Late night thoughts of a clinical scientist.

by reestheskin on 30/07/2020

Comments are disabled

More accurately, late night thoughts from 26 years ago. I have no written record of my Edinburgh inaugural, but my Newcastle inaugural given in 1994 was edited and published by Bruce Charlton in the Northern Review. As I continue to sift through the detritus of a lifetime of work, I have just come across it. I haven’t looked at it for over 20 years, and it is interesting to reread it and muse over some of the (for me) familiar themes. There is plenty to criticise. I am not certain all the metaphors should survive, and I fear some examples I quote from out with my field may not be as sound as I imply. But it is a product of its time, a time when there was some unity of purpose in being a clinical academic, when teaching, research and praxis were of a piece. No more. Feinstein was right. It is probably for the best, but I couldn’t see this at the time.


Late night thoughts of a clinical scientist

The practice of medicine is made up of two elements. The first is an ability to identify with the patient: a sense of a common humanity, of compassion. The second is intellectual, and is based on an ethic that states you must make a clear judgement of what is at stake before acting. That, without a trace of deception, you must know the result of your actions. In Leo Szilard’s words, you must “recognise the connections of things and the laws and conduct of men so that you may know what you are doing”.

This is the ethic of science. William Gifford, the 19th century mathematician, described scientific thought as “the guide of action”: “that the truth at which it arrives is not that which we can ideally contemplate without error, but that which we may act upon without fear”.

Late last year when I was starting to think what I wanted to say in my inaugural lecture, the BBC Late Show devoted a few programmes to science. One of these concerned itself with medical practice and the opportunities offered by advances in medical science. On the one side. Professor Lewis Wolpert, a developmental biologist, and Dr Markus Pembrey, a clinical geneticist, described how they went about their work. How, they asked, can you decide whether novel treatments are appropriate for a patient except by a judgement based on your assessment of the patient’s wishes, and imperfect knowledge. Science always comes with confidence limits attached.

On the opposing side were two academic ethicists, including the barrister and former Reith Lecturer Professor Ian Kennedy. You may remember it was Kennedy in his Reith lectures who quoting Ivan Illicit described medicine itself as the biggest threat to people’s health. The debate, or at least the lack of it. clearly showed that we haven’t moved on very far from when C P Snow (in the year I was born) gave his Two cultures lecture. What do I mean by two cultures? Is it that people are not aware of the facts of science or new techniques?… It was recently reported in the journal Science that over half the graduates of Harvard University were unable to explain why it is warmer in summer than winter. A third of the British population still believe that the sun goes round the earth.

But, in a really crucial way, this absence of cultural knowledge is not nearly so depressing as the failure to understand the activity rather the artefacts of science. Kennedy in a memorable phrase described knowledge as a ‘tyranny’1. It is as though he wanted us back with Galen and Aristotle, safe in our dogma, our knowledge fossilised and therefore ethically safe and neutered. There is, however, with any practical knowledge always a sense of uncertainly. When you lift your foot off the ground you never quite know where it is going to come down. And, as in Alice in Wonderland, “it takes all the running you can do to stay in the same place”.

It is this relationship, between practice and knowledge and how if affects my subject that I want to talk about. And in turn, I shall talk about clinical teaching and diagnosis, research and the treatment of skin disease.


Reductionism and pragmatism

I came to this university in 1976. In those days, before the discovery of AIDS, the biggest threat to a medical student’s health was the multiple choice question. The reduction of human experience to True/False/Don’t know2, this ritual humiliation and corruption of the minds of the young, is something that I have had to accept from the medical Royal Colleges as well. Their excuse is that at least they make money. Yet it is an activity, that I have always found curiously disturbing for a University, I mean the Multiple choice exam, rather than making money …

Nevertheless, students come to University not just to worship what is known but to question it. And in my case, apart from allowing me to fail one or two class exams, the multiple choice question spurred me on to trying to develop an educated irreverence for my teachers. I discovered the work of Lewis Thomas, from whom I take the title of this lecture3, and also the work of Alvan Feinstein, currently professor of Medicine at Yale4. In his book Clinical Judgement, published in 1967, Feinstein described a clinician as somebody who “depends not on a knowledge of causes, mechanisms, or names for diseases but on a knowledge of patients”.

I have been intrigued by this statement ever since. What is Feinstein saying? What does he mean? That the essential elements of practice, the knowledge that in Clifford’s words is our ‘guide to action’ comes not from say, molecular biology or epidemiology, but from patients? Initially this seemed a strange idea. The view from the medical school was clear on this point. You learn about the atoms and molecules, then the cells, then the organs, stacking layer upon layer, finally adding disease, and possibly patients some time later if time allowed. Medicine was reducible to the molecules: if only you were clever enough5.

Nevertheless, one of the delights of the Newcastle curriculum was that you spent time early in your career with General Practitioners. It was nothing particular about going into the community, although I sadly note that this is now seen to be politically correct6. But one of the advantages of GPs, at least in those days, was that they were financially independent of the University. Salary, departmental structure, devolvement of budgets, these structures always induce a subtle, or not so subtle, conformity on one’s thinking. And it was obvious when you observed clinicians working or when you discussed their work with them, that the medical school model of medicine reducible to the molecules was wide of the mark, and that Feinstein’s judgement was not so wild after all.

The battle between medical educationalists on the one hand and rationality on the other is a long one. And much as the Church in a similar battle keeps moving its position — they wouldn’t make Galileo recant nowadays and so on — and however frustrating this is for my argument, things have moved on. In particular, the proponents of the other extreme have marshalled their forces. Now the threat to clinical medicine comes not only from the arch-reductionists. In the words of a recent leading article from the New England Journal of Medicine, the “structures of biological science are almost irrelevant for the practice of medicine”7. The new kind of pragmatic knowledge is, and I quote, “superior to the cause and effect reasoning of traditional medical science”.

The results of this nihilistic denial of natural science are seen all around us : in what (Petr) Skrabanek has labelled the epidemiology of errors — the idea that associations or ‘risk factorology’ without experiment or a biological model are useful; the audit bicycles we see leaning against our hospital walls; and in the extreme view the idea that giant databases obviate the tedium of having to think8. The pragmatic model works like this: you collect lots of information, lots of data, and somehow that tells you what is going on — what the sociologist Jerry Ravetz christened GIGO science — garbage in garbage out. Medical practice becomes a problem of operant conditioning. Like Skinner’s rats we (meaning the doctors) just need the right incentives.

These two themes — on the one hand the necessary reductionism of biological science (and make no bones I am a reductionist, even if it looks as though I have only just come out of the closet) and on the other hand the pragmatic argument that clinical medicine has no structure, is devoid of concepts or mechanisms based on biological science — these two themes have started to dominate the ideology of medicine. Both these views in one sense relegate if not subjugate the practice of clinical medicine.

But medicine is not like a pocket calculator, nor does it run programmes from any other discipline but its own. An analogy may help: lovers of the ongoing debate about whether pocket calculators are intelligent or conscious, will know the critique made by the Berkeley philosopher John Searle. His metaphor is delightful. Talking about logical machines, he says, imagine a non-Chinese speaker in a closed room who receives questions written in Chinese through a hole in the wall. Available in the room is a code which enables the Chinese symbols to be matched against the second set, which constitute replies to the questions being asked. The replies can be fed out through the hole in the wall to the outside world.

To the observers, however, it is clear that questions in Chinese have been answered appropriately in Chinese. So in any meaningful way can we consider the operator conscious or intelligent? Because if we do we are faced with all these delightful ethical problems such as : is it ethical to switch your pocket calculator off? Can you replace the battery of one machine with a different coloured battery. And so on…. Searle’s point is that in no way can one infer that the person in the room understood, or is conscious of, or is intelligently responding to the content of the messages. It is nothing more than a purely automatic operation. But what about medicine? Does the metaphor hold? Can we consider the doctors to be conscious or even intelligent? Or are they just logical operators on inflated salaries?


The nature of medical practice

What is diagnosis? Diagnosis is an attempt to order experience, an attempt to cleave nature at its joints. An attempt to reveal the underlying likenesses, to map out the structure of disease such that one can predict an outcome. In Clifford’s words, “that which provides a guide to action”. Dermatologists are collectors of rashes. The antiquarians of a nation of shopkeepers. Do we therefore imagine that the diseases are real, existing independently in their own right? That they are ‘entities’, a word loved by dermatologists? No, of course not. Diseases are concepts to order experience. Psoriasis is as real as gravity. Yes, the apple drops to the ground— yes, the rash exists. But the idea that links the apple with the planets and allows the concept of gravity is like the structure of diagnosis. Diagnosis is a metaphor for disease.

Dermatology was established as an academic discipline in the second half of the 19th century. Within a 20 year period around 1850, the first journal, the first society and the first chair of dermatology was established. If you ignore the post war period, and then exclude all but the Viennese, you would still be left with many if not most of the clinical descriptions that we now know. How did the subject develop so rapidly? By clinicians recording all that they could see with their eyes? Recording all the redness, itch and scale and soon. No. It would be as ludicrous to suggest this as to suggest that Newton went round the world measuring how quickly apples fell from the trees.

Diseases, syndromes, are characterised by their ability to explain and predict behaviour. Despite numerous descriptions of rashes neither Hippocrates or Galen described psoriasis. Why? Didn’t they see it? Given its prevalence I think that most unlikely. The rash will have been there in front of their eyes on many occasions. And it isn’t so much that they didn’t discover it, but that they didn’t invent it. How do you, in everyday practice get from the rash to the diagnosis? How do you get from this picture to the information somewhere in this medium-sized textbook of dermatology?

If you carry out an experiment putting about 20 objects on a tray, ask people to remember them, people can recall perhaps 18 a couple of hours later. But if you do similar but not quite identical experiments with human faces, it is clear that we can easily keep in our brains upwards of 10,000 images. I don’t want to suggest that the ability to recognise patterns in Dermatology is identical to that which we use to identify faces, but nevertheless I think the analogies are great. And of course the key here is meaning and context.

This is not just true of Dermatology. There is a widespread view that we are asking students to remember ‘too much’: we must reduce our clinical curriculum. It’s all about facts … And there appear to be too many of them and therefore we must tease out core facts and core curriculum and so on. I take the fairly conservative view that diagnosis is central to medicine and that it underlies the trade of a physician. And whereas I don’t doubt that the trivia that makes up many of our multiple choice questions may quickly saturate one’s storage capacity. I suspect our students are capable of achieving far more in terms of clinical diagnosis than we suspect. The crucial aspect is to remember that the act of diagnosis is not simply a listing of sentences: In clinical medicine apparent problems with factual overload are actually a reflection of the poverty of both concepts and context. Students are not blank pieces of paper.

I am making two points. Firstly that many of the skills required for medicine are procedural rather than declarative. And secondly, that the context is all important. Facts have become expropriated from the context of the clinic where they belong. In postgraduate medicine we see the same phenomena — people walking round with books full of lists and crammers. And yet what is intriguing is that if you ask competent clinicians those lists, they don’t know them. Nevertheless, these little educational viruses, ‘memes’ in Richard Dawkins’s definition, go on infecting one generation after another, one vector being the multiple choice paper.

But I have digressed a little. Let me return to diagnosis. How is the diagnosis made? … Simple. Listen to the Seminar Room on Ward 219 at the Royal Victoria Infirmary every Thursday or Friday morning: 15-20 dermatologists are quarrelling, and that’s exactly as it should be. If you go round thinking that diagnoses are fixed immutable objects rather than concepts, you are profoundly mistaken. If you look at a picture from von Hebra10 it is quite clear. These pictures are an attempt to organise activity. And if you read the old textbooks, if you follow the descriptions through, you find them constantly changing. Our concepts change. Psoriasis is not a satisfactory diagnosis. ‘Basal cell carcinoma’: this is not satisfactory either — it is the best we have – but the activity of diagnosis is to continually remodel diagnosis.

I have talked abut the reductionist approach. But the opposing voice, is now, I think, becoming dominant. And this is the view that you almost don’t need to do experiments, that the difficulty, the frustration, the close observation of cause and effect, is not necessary. In the words of the New England Journal, cause and effect, the reasoning traditional of medical science, is inferior to the statistical analysis of large databases.


Clinical Science

But human disease is not just any old biology. Yes, since we are all made of molecules, the disease must be explainable at the molecular level. The point, however, is that the scale and the nature of biological systems is such that to make accurate predictions from one level of explanation to another is fraught with difficulties. Furthermore, predicting interventions on the basis of a myriad of pathways has to be performed experimentally on humans. This is clinical science. Lewis Thomas pointed out confusion over the word ‘multifactorial’. This is a word loved by epidemiologists, and is all too often used by medics to justify ignorance. He points out that the most complicated multicell, multi-tissue and multiorgan diseases he knows were – syphilis, TB and pernicious anaemia – in each numerous organs and tissues are involved, and each appears to be affected by a variety of environmental influences. But before they came under scientific appraisal, each was thought to be multifactorial. The diseases seemed far too complex to allow for a single causative mechanism, yet when the necessary facts were in, it was found that just by switching off one thing, the spirochete that causes syphilis, the bacillus that causes TB, or Vitamin B12 that underlies pernicious anaemia, the whole array of disordered and unrelated mechanisms can be switched off. One of the major ways we learn to understand disease systems is by doctors’ undoubted ability to poison patients. The reason this approach is fruitful is partly because the observations are made at the correct level, i.e. not either the atom nor the population, but the actual patient; and partly because the drugs and therapy work at the ‘target species’ of all medical research, the human being. Consider, for example, the major therapies for psoriasis? Some of them are old, some of them are very new. Yet the story of their development is remarkably similar. The description of the pathways and molecular events underlying psoriasis have moved forward enormously in terms of description in the last 50 years. But textbooks still refer to it as a multifactorial disease. In Lewis Thomas’s words, we still don’t know the key mechanism, except for the clues that have been provided by therapy. Tar and dithranol were used in the last century. Tar used to be applied to anybody with skin disease, to anything that moved — or was stationary for that matter.

Dithranol was introduced to treat fungal infections. However, one of the hazards of over the counter medicines is that you allow the patient rather than the doctor to get the diagnosis wrong. When dithranol was mistakenly applied to psoriasis instead of ringworm the rash was noted to improve. The rationale for Ultraviolet radiation and PUVA was that patients and doctors noted that the rash improved in summer/or on holiday.

Am I saying this is just luck?… No. Absolutely not. The striking thing about these observations is that they are made by researchers. They are made by people looking for opportunities. It’s just that one of the relevant laboratories for human disease is the clinic. Secondly, do they rely on basic science?… Yes of course they do. The arguments, the phrases, the thinking I have used to describe these compounds are heavily dependent on basic biological science, although perhaps even more dependent on chemistry.

Do I think this approach is confined to Dermatology? … No. Sir Colin Dollery, in his Rock Carling Lecture in 1978 speaking about this very matter pointed out that we have been constrained by the editors of journals and philosophers of scientific discovery to present things in a particular way, when we all know that this is not how it works in practice. So how come useful things came out of ideas that went wrong? The answer is a cliché but true: all knowledge is provisional and ultimately flawed, but experiment — especially when you use the target organism of medical research — provides a guide to action.



I want to draw the strands of my argument together. I have argued that there are at least two parts to the morality that underlies medicine. Firstly the sense of compassion — the projection of and identification with the patient; and secondly the ability accurately to predict the results of one’s action, without a trace of deception.

I suggest that diagnosis is not a simple listing of disease, a photographic image, but an activity that orders our experience; concepts that make binding the physical signs of disease with outcome. Acquiring diagnostic skills is not a matter of learning lists, but rather rests in the recognition of shape and form; much of learning is procedural rather than declarative, and context is all important.

I have suggested that clinical practice on the one hand is not reducible to molecular biology, nor can I tolerate the absurdity of the opposite belief that somehow one should ignore the shape and form that biology must impose. Clinical science provides the major causal pathways that allow us to act against disease. The concepts we use however, the very language I have used in this lecture, are of course those of experimental science; it cannot be otherwise. Science is reductionist in methodology, but not conception.

The limits to knowledge, and therefore our actions, are the confidence limits attached to our predictions. It will come as no surprise that I have argued for the central role of clinical science. This reflects the curious biology that is human disease, and the experimental constraints that are imposed by human pathology11.


  1. Echoes of we don’t need experts?
  2. The days of negative marking are now over.
  3. Lewis Thomas, wrote a number of books that influenced me. One was simply titled Late Night Thoughts with a moving last chapter Late Night Thoughts on Listening to Mahler’s Ninth. Wikipedia describes him as an ‘American physician, poet, etymologist, essayist, administrator, educator, policy advisor, and researcher’.
  4. Feinstein died in 2001.
  5. Reductionism: The late physicist John Ziman quotes John Tyndall believing that it would be possible to predict Hamlet from a knowledge of the forces between the atoms in a mutton chop.
  6. Newcastle moved teaching into what was then called General Practice and now is referred to as primary care ahead of many medical schools. I liked it. The Scottish government issues instructions to medical schools about how much teaching has to take place in primary care. People tend to keep their head down when the state is dictating how and what is taught: history suggests trouble lies ahead. James McCormick, a distinguished Professor at Trinity College, Dublin, debunked much lazy thinking on this topic in the book by Downie and Charlton, The Making of a Doctor: Medical Education in Theory and Practice OUP 1992. I doubt if anybody in the Scottish Government has read it. I forget how old the ‘PC’ term is.
  7. I am afraid I have not searched for the citation behind this statement.
  8. I forget how old this delusion is. As Alan Kay said, we don’t need big data as much as we need big ideas.
  9. The dermatology seminar room was at the end of ward 21 in the old part of the RVI, Newcastle.
  10. Ferdinand Von Hebra (1816-1880), Professor of Dermatology, University of Vienna, and author of the Atlas der Hautkrankeiten
  11. This article (with a few minor changes for this posting) was adapted and edited from my inaugural lecture in January 1994 by Bruce G Charlton and published in the Northern Review in 1995 | 1 | 41-50.