1.2.2 Scientific thinking and ideological attitudes

Our knowledge of the world does not depend solely on our observational techniques, which are always open to improvement. It's more than that. We are constantly evaluating events in our minds. We constantly associate them with concepts in order to understand them. We mechanically translate them into language to find meaning, as if reality were just a series of sentences. "In the process, we constantly mix up two different dimensions: that of our thoughts and that of reality. This mixture causes confusion in our understanding of the world, with far-reaching consequences" [1]. Scientific knowledge and its laws remain in the realm of language, even if it is as rigorous as mathematical language. They have all the limitations of language, as Wittgenstein so clearly demonstrated [2]. We have no direct access to reality, only to our mental representation of it. Defining a phenomenon such as a heart attack, for example, means isolating it from reality and transforming it into a concept through thought; after this initial process we are left with nothing more than a language game, or a screen that masks reality.

Scientific thinking was built against religious beliefs, but the attachment and respect that the West has for scientific theories is itself a religious attitude. If we only trust - but absolutely trust - randomised trials or evidence-based medicine, we ourselves lack judgement and rigour, and we cut ourselves off from a part of reality that does not fit into this framework [8]. That belief in science itself becomes an ideological attitude can be seen from the faith with which the latest publications are accepted, provided they remain within the framework of a certain well-established convention and enjoy the backing of a journal with a high impact factor. The resulting propositions quickly become dogmas that a few thuriferous individuals take it upon themselves to enforce under the threat of a new Inquisition. Everything seemed to confirm the data; it was a time of euphoria. And when the conclusions are subsequently questioned or disproved, the same enthusiasm for condemning what has been worshipped develops. A review of ten years of publications in the New England Journal of Medicine found that 60% of the 1,344 studies in the field of medical therapeutics proposed the opposite of what had previously been accepted or failed to reach valid conclusions, and that 40% of the 363 articles testing practices considered standard concluded that they were no better than simpler, less expensive therapies [5]. The more fashionable the field, the wider and faster the back-and-forth [3].

In a trial, we ask a question that we hope to find an answer to - for example, is a particular substance more effective than a placebo for a particular condition in a particular set of circumstances? But the very fact that we ask a question in a certain way determines the answer we get, even in the absence of any outside interference. The design of the trial itself is a fundamental bias that will inevitably affect the results, turning them in part into a very precise measure of the original bias. There is another bias: the hotter the topic and the greater the economic interest, the less likely the results are to be accurate [3]. What's more, the meticulousness with which an event is observed introduces a new element, known as the Hawthorne effect: the meticulous observation of a phenomenon changes its probability of existence by taking into account irrelevant events that would otherwise go unnoticed. In a randomised controlled trial, what we are looking for is induced by the rigour of the framework, the precise choice of observations and the highly selective recruitment of patients; sometimes less than 10% of potential candidates are selected for the trial [4]. In a therapeutic trial, patients are observed and personally involved with far greater rigour than in normal monitoring; they do not miss a dose, they do not miss a check-up and they are never out of sight. In the perioperative period, careful monitoring can have as much impact as the observed effect of the drug. This is why the first publications on a phenomenon usually show extraordinary results. Subsequently, those who try to repeat the experiments and approach the problem from a different angle, or in the real world of everyday practice, generally obtain less striking results. Finally, the element may even be abandoned because it is invalid or because its defects outweigh its benefits. Many drugs and techniques have gone through this sequence: atropine in premedication, thiopental in cardiopulmonary resuscitation, aprotinin in cardiac surgery, the Bain cycle in ventilation, and so on... Perhaps preconditioning, β-blockers and antiplatelet agents for the prevention of perioperative ischaemia will one day suffer the same fate....

In clinical practice, the validity of a conclusion will ultimately be demonstrated by the introduction of therapeutic changes that significantly improve the quality of care provided to patients. The fact that these improvements are significant will in turn be demonstrated by quality control studies of patient outcomes. There are still very few of these in anaesthesiology, and they do not always provide clear enough answers to establish a real consensus on the subject. Those that do exist are often inconclusive. This is not necessarily because the trials are too small, because the larger the trials, the greater the chance that the results are the result of significant bias [3].

Looking at the medical literature over the last twenty years, it is sad to say that the power of the results is getting lower and lower. For example, dual antiplatelet therapy, which is universally regarded as essential in the secondary prevention of cardiovascular risk, only improves patients' prognosis by 20-30%. That's still an improvement, but it pales in comparison to the huge sums of money invested in research and treatment. Yet the entire management of a disease is being changed in the name of a trial that, under ideal experimental conditions, shows a benefit of around 25% (hazard ratio 0.25) compared with placebo. It is hardly surprising that half of the therapies tested in the last decade are no better than simpler, less expensive treatments [5], and that 45% of the proposed therapeutic interventions should not be used clinically because of the weakness of the evidence; of the trials whose conclusions indicate a probable benefit, only 2% have results that are sufficiently clear to not warrant further investigation [7]. As for the recommendations published in cardiology (ACC/AHA guidelines), 48% of them are based on only level C evidence (see Figure 1.9) [6]. Controlled by the editorial boards of the major journals, supported directly or indirectly by industry, and having become a prerequisite for any academic claim, clinical research has less and less chance of producing new ideas and revolutionary discoveries, so much so that it hums in a vacuum within an intellectual establishment bound by convention. Conflicts of interest are not just a matter of industry or academic politics; they already exist in the researcher's belief in his or her own theory and in the rational presuppositions of the scientific paradigm.  

© PG Chassot  April 2007, last update September 2019

References

1. GALATIS A. L'indicible. Paris: Editions Accarias L'Originel, 1997, p 5
2. HADOT P. Wittgenstein et les limites du langage. Paris: Librairie Philosophique J.Vrin, 2005,126 pages
3. IOANNIDIS JPA. Why most published research findings are false. PLoS Med 2005; 2:e124
4. POLDERMANS D, BOERSMA E, BAX JJ, et al. The effect of bisoprolol on perioperative mortality and myocardial infarction in high-risk patients undergoing vascular surgery. N Engl J Med 1999; 341:1789-94
5. PRASAD V, VANDROSS A, TOOMEY C, et al. A decade of reversal: an analysis of 146 contradicted medical practices. Mayo Clin Proc 2013; 88:790-8
6. TRICOCI P, ALLEN JM, KRAMER JM, et al. Scientific evidence underlying the ACC/AHA clinical practice guidelines. JAMA 2009; 301:831-41
7. VILLAS BOAS PJ, SPAGNUOLO RS, KAMEGASAWA A, et al. Systematic reviews showed insufficient evidence for clinical practice in 2004: what about 2011? The next appeal for the evidence-based medicine age. J Eval Clin Pract 2013; 19:633-7
8. WALLACE BA. The taboo of subjectivity. Oxford: Oxford University Press, 2000, 218 p