan Doyle, Dreyfus and Dreyfus present intuition as a method of problem solving that marks the expert out from the novice, and they acknowledge the elusive nature of the intuitive method. Experts themselves can rarely provide an immediate, rational explanation for why they behaved in a particular way.
TABLE 1: FOUR CLERKINGS |
|
Examples of clinical clerking styles of health professionals at different stages of training |
|
3RD YEAR MEDICAL STUDENT"Mr Brown is a 38 year old computer operator who attended the Accident and Emergency department with a bad feeling in his eye. The history of the presenting complaint was that it was there when he woke up at 7.15 am on Wednesday morning. When he was a little boy he had an operation on his eyes for squint. He is up to date on his jabs....." |
5TH YEAR MEDICAL STUDENT"This 38 year old male attended with a feeling of grit in his right eye. The eye also had a yellow discharge. He could still read the paper with that eye. He had not had any previous episode like this. His visual acuity was 6/6 bilaterally. His pupils were equal, concentric responding to light and accomodation...." |
CASUALTY OFFICER "38 year old male |
GP"Rt conjunctivitis Chloramphenicol drops See S.O.S." |
Table 1, which is based on real life observations, shows excerpts from four clinical 'clerkings' of a single patient. The different problem-solving approaches adopted by clinicians at varying stages of training illustrate stages in the classification that Dreyfus and Dreyfus derived from observations of professional engineers. According to them:
The novice practitioner is characterised by:
-
Rigid adherence to taught rules or plans
-
Little situational perception
-
No discretionary judgement
The competent practitioner
-
Is able to cope with 'crowdedness' and pressure
-
Sees actions partly in terms of long term goals or wider conceptual framework
-
Follows standardised and routinised procedures
The expert practitioner
-
No longer relies explicitly on rules, guidelines and maxims
-
Has an intuitive grasp of situations based on deep, tacit understanding
-
Uses analytic (deductive) approaches only in novel situations or when problems occur
In general, we are at our most intuitive when doing our regular job and dealing with patients whom we know well. In unfamiliar situations, we resort to a more formal and rational approach based on explicit (and defensible) professional rules. The skill of the expert is to respond to the subtle cues that signal a need to shift between the two approaches. The GP who writes the single word 'conjunctivitis' in a medical record may live to regret it when the patient subsequently sues for a missed diagnosis of uveitis.
Objective science or intellectual imperialism?
Evidence-based medicine, which rose to prominence in the late 1990s, was based on principles developed by Sackett and colleagues and listed in the box on p25. Given their clear commitment to raising standards of clinical care, it is hard to understand precisely why the evidence-based medicine movement has drawn both passionate criticism4,5,6,7 and undisguised mirth 8 from so many sources. Clinicians in particular have claimed that the patient's unique and complex predicament is poorly served by applying a recommendation derived, however objectively, from an average result in a select population sample. As Goodman puts it, just because the average woman in the UK is a size 16 does not mean that all women should wear that size clothes. 5
Other critics have challenged the primacy accorded by evidence-based medicine to the randomised controlled trial for assessing the efficacy of interventions, and to epidemiological data for predicting the course of illness. They claim that decision makers in both these fields should also look systematically and rigorously at the evidence from basic science that underpins physiological and pharmacological processes10, economic evidence (about the costs of interventions and the values through which society allocates limited resources for health care11,12), philosophical and ethical evidence (about the nature of clinical method and the priorities of good clinical practice13,14), psychological evidence (about the patient's values and preferences and how to establish them15,16), contextual evidence (about the family and social systems in which any individual's illness experience and health goals are embedded17.18), and organisational evidence (about the professional and administrative systems within which health interventions are developed, evaluated and delivered19).
The contribution of evidence-based medicine to improved patient outcomes in general practice is incontestable. The old-fashioned family doctor, for example, was frequently to be heard reassuring patients that although their blood pressure was 'a bit high', it was 'nothing to worry about at your age and certainly not worth treating'. Large, prospective randomised controlled trials have shown precisely the opposite - that even the old elderly have substantial reduction in their risk of stroke if blood pressure is treated to below the target level of <160/90.20 Differences between groups are large and consistent throughout many trials. It has, furthermore, been estimated that finding and treating moderate hypertension in older people will save more lives than any other single intervention in primary care.21
Of course, we all know that certain elderly people would be better off not taking the standard recommended tablets for their blood pressure, perhaps because they have expressed a preference for non-drug measures or because they suffer from relevant comorbidity. 'Breaking the rules' in individual cases is entirely justified. Despite claims to the contrary by their critics, the founding fathers of evidence-based medicine never claimed that evidence from clinical research trials should be sought or applied in a vacuum.
Indeed, in one of the most widely cited paragraphs ever published in the British Medical Journal, they state: 'The practice of evidence-based medicine means integrating individual clinical expertise with the best available external clinical evidence... By individual clinical expertise, we mean the proficiency and judgement that individual clinicians acquire through clinical experience and clinical practice'. 22
The emotive nature of the ongoing debate is helped neither by both sides' sanctimonious claim to the moral high ground, nor by the inconsistent way in which writers use the term 'evidence'. In some publications, 'evidence' has a narrow definition related to the results of rigorous clinical trials and observational studies (as in 'the hierarchy of evidence'9). But in others, the term is used to refer to any factor that can and should influence clinical decision-making - including costs, patient preferences, ethical considerations, and socio-political context 23,24.
Another source of confusion is the gap between what practitioners of evidence-based medicine say and what they actually do. For example, the spate of research trials inspired by Sackett that classified clinical decisions as 'evidence-based' or 'non evidence-based' on the basis of whether the medication prescribed in each case had ever been the subject of a randomised controlled trial,25,26 were examples of the very 'cookbook medicine' that these authors had previously dismissed as naive and reductionist.22,27 An article that purported to be a systematic review of, and response to, arguments against evidence-based medicine remained within this limited conceptual framework and failed either to acknowledge or address more fundamental epistemological criticisms.28
Other criticisms notwithstanding, if evidence-based medicine really values 'clinical experience and judgement', its leading proponents surely have a responsibility to take that concept out of its black box and define precisely how it interacts with evidence in the decision-making process. What follows is an attempt to do just that.
The nature of clinical decision making
There are three widely held myths about clinical decision-making - first, that it is an entirely logical and deductive process; second, that experts think more logically than novices; and third, that more knowledge leads to better decisions.
PRINCIPLES OF EVIDENCE BASED MEDICINE |
|
The research literature on clinical decision-making tells us otherwise. Firstly, the critical importance of experience, context and familiarity have been persuasively demonstrated by Kathryn Montgomery Hunter, a professor of literature who spent several years watching, and listening to, doctors going about their duties. As her detailed fieldwork showed, clinical decision-making occurs by the selective application of general rules to particular individuals and contexts.29 The uniqueness of the individual (comorbidity, values, context, and the physiological idiosyncrasies that give rise to murmurs in the normal heart and make one person's pain another's 'tingling' or 'pressure') preclude any purely rule-based method for assigning diagnoses or selecting treatments.30
Hunter concluded: 'Clinical education is preparation for practical, ethical action: what best to do, how to behave, how to discover enough to warrant taking action, which choice to make on behalf of the patient. [These] choices are governed not by hard and fast rules but by competing maxims...As lawyers, literary critics, historians and other students of evidence know well, there is no text that is self-interpreting. As rules, these maxims are relentlessly contextual.' 29
Secondly, studies on the development of expertise in clinicians confirm the Dreyfus' taxonomy of problem solving: the more experienced a clinician gets, the less logical their decision-making processes are shown to be.31
Thirdly, there is a striking absence of studies showing that knowledge per se improves decision-making. It is well documented that merely having the knowledge at their fingertips in the form of guidelines generally makes no perceptible impact on clinicians' behaviour.32,33 For significant and sustained change in clinical behaviour, knowledge of best evidence must be combined with a change in attitude and motivation; structural barriers (such as availability and affordability of a new treatment) must be overcome; and the new clinical behaviour must be both prompted and reinforced.33,34,35,36
Even in something as clinically clear-cut as the management of hypertension, the naive application of 'evidence' without regard to the unique predicament and priorities of the individual patient soon makes the decision evidence-burdened rather than evidence-based. 37
In the words of Sir John Grimley Evans:7 'A clinician might be forced by guidelines to make use of evidence that is only doubtfully relevant, generated perhaps in a different grouping of patients in another country at some other time and using a similar but not identical treatment.... [This is] to use evidence in the manner of the fabled drunkard who searched under the street lamp for his door key because that is where the light was, even though he had dropped the key somewhere else.'
We are at our most intuitive when doing our regular job and dealing with patients whom we know well. In unfamiliar situations, we resort to a more formal and rational approach
A number of studies by educationalists have begun to throw light on the process by which clinical expertise accumulates. We start by learning detailed 'rules' about the cause, course and treatment of each condition. As we gain knowledge we convert these rules to stereotypical stories ('scripts'). We refine our knowledge by accumulating atypical and alternative stories via experience and the oral tradition (such as grand rounds, 'corridor consultations', and so on). Furthermore, there is growing evidence that clinical knowledge is stored in our memory as stories rather than as structured collections of abstracted facts.29,38,39,40
A 'zero-sum' equation?
Over the past five years, I have given around one hundred seminars, workshops and lectures to different groups of clinicians and academics on the general theme of evidence based practice. I rarely escape without being asked to take sides in the perceived adversarial relationship between those who view themselves as experienced clinicians - proud to be practising 'old-fashioned' medicine based more or less on intuition - and those (generally younger, computer-literate and possessed of formal postgraduate qualifications) who support the rational, explicit and systematic use of research evidence in the clinical encounter and who seek to 'convert' their colleagues to this approach.
This polarity of allegiance has generated another popular myth among clinicians - the notion of a 'zero-sum' relationship (i.e. more of one implies less of the other) between the deductive steps of evidence based medicine ('the science') and the subjective interpretation of the patient's story ('the art') in clinical encounters. This was probably never explicitly argued in print - though some have come close.41.42
Advocates of a humanistic, intuitive approach to clinical care appear to assume that at a philosophical level, it is impossible to integrate the 'science' of evidence-based medicine with the intuitive 'art' of clinical judgement - and that while the rules of evidence-based medicine can be taught, clinical intuition is an unfathomable phenomenon that simply 'happens'. The false dichotomy between evidence based medicine and clinical intuition, with the former defined as the 'scientific' element, has no sound theoretical basis.
Back in 1958, Sir Peter Medawar argued: 'Clinicians who discourse upon the 'spirit of medicine' will always point out that, while there is a large and profoundly important scientific element in the practice of medicine, there is also an indefinable artistry, an imaginative insight, and medicine (they will tell us) is born of a marriage between the two. But then the speaker spoils everything by getting the bride and groom confused.'43
It is well documented that merely having the knowledge at their fingertips in the form of guidelines generally makes no perceptible impact on clinicians' behaviour
The bride is Medawar's metaphor for the imaginative, intuitive ('female') approach to problems; the groom corresponds to the objective, deductive rationality conventionally associated with the (male) laboratory scientist. But such a polarity, suggested Medawar, misrepresents the nature of the scientific method. He argued that scientific reasoning is an exploratory dialogue that can always be resolved into two voices: imaginative (hypothesis-generating) and critical (hypothesis-testing).
He deplored the popular image of the white-coated scientist doggedly doing experiments to contribute fact after fact to the unproblematic, unidirectional march of scientific progress. 'The initiative for scientific action,' he wrote, 'comes not from the apprehension of facts but from an imaginative preconception of what might be true' - in other words, it is the 'bride', not the 'groom' that drives progress in science.
As anyone who has tried to generate research ideas knows, a creative imagination is indeed an essential precursor for a worthwhile research study. But good research ideas cannot be picked out of the sky. Medawar's contention that the possibility of truth distinguishes the scientific imagination from the fanciful 43 echoes Kant's declaration that, 'It must certainly be true of every hypothesis that it could possibly be true.'44
The importance of the patient's unique story is no longer in doubt. According to Macnaughton, the patient's narrative, interpreted by the skilled clinician, provides relevance and context for the illness and offers insights into their beliefs, priorities, and life choices. 40 But Medawar's axiom, derived from Kant's philosophical writings, suggests that the illness narrative contributes to the science as well as the art of medicine. The unique, individual, contextual, interpreted story provides insights into what could or might happen and as such is the raw material from which plausible hypotheses about causation, prognosis and therapy are generated before being tested against research-derived evidence.
Towards a science of intuition
Let us return to the anecdote with which I began this essay and see where the scientific imagination meets the patient's personal narrative. My subconscious hypotheses about what might happen incorporated both generalisable, research-derived truths (such as the known risk factors for suicide in young men, which include poor social circumstances, relationship breakdown, depression and what is termed 'cynical hostility') and unique, contextual ones (including the lack of any previous consultations, recent eviction by his long term partner, and his refusal to be drawn into formal mental state examination).
When I anticipated his impending death (or at least, recorded serious concerns in the notes), I was not consciously aware of the intermediate steps that led me to my hypothesis. But when I learnt the outcome and sought a debriefing with a colleague who also knew him, the pieces of the jigsaw were revealed to both of us.
The work of Benner 45 and Eraut 38 (among others) suggests that the insight I gained from critical reflection and discussion with a professional colleague is to be expected. Reflecting retrospectively on the process of clinical intuition (asking, for example, 'Why did I make diagnosis X rather than diagnosis Y at that point?' or 'What prompted me to start/stop that drug?') is a powerful educational tool.
In particular, critical reflection on past intuitive judgements highlights areas of ambiguity in complex decision-making, sharpens perceptual awareness, exposes the role of emotions in driving 'hunches' (perhaps also demonstrating the fallibility of relying on feelings alone), encourages a holistic view of the patient's predicament, identifies specific educational needs, and may serve to 'kick-start' a more analytical chain of thought on particular problems.
Educational methods aimed at supporting evidence-based practice have been systematically explored and evaluated,46 and now feature prominently on the curricula of medical schools, Royal Colleges and service organisations such as primary care trusts 47. But educational methods for improving clinical intuition are less well-developed, less easy to evaluate in behavioural terms, and generally less credible with decision-makers and policymakers in the health care sector (although there is an emerging literature on their use in management education 48).
Yet, as the enduring tradition of 'Balint groups' indicates, storytelling and reflective discussion in groups is a time-honoured method for professional development (though they tend to occur outside of formal educational settings and are rarely 'PGEA approved'). 38,49, 50 Surely it is time to acknowledge, and take steps to overcome, the false dichotomy between the science and art of clinical practice as taught in undergraduate and postgraduate settings? Having quite rightly placed the principles and methods of evidence-based practice on the mainstream educational agenda, it is now time to raise the status of intuition as a component of expert decision-making, and begin to integrate both group discussion methods 51 and individual reflective writing 52 alongside the teaching of these skills.
The false dichotomy between evidence based medicine and clinical intuition, with the former defined as the 'scientific' element, has no sound theoretical basis
Once we have recognised that none of us ever needed to choose between evidence-based medicine and old-fashioned clinical intuition, the stage is surely set for developing an educational method that draws productively on both traditions. Far from being a contradiction in terms, the evidence-based Balint group is surely the epistemological marriage we have all been waiting for.