Georg Ivanovas From Autism to Humanism - systems theory in medicine
2.1 The medical paradigm
Another inconsistency in the logical context of today’s medical thinking is that descriptions are regarded as truth. In epistemology this has been described as confusing the map and the territory (chap. 3.8). Descriptions of a disease according to certain criteria (for example pathophysiology) are but maps for an unknown territory (sick patient) established according to a legend (basic theory). This process of mapping is necessarily insufficient as a map is never able to represent the complexity of a territory. That is, a diagnosis is often not appropriate to describe what the physician really encounters. Actually, many a practitioner is happy, if (s)he sometimes sees a patient resembling the definitions. The specialist sees more typical cases, but also more where the concepts do not seem to fit either. This shall be demonstrated with arthritis.
“Anyone who has seen half a dozen examples of common lupus and lupus erythematosus is able with ease to distinguish one from the other…, but let him wait awhile and see more, and he will find before long that there are examples of mixed forms of the disease which it is impossible to denote correctly without employing hybrid names.”
Jonathan Hutchinson, F.R.C.S.British Journal May 1880 (Kelly et la: 1115)
What is true for lupus erythematosous, one of the main forms of arthritis, is true for all kinds of arthritis. The most common form is rheumatoid arthritis (RA) which is today diagnosed according to the American Rheumatism Association (ARA criteria). There are older and the newer criteria:
older ARA criteria for Rheumatoid Arthritis
ARA criteria for rheumatoid Arthritis 1987
Although such criteria are helpful in establishing a diagnosis their value should not be overestimated. It is easy to diagnose the disease if many criteria are present. But then no criteria are necessary. In early stages, where the diagnosis and the differential diagnosis to other forms of arthritis is often difficult, the older criteria spoke of ‘probable’ RA. Newer criteria either diagnose RA or not. But they do not exclude RA if only a few criteria are present.
From a theoretical point of view this is an interesting situation: RA in an early stage is difficult to be distinguished from other forms of arthritis and has also a good tendency for a total remission. The question is what we understand under ‘having RA’. Is it possible to have a disease without a clear diagnosis? Is it an undetected disease? Or is it no disease yet? Actually, every answer depends on the criteria of the observer and “one of the most significant determinants of prognosis in RA may be which criteria….are used to establish diagnosis” (Kelly et al: 917).
This lack of security on a large scale is typical for the medical profession. The problem arises, when our models are taken for real, when the map is taken for the territory. The distinction disease no disease is only a guideline and creates no truth. Of course, such logical subtleties are of no importance in an outbreak of cholera in a refugee camp. But they have a lot of consequences under other conditions.
In arthritis things are complicated by the fact that each of the ‘diseases’ RA, systemic lupus erythematosous and sclerodermia have a different pathomechanisms and their own criteria to diagnose them, but they can only to be diagnosed and distinguished from each other when they are clearly expressed. Furthermore, there are so called overlap syndromes, containing characteristics of two or three forms. One of these overlap syndromes is today called mixed connective tissue disease (MCTD) diagnosed according to clinical manifestation and biochemical findings. But, again, such a distinction is more a rule of thumb than a logical category.
Of course, there have been attempts to define rheumatic diseases through pathophysiological mechanisms, by certain antibodies, rheuma factors, etc. But no attempt leads to clear results. Genetics suggest that different genes play a role in the development of rheumatic disease (Morel 2004). One gene called RUNX1 involved in the development of blood cells and playing a regulatory role in the immune system, has been connected to different forms of arthritis such as RA (Tokuhiro et al 2003), psoriatic arthritis (Helms et al 2003) and lupus erythematosus (Prokunina et al 2003). The same ‘causal relation’ between a gene and RA and Lupus seems to be true for the STAT4 gene (Remmers et al 2007). Another gene (MHC2TA) is connected not only with RA but also with multiple sclerosis and myocardial infarction (Swanberg et al 2005). So even on a genetical level rheumatic disease seems to be a manifold expression of a basic process.
It is known that the imbalance of Th1 and Th2 helper cells play a crucial role in the genesis of rheumatic diseases. A type1 imbalance with a predominance of Th1 cells corresponds with RA and a type 2 imbalance with a predominance of Th2 cells corresponds to systemic lupus erythematosus (Kreutzfeldt/Müller 2001). That is, the different types of imbalance might lead sometimes to different, sometimes to similar clinical expression. In this sense arthritis would just be a certain morphological pattern of a more general imbalance, a network pathology (chap. 6.2, 6.7).
“Diagnosis in the rheumatic disease has been overemphasized. On the one hand, current classifications contain patients with features of several diagnostic entities, often termed ‘overlap’ patients, with no absolute diagnosis possible. On the other hand, patients with the same diagnosis often should managed very differently. Modern management individualizes therapy with diagnostic categories, based upon subgroups of patients with differing prognoses and therapeutic requirements……Management in rheumatic disease is more closely linked to the underlying pathophysiological process than to the specific entity….Even such a basic concept as ‘inflammation’ has different therapeutic implications, depending upon underlying pathophysiology” (Kelly et al: 361).
That is, the classical approach of diagnosing is a restricted tool in order to develop an individual strategy which has to take many factors into account.
The American Rheumatic Association does therefore not provide a definite map for a definite diseases. It standardizes only one way to classify morphological and pathophysiological changes. To some extend this is true for most of our diagnoses.
A diagnosis might be of doubtful value in the daily experience of the physician. He is always trapped between two extremes. One extreme is to understand the patient in all of his individual expressions, the other is to come to a clear and precise conclusion. This problem became more obvious since physicians in many countries are obliged to define their diagnosis according to the International Classification of Diseases (ICD) key. Working with the ICD key, it is often difficult. Sometimes it is even impossible to assign a patient to a given diagnosis.
The new key, the ICD-10, currently introduced in most industrialized countries, is said to solve this and other disadvantages. The new key shall provide a greater specifity. It also abandoned the notion of disease and replaced it through ‘disorder’. Through a phenomenological nature it would have had the advantage to be closer to the clinical practice. But ICD-10 tries the impossible. It strives for an operational diagnosis with a phenomenological description. This is a logical impossibility or it requires a polyvalent logic (chap. 3.5) with a totally different structure of thinking. It is impossible to be specific and relational at the same time. The dilemma is not solvable in the given epistemological frame. So ICD-10 does not hold what it promises (Küchenhoff 2001). It cannot. Therefore it treads the usual paths by counting criteria, something that is neither specific (as it does not refer to pathophysiological mechanisms) nor relational.
The crux in establishing a clear-cut diagnosis can be seen with the diagnostic manuals of psychiatric diseases. It underwent the change from the extreme of an unspecific description (not able to provide comparable diagnosis) to the extreme of a clearly classifying system. After WWII the catalogue of American Veteran’s Administration (the forerunner) listed 26 mental disorders. Too few to describe all the different mental alterations seen in practice. Its follower the Diagnostic and Statistical Manual of Mental Disorders (DSM) listed more diseases with every new edition. But still before the third edition, DSM III, 1980, “a patient categorized as schizophrenic in New York would take on a diagnosis of affective disorder upon flying to London” (Helmuth 2003).
This has changed totally with DSM IV which lists 395 disorders with decisive, countable criteria. Everywhere in the world it is now possible to come to the same diagnosis for the same patient. And all diseases are clearly distinguished. “Anxiety and depression had become as different as chalk and cheese” (Shorter/Tyrer 2003).
But, as it could be expected, this did not solve the problem. The map might be very strict. The patient isn’t. Therefore it is no wonder that 20 to 50% of patients fall short of DSM IV criteria and get thrown into the nos (not otherwise specified) category (Hellmuth 2003). This means that these people are classified as ill, but cannot be diagnosed correctly. According to the principles of evidence based medicine there is no evidence at all for their treatment. This is hardly acceptable in a strict scientific thinking.
But the price for decisive diagnostic criteria is even higher. Adding up all the epidemiological occurrences for mental disorders according to DSM IV, in Germany about 58% of the population is suffering of a disturbance of personality (Blech 2003). Nearly nobody is left as healthy. Everybody has an abnormality, making it difficult to find “the last well person” (Hadler 2004). In fact, "medical research has made such progress, that there are practically no healthy people any more" as Aldous Huxley shall have said.
The reason is probably not only that half of the authors of DSM IV had a commercial conflict of interests (Cosgrove et al 2006). It has also to do with an incertitude about the nature of a diagnosis (Ghaemi 2009). For the moment it is sufficient to state that the DSM and the ICD-key just standardize the use of certain terms according to criteria, given by committees. They do not create a reality (1).
The ‘chronic fatigue syndrome’ might serve as an example for the difficulty with the realty of a diagnosis. We all have seen such cases with chronic fatigue, most of us more than of tuberculosis. It is futile to discuss, whether a disease really exists (Raine et al 2004), after some criteria are defined. Of course it exists, as it has been diagnosed according to criteria. It does not matter that there are no pathophysiological ‘causes’ for the disease (White 2004), as a lot of other diseases, as most cases of hypertension have no other cause either, except of certain criteria. The ‘advantage’ of hypertension is that there are, at least, some altered reference values. ‘Existence’ is not the issue. Karl Krauss shall have said: “The main disease is the diagnosis”. And this is 100% true when the map is taken for the territory (2).
In older times many diseases have been called essential or functional. But this always has provoked the anger of physiologists. “No longer able to find an anatomical relation, men said then, that the disease was essential, i.e., without any lesion; which is absurd, for it amounts to acknowledge an effect without cause” (Bernard: 113).
This is a crucial point in the understanding of epistemology. If a linear, causal approach is chosen then there has to be something to cause an alteration, a disease. Without such a definable cause things seem to become unsound. Therefore, the ‘functional somatic syndromes’ are ‘not medically explainable’ states (Henningsen et al 2007). That is, the disease seems to be uncaused. It does not help to call it psychosomatic, as this is only wording where a fictive cause for an otherwise unexplained state is postulated (chap. 2.3).
However, functional problems exist. Autoimmune diseases are a result of impaired functions. So even anatomical lesions and physiological alterations might be nothing else than an expression of an imbalance. But the current medical paradigm of cause and measurement does not provide a suitable tool to understand these processes. Imbalance is not a measurable item. This is why current diagnosis is more based on criteria than on function.
(1) The question how self-created criteria induce a ‘truth’ has been investigated by the ‘Deflatory Theory’ (Stoljar/Damnjanovic 2007).
(2) In how far a diagnosis is just an intervention will be discussed later (chap. 4.6).