AI Guidelines for Clinical Care

We consult professional healthcare when we have experiencees we cannot understand, tolerate or resolve. The clinician listens to our account, makes a physical examination and often orders further special clinical investigations. But which ones and in what order? Guidelines exist but are too complex for current clinicians. AI will navigate them with ease. Clinical investigations in future would be most effectively utilised if selected and sequenced by AI . The Electronic Health Record will comprehensively record the patient details, the outcomes of previous similar consultations and initially at least "suggest" useful next steps. 

End of an Era

An era is coming to an end. Bracketed by the Flexner Report of 1910 and the arrival of Artificial Intelligence in Healthcare, medicine depended daily on the doctor being well schooled in bioscience and charged with delivering good decisions based on rationale, existing agreed best practice, personal and communal experience and respecting the patient's preferences and the healthcare system's capabilities and limitations. At the outset this was as good as it might be. But it wasn't that good. Existing best practice was no more than a consensus view with little evidence to support it. Personal experience was seldom recorded and relied on the recall of the practitioner who had few numerical data to help and whose recall of outcomes would be partial and most likely selective. Rationale, rooted in a thorough grasp of best scientific knowledge seemed the key redource yet has subsequently been shown to be an unreliable source of effective guidance. The capabilities of the system to deliver were by modern standards extremely limited. The patient's preferences were of course available and could be respected but limited by the aforementioned fragile contributions. Now we have robust evidence on best actions in an increasing range of situations, with AI to carry and dispense the data and guide non science based practitioners.

Evidence and Psychotherapy

In 1980 Smith Glass and Miller published this book: The Benefits of Psychotherapy https://archive.org/search.php?query=external-identifier%3A%22urn%3Aoclc%3Arecord%3A1244213270%22

They were supportive despite recognising small effect sizes.

In 1983 Prioleau Murdock and Brody published: An analysis of psychotherapy versus placebo studies - a scrutiny of a subset of the former's data. They wrote "It was concluded that for real patients there is no evidence that the benefits of psychotherapy are greater than those of placebo treatment."

We are still engaging in this. More to benefit the therapists than the patients?

After Flexner

We have seen how Abraham Flexner structured his view of medical education's future predicated principally on  the idea that doctors possessed of a fair grounding in assorted scientific disciplines would on clinical occasions be best able to derive and deliver a useful response for the patient. That was an implementation of the principle of rationale, the notion that, having considered what we know, we believe that this is likely to be a useful response.  While it is an integral part of current medical practice it is incomplete.  The second stage that was addressed in the years following the  Flexner report is the quantitative measurement of our past activities. We need to  know or have a record of what clinical decisions were made and recorded and what their outcomes were. Thee task was then to construct the methodologies and implement them to review past performance to discover successes and failures and to use that as an additional dimension of future practice. The idea that we should rely only on the individual experience of the clinician was to be supplemented by the accumulated communal experience of past activity. This activity was known as clinical epidemiology and was growing as we added the third element.

Forthcoming Election in UK

Stirring times. Polls suggest major change in government. Yesterday the Daily Telegraph, a resoundingly Tory newspaper had portentous headlines. In England it proclaimed "Tory Wipeout" and in Scottish editions "SNP Wipeout". It carried two constituency maps of the UK with the constituencies coloured by current or poll projected post election party. The current map is largely blue for the Tories in England and Yellow for the SNP in Scotland. On the map based on poll projections of the outcome of the election on 4th July Both Blue and Yellow are much reduced in a large swathe of Red for Labour. We shall see.

A New Flexner

 In 1910 Abraham Flexner, who was not a doctor, published his massive report on the education of doctors for the Carnegie Foundation in response to a growing concern at poor quality doctors being produced.  This triggered major change and perhaps intuitively recognised the core routes for doctors to produce useful advice for patients. These are three in number and presented in order  of increasing efficacy (and, at the time, decreasing accessibility)

Rationale "What should be worthwhile"

Experience "What has seemed to be worthwhile in previous similar situations"

Evidence "What has on average been shown to be worthwhile in controlled studies of similar situations "

In 1910 a programme of education could only be  securely based on  the first of these.  There was no structure to gather and analyse and disseminate clinical experience and no established methods of clinical science to mount prospective inquiry.  

There was however a burgeoning biody of bioscientific findings such that suitably armed a doctor could construct a logical argument to support specific conclusions in a given clinical situation. Such conclusions had no guarantee of being correct but we're likely better than what had gone before. 

Flexners report was widely adopted and this model of clinical practiced dominated unchallenged until 1985.

Abiding Clinical Truths

Abiding Clinical Truths

Abiding truths there are in medicine and though few in number they may yet bring comfort to our embattled colleagues. We are in a time of change in medicine more profound perhaps than in any other part of life. Nothing like this has been seen before. Doctors are reasonably and increasingly uneasy and unhappy with current system dysfunctions but beyond that there is an unspoken fear of some looming threat that lurks unseen or if dimly perceived evokes widespread denial. Diagnoses there are aplenty but none consoles and leadership is absent. Prognoses, as ever uncertain and frequently doom-ladened, lead away from what we desire. This is natural. We are in the early stages of a millennial change in how we contribute to human well being yet armed with our abiding truths we shall comprehend and succeed in our honourable mission.

To map the future it can help to review the past. Our remote predecessors started from nothing and constructed a clinical system based on appeals to intuitive overarching principles and at times supernatural authority. It was however imbued with unspoken truths which still abide. With a significant incidence of self limiting illness, a modicum of luck and a pervasive fatalism, it sufficed for thousands of years. Progress was made with the growing adoption of empirical observation, understanding the basics of biology, and the application of the scientific method. Mainstream medical progress was to follow this path while alternative practitioners continued to appeal to principle, system and invisible influence. The two traditions survive today and both have a role.

The tenets of scientific medicine came to be implanted in the minds of high achieving individuals trained as bioscientists and sent into independent practice to make the best judgements they could. This model, formalised by Abraham Flexner in his report of 1910, though imperfect, has served as well as any that might have been proposed. Today’s practitioners were selected, educated and trained in this model, rejoicing in its strengths while ignoring its weaknesses. We were to do our best armed with what we had been told was established knowledge from the past, rationale applied to the clinical situation in the light of our scientific knowledge and a growing personal experience, tempered always by a deference to the patient’s preferences and a submission to the practical constraints of the system within which we practiced. Thus was the professional doctor, educated, independent, autonomous in selection of approach and in command of resources which for long enough were scarce, largely ineffectual yet mercifully inexpensive. These features, common to us all this last century or so do not however constitute abiding truths. They were the best fit for the times. It was a shaky achievement; reliable knowledge was scarce, the application of rationale yields many false conclusions and personal experience is not only patchy and fragmented but in the absence of searchable records is likely to be incompletely and selectively recalled.

The 20th century brought great advances which accelerated after 1945. Clinical practice costs grew as more effective responses to the patient’s distress became available. More complex investigations, biochemistry, pathology imaging and endoscopy fuelled more effective surgery, radiation and chemotherapy driving up costs. Patients responses to mishap became more demanding and their settlements more expensive. Funders started to take notice and tried to act. Their ability to effect change was stymied by the lack of any knowledge or agreement within the profession of how specific illnesses might be most effectively (or even most economically) treated. Healthcare organisations busied themselves with reorganisations and tinkered with financial management and infrastructure.

Change was coming. In 1983 the United States federal government introduced Diagnosis Related Groups to set Medicare payments for inpatient stays. This stepped away from funders unquestioningly paying for whatever the doctor ordered and offered incentives to hospitals to question the details of the clinical process. In 1990 the world changed in UK medicine with the transfer to the NHS of responsibility for claims of negligence by hospital doctors. Globally this would seem a suitable zero in the chronology of change. Organisations now had responsibility for increasingly expensive aspects of the outcomes of clinical practice and surely could not long accept that responsibility without commensurate authority. The bomb was ticking. The following year in the UK we had not only, responding to burgeoning consumerism, the Prime Minister’s “Citizen’s Charter”, which led on quickly to waiting time guarantees, but also the “New Deal” for junior doctors which though long delayed in the implementation offered the prospect of safer and more humane working hours over the strident protestations of much of the profession both senior and junior. The status quo has enduring appeal. Management was on the move asking unprecedented questions about the basics of clinical decision making, entering new and unfamiliar territory with new ambitions and expectations but still lacking the crucial ability to question the utility of individual clinical decisions since there were largely no agreed answers to any questions on the details of clinical practice. That year however a young Canadian doctor, Gordon Guyatt, delivered the keys to the kingdom of medicine with a short paper promoting the notion of “Evidence Based Medicine”. He has not been forgiven. To this day many column inches are given over to those hurt, confused and in denial of the changes that EBM has brought. His work (and that of his mentor David Sackett) symbolised by that initial paper was akin to the assassination of the Archduke at Sarajevo in 1914. It was not the cause of what followed but it was the trigger. The next 25 years saw tumultuous change. Yet this was only a prelude.

But what of abiding truths? Humankind has always experienced symptoms though the concept took time to be clear. Everyone has distressing experiences which suggest disordered function of the body or the mind. These are common almost universal and the majority are probably unknown to medicine. With the advent of professional therapists symptomatic individuals could seek help beyond their own resources or those of their immediate circle. Such a person became a patient. Patients come to the clinical encounter with two objectives; to gain relief from current distress and to receive reassuring advice on improving future prospects. To this end they have three tasks; to recount their experiences, to select from suggestions made by the clinician and to instruct the implementation of chosen interventions. The clinician can usefully also be conceived to have three tasks; to assess the situation, to make recommendations for action and to enact the patient’s instructions. That structure has been there from the flickering lamp of the cave shelter of our remote ancestors and is still there in the brightly lit technology rich clinic of today. It will survive for as long as mankind does and well beyond our survival as traditional medical practitioners. To supplement this structural arrangement we might note the patient’s objectives as to the manner of the encounter; to achieve relief and to be afforded assurance in an accessible, safe, effective, convenient, comforting, comfortable, confidential and economical manner. And these are the abiding truths of medicine.

But, in our distress, we insist that we are doing all that so what is the problem? Simply put, for our patients the current methods of delivery of care are unnecessary and overly expensive and for the traditionally educated and trained doctor they are increasingly trying and frustrating. The removal of clinical autonomy anticipated and realised is as burdensome to us as is its progression and inevitability. Newer types of clinician, drawn from more varied backgrounds, already supported by the products of EBM, and with clinically responsive information technology and ultimately artificial intelligence to come, can and increasingly will provide clinical care of a high standard. The current “Flexner” doctor in much smaller numbers will be more usefully and satisfyingly deployed for the greater part in the exploration and development of new diagnostic and therapeutic possibilities. Every patient to some degree will participate in clinical R&D. With robust electronic health records, accurately captured and summarised, the accumulation of the collective experience of healthcare organisations will become increasingly available to support clinicians. Behind the clinical endeavour sophisticated IT will extract and present usable information from the free text repository of both clinical practice and research and use it to construct more complex algorithms to guide clinical advice. Clinical care will move progressively from the varyingly effective, autonomous process of the past and present to structured evidence based care delivered by a much wider range of clinicians with varying skills, progressing to automated responses and ultimately fully fledged artificial intelligence.

Yet still we hear the cry, “Patients want more than this.” That is a view which is understandable yet in part misguided. It is true that individuals today spend significantly on alternative medicine which by any reasonable assessment is ineffective. We are told it brings a level of comfort absent in orthodox medicine but such claims are nebulous and unquantified. To unlock this quandary we need to step back for a moment and reflect that modern orthodox medicine has large areas where it also is ineffective. Bluntly put, medicine has made signal progress in areas where the patient’s illness can confidently be attributed to demonstrable pathology, to physico-chemical disorder, to disease. But much illness is not of this sort and orthodox medicine is largely powerless to help beyond a kind and attentive and at times perfunctory response. Alternative medicine shrewdly and yet intuitively recognises our weaknesses and the value of charisma and theatre in evoking gratitude and a sense of wellbeing, while silently accepting its shared limitations in effecting significant improvement. The best traditional doctors always had a double approach but with the advent of effective therapies we correctly emphasise these aspects and proceed. It is also harder to have an unambiguously healing persona when the core therapies bring short term distress often well before long term benefit. Surgeons before 1847 and the advent of anaesthesia were appreciated by society yet socially were held somewhat apart in recognition of imagined possible future relationships. Do modern oncologists evoke a similar ambivalent attitude? Psychiatrists even cause distress in their efforts to bring relief. Alternative medicine practitioners have no such barriers to universal approbation. As they have no upside apparent to us they have no downside apparent to their patients.

However that may be, our abiding truths will survive and will be ever present despite radical transformations in the methods of clinical practice. We are moving to a new model where our traditional scientific doctoring skills will be more focussed and in the hands of fewer individuals working with a wider range of clinicians all of whom will however continue to exhibit the core clinical skill of consistently and confidently making correct predictions. The patients will have more success in pursuit of their objectives and whether by the resolution of disease or a better understanding and response to existential dread, medicine will have moved on past the current uneasy compromises.