Georg Ivanovas From Autism to Humanism - systems theory in medicine
2.4 The placebo effect
A paradox is the discrepancy between reality and the idea how reality should be.
My first contact with placebos was during an internship in a rehabilitation clinic. I was called to a patient (5th day after an operation of an intervertebral disc) at 2 a.m. She had severe pain, although she had already taken all oral drugs (all RCT proven) left for emergency. She said that only an injection could help her. So I injected an ampoule of physiological salt solution. Already during the injection she relaxed and slept in my presence. The next day she was refreshed and nearly free of pain.
This is an undisputable example of a placebo effect that corresponds to all the parameters under which a placebo effect is to be expected: previous medication, strong pain, expectation and injection (Gauler/Weihrauch 1997: 13-16).
How does this fit into a scientific concept? What is the basis to see here a placebo effect? If I had injected half of an ampoule of diazepam (still often in use then and the most probable alternative), this would not have been spectacular. But as it was salt solution we define the reaction as a placebo effect. Colloquially one could say that salt solution does not give rise to such effects. This is the usual, mostly used logic. However, it is not logical. This example proves that salt solution gives rise to such effects. This is a main, mostly overseen phenomenon in the discussions of placebos. A therapeutical effect is declared as not therapeutic. The reversal of causal logic is the definitory foundation of the placebo effect. Only by a therapeutical force something becomes a placebo. An injection of physiological salt solution has an analgetic and sedative effect. My example is prove. The only objection would be to say that the patient would have fallen asleep anyway without my therapy, which is not very probable.
What looks like a piece of juggling is nothing but a change of the frame of observation. A verum (the drug containing an effective substance) is expected to have an effect, therefore only the substance is seen. A placebo (or what is defined as such) is expected to have no effect, therefore the frame is observed. It is a (metaphysical) change of strategy of observation.
Normally the problem is solved by modelling. It is assumed that physiological salt solution on grounds of our physiological knowledge is not able produce analgesia. Although logical, it does not solve the main problem. An effect is not the result of a model but of an observation. A model has to explain the facts not the facts the model.
This is a widespread mistake. A substance is defined as placebo because the scientist or author just cannot imagine that the substance might have an effect. So Shapiro (the father of placebo research) states - a little bit complacent - that in ancient times “the astonishing total of these ancient remedies is about 4,785 drugs and 16,842 prescriptions. Even more startling is that with only a few possible but unlikely speculative exceptions, all were placebos” (Shapiro/Shapiro 1997: 13).
He subsumes under placebos e.g. mandrake, a plant with a high concentration of alkaloid substances which in ancient times was used as anaesthetic drug during operations (Baumann 1999: 108) and is the emblem of the Hellenic Society of Anaesthesiologists.
It is the lack of knowledge paired with a false certainty that old medicine was helpless that leads to such wrong conclusions. Similar mistakes happen if the theory of a remedy is rejected. “Paracelsus…used the placebo effect of his remedies: ….Chelidonium because of the colour of its sap as a remedy for liver and gall bladder” (Gauler/Weihrauch 1997: 172, my translation). However, Chelidonium, that’s sap is a little bit yellow as bile, contains a big amount of biological spasmolytic substances, and is effective in cholecystopathia and cholecystolithiasis (Gessner 1974: 55-56).
Moreover, not so rare the model of physiological and pharmacological pathways is wrong “Until recently, medical students were taught the pathophysiological reasons why β-blockers are contra-indicated in heart failure (they are a good treatment for heart failure); why colloid is more effective than crystalloid for fluid replacement (it is worse) and that because the vascular supply of the scaphoid places it at risk of non-union, any suspected fracture requires a cast (active mobilisation results in better outcomes) ” (Doust/Del Mar 2004).
It is common practice to call substances a placebo that do not work according to the beliefs of the scientist. We know nearly nothing on the medical power of a list of ‘ancient placebos’ as “theriac, mattioli, mithridate, bile, blood, bee glue, bones, bone marrow, claws, cuttlefish, cock’s comb, cast-off snake skin, fox lung, fat, fur feathers, hair, horns, hoofs, isinglass, lozengs of dried viper, oil of brick, ants, and wolves, powder of precious stones, seasilk, sponge, scorpions, swallow’s nest, spider webs, raw silk, teeth, viscera, worms, wood lice, human placenta etc” (Shapiro/Shapiro 1997: 14). As long as there are no RCTs it is nothing than a belief and even then there is no proof, if, as with Viagra (chap. 2.5.c) not the right parameter is investigated. The conclusion that ancient “therapy rested on placebo foundations, despite the tendency of historians to glamorize, sentimentalize, and exaggerate the significance of primitive and prescientific medicine” (Shapiro/Shapiro 1997: 15) is no valid statement. There might be highly effective drugs among the mentioned list. There might have also been a precise knowledge how to administer certain drugs..
It is habit to call a remedy a placebo not in line with the current hypothesis of effectiveness. But to take the model of a therapy as the basis of placebo definition leads to a variety of paradoxes. It would mean that having a model makes a remedy an active substance (verum) and the lack of a model a placebo. So Jenner’s vaccination would at first have been a placebo and after understanding the physiological process a verum. The same with aspirin. It was prescribed long before we understood how it relieved pain (Spiro 1997: 40).
The next mistake in the ‘model hypothesis’ is that there is a model for nearly all medical activity. Immediately the question arises whether or not someone believes the model. Do I believe that venesection frees an organism of unhealthy substances? Do I believe that digitalis strengthens the heart? Do I believe the models that the peoples in the rain forests (often using forces of gods) have for their highly effective plants?
But it is even more complicated: One placebo definition says that a placebo is “a pill deemed inactive by a physician but given to relieve the complaint of a patient” (H. Shapiro cited by Spiro 1997: 44). According to this definition a substance is a placebo, when the doctor believes it to be a placebo but the patient believes it to be a verum. However, a placebo works better, if the practitioner believes it to be a verum (Benson/Epstein 1975). Then a placebo is a substance where the practitioner believes it to be a verum but the scientific community believes it to be a placebo. But effectiveness is not a decision of one person or a community but a result of observation.
Physiological salt solution is a placebo and not an analgetic drug not because nobody believes it to be an analgetic drug nor because there is no model, but because normally it is not effective as an analgetic drug.
From the point of observation it only can be said that a placebo has normally no certain effect (because sometimes it has one) and a verum normally has a certain effect (because sometimes it does not have one). Theoretically (and practically it is the same) there is no clear distinction between a placebo and a verum. The expressions describes an arbitrary judgement on a coordinate of effectiveness in a certain frame of observation. A verum becomes a verum if it is more effective than a placebo. Verum and placebo are but statistical parameters that are mutually conditional. That is, there is no placebo effect as such. Verum and placebo are probabilities of effectiveness, terms of describing collectives without semantic relevance (chap. 2.1.d). Therefore they cannot be applied to the individual case. This is according to the current theory that a causal connection can never be stated in a singular observation (Kiene 2001: 10-16).
The scientific correct statement would be: “ As physiological salt solution normally (or: statistically) has only a small analgetic effect, it is in this case probably a placebo effect”.