Georg Ivanovas From Autism to Humanism - systems theory in medicine
For a lot of reasons it became necessary for our current medicine to adopt systemic concepts. The most important are in my opinion:
Complexity management in basic research
Basic research in biology and physiology has reached a point where no further progress seems to be possible without systemic tools (Reeves,/Fraser 2009). Polycontextural concepts (chap. 3.5) are necessary to model recursive and complex processes. This is especially true for brain research and genomics (Ahn 2006a; 2006b), but concerns all other medical topics. The European Science Foundation sees in systems biology the ‘key to future medical breakthroughs’ (European Science Foundation 2009).
Limitations of the reductionist approach
The machine model in medicine (chap. 2.1.a) expecting that the correction of a subsystem has no further impact onto the whole has led to a tremendous improvement of short term therapies. In the treatment of chronic diseases the contemporary approach was less successful. Too often current medicine provides only a symptom management. People live longer, but “they experience poor health for longer” (Cole 2005) and it is even uncertain whether medicine contributed to this longevity (chap. 2.2). For some time there was the hope that the engineering of hormones, enzymes or the genome would solve this problem. But these hopes slowly fade away. The manipulation of certain physiological mechanisms does often not lead to a long term benefit. Already in 1968 Bateson provided a detailed model why this is so. Based on the cybernetic research of Ashby he demonstrated that linear interventions, which he called conscious purpose, might contribute to a decline of health (Bateson 1972: 426-447) by inducing ‘network pathologies’ (chap. 6.2, 6.7), as they are called today.
Reductionism creates a situation which has been branded as an autistic-undisciplined thinking (chap. 2.1.c), making it often difficult to come to clear decisions (Hu et al 2004). But also ‘clear evidence' is often more a result of which facts are taken into account and less of the encountered situation. Much depends on the underlying beliefs and values of the physician (Geneau et al 2008). However, value decisions are normally not addressed as such. They appear in the disguise of ‘clinical decisions’ (Parker 2004, Sennet 2003).
An example: The hygiene hypothesis says that the contact with a variety of germs is beneficial for the health of asthmatics. A value decision would be: “Eating dirt or moving to a farm are at best theoretical rather than practical clinical recommendations for the prevention of asthma” (Weiss 2002). Nevertheless, some physicians expect from their patients with chronic disease a fundamental change of life-style. The administration of probiotics (chap. 2.8.b) would be a method of eating dirt, even in line with the prevailing opinion that therapy is drug therapy. Some scientists also proposed to vaccinate atopic patients in order to induce an appropriate immune response (Watts 2004). But does it makes sense to vaccinate against a large number of germs which might contribute to atopic disease (Bach 2002) and then to vaccinate other germs to counter-balance this effect? It is not important to discuss here whether these theories are valid or not. The example shall only demonstrate an important question: what kind of control or therapy is intended?
Other examples of value decisions are: Is polypharmacy in a 95 year old lady really an improvement for her life (Fitzpatrick 2003)? In how far are the aims of the physician in line with the values of the patient? Might an unspecific therapy be better than a specific therapy or even no therapy at all? What kind of supporting measures are necessary? What will happen under changing conditions? After 5, 10 or even 20 years? How is the patient to be convinced to follow the directions of the physician? By authoritarian behaviour? By frightening the patient? Or by discussing pros and cons? Who decides under which circumstances?
These are crucial and everyday questions in medical practice. They include the question of power (chap. 4.6.a), as well as question of the order of life (chap. 5.4.b). For their decision the normal ‘evidence’ provides at best a supporting help.
Most physicians are not aware that they make continuously value decisions. Such issues are rather seen as an ethical or humanistic affaire, not as scientific medicine (Cousins 1979: 125 160).
The current epistemology (the reductionist and specific approach) creates a picture where the therapist (the most relational aspect in medicine) should be irrelevant for the therapy (Hammer: 49). The therapeutic relationship seems to be just an appendage of drug therapy subsumed under the label of the placebo effect. In its final consequence the specific medicine can be done by machines. There are already proposals to connect ever physician via internet with the National Library for Health providing him/her with the best possible evidence for the patient. The only task left for the physician would be to help the patient to consider the harms and benefits of a therapy (Evans et al 2006: 96-98). But no physician is necessary to do this. It could be done by a nurse or an interactive computer program. Also the specific diagnosis could be established through a questionnaire leading to a series of machine based tests. Even googling leads already to a quite successful specific medicine (Tang, Kwoon Ng 2006).
Such proposals and developments show that the ‘specific’ medicine is blind for the therapeutic reality. Many things happening in the therapeutic context (chap. 5.7) just do not appear in the scientific context. These issues seem to represent a kind of not reliable soft science (Nature editorial 2005a). It is therefore no surprise that there is a steady decline of patients’ confidence in the physician (cf. Chen 2008c and the related bolg of the New York Times). Although the psychosomatic movement of the last century has investigated such topics with certain results, the whole approach remained somehow vague, probably because the specific paradigm has rarely been abandoned.
The psychoanalytical concept of transference and counter-transference was a first attempt to describe the pattern between the therapist and the patient. For general practice the Balint groups (chap. 2.2) train the physician to understand, or better, to perceive that a lot more happens in the therapeutical setting than suggested by our models. There can be no doubt that the process of mirroring (chap. 4.2) has an enormous impact for the therapeutical relation, as well, but we have no idea in how far a therapy is influenced by the structural coupling of humans. May be there is even a ‘conjoined physiology’ (chap. 5.2) between physician and patient.
A scientific medicine has to render an account for such processes. But this is impossible with the current epistemological tools. Therapies like hypnosis or suggestive therapy represent somehow the science of the unspecific. They focus mainly on the context of the therapeutical relation whereas the specific approach focuses on the content. But the two remain unconnected.
Limitations of medical cognition
Some formal aspects influencing medical cognition (chap. 2.2) are:
- In the specific concept a learning of a higher order cannot be modelled (chap. 4.3).
- The specific model is not able to understand fluctuations far from the equilibrium (chap. 6.4).
- Is not able to understand complex rhythms (chap. 5.3.b).
- It is not able to distinguish between robustness and rigidity (chap. 6.4).
Statistics have no semantic relevance for the individual
Statistical medicine with its emphasis on randomised controlled trials is a kind of deadlock in regard to the individual. It is only able to provide a general impression. Individual statements are impossible (chap 2.1.f; 2.5.d). They have no semantic relevance (chap. 2.1.d). A therapy might be beneficial, harmful or without consequences for an individual despite the effects found for a population. There is, for example, no possibility to predict immanent side effects of a therapy. “Over the past 30 years attempts have been made to enhance the recognition of adverse effects by "data dredging" or "data mining." But the results have been modest (Stricker/Psaty 2004). A solution cannot be found in statistics. What seems to be necessary is a model of individual prognosis. This might be developed from a combination of empirical experience with systemic concepts including biological hierarchies (chap. 6.12).
Such a new model provides a language that allows the investigation of relations, generates questions and becomes a tool for comparative studies of different fields of phenomena (Bateson, 1988: 37). For example, until recently the biological theory excluded the possibility of the transmission of acquired characteristics. Now, as the theory has changed, observations supporting this theory proliferate. “There is nothing more practical than a good theory,” as Lewin has said.
In order to understand a systemic medicine and to initiate further developments only one main prerequisite is necessary: health, disease and therapy have to be understood as regulative, not as statical. Although probably nobody would really doubt that medicine is regulative, the regulative approach has certain implications which some might hesitate to accept. First, regulation is an emergent process (chap. 4.10) where simple causal attributions hardly describe the observed events. Second, changes (and even the status quo) have to be observed in time series, and cannot be judged from a snapshot.