Hydroxychloroquine for the treatment of COVID-19: an approach based on the philosophy of science and heuristics

How to cite this article: Solla DJF. Hydroxychloroquine for the treatment of COVID-19: an approach based on the philosophy of science and heuristics. J Évid-Based Healthc. 2020;2(1):28-31. doi: 10.17267/2675-021Xevidence.v2i1.2890 Submitted 04/23/2020, Accepted 05/07/2020, Published 05/08/2020 J. Évid-Based Healthc., Salvador, 2020 June;2(1):28-31 Doi: 10.17267/2675-021Xevidence.v2i1.2890 | ISSN: 2675-021X Hydroxychloroquine for the treatment of COVID-19: an approach based on the philosophy of science and heuristics Concept Article


Concept Article
It is unprecedent such a case of a medication with so few and flawed studies that has so quickly influenced the clinical practice of so many physicians around the world. Unfortunately, this is not the first time we witness a treatment to be picked up and go mainstream despite lack of good evidence, but none has done so in such a brief period of time. Here we will discuss some basic principles on the philosophy of science that should guide our critical appraisal of the available evidence and how heuristics may be biasing medical practices and shaping public opinion.

The null hypothesis principle and the inductive problem
As a basic assumption, we should be aware of the probability of the null hypothesis to be higher than the alternative hypothesis. In other words, most published research findings are false 1 . Most results from pre-clinical and phase I/ II studies are not confirmed in phase III studies.
One of the main reasons for that is related to the primordial conception of the hypothesis. It is not uncommon (not to say, routine) to medical treatment hypotheses to be originated from inductive models (as opposed to the hypotheticaldeductive reasoning) 2 . That is, many hypotheses are formulated a posteriori (alongside an attractive physiopathological explanation) after empirical observations, ex post facto, although not always self-admitted 1,3-5 . This practice is known as HARKing -Hypothesizing After the Results are Known 6 -and contrasts with an a priori hypothesis followed by a study specifically designed to refute it or not.
The first reported uses of chloroquine for the treatment of an infectious disease (infectious mononucleosis) dates from 60 years ago, before its viral etiology was discovered 7 . Since then, numerous studies have followed for several viruses (influenza, Epstein-Barr, Ebola, Zika, Chikungunya) with heterogeneous results, including deleterious ones (whether viral replication increase or clinical worsening) 8 . Biological systems are much more complex and dynamic than the inert environment of in vitro studies. It is not always possible or safe to achieve the preclinical therapeutic dose of the study drug and its pharmacodynamics when interacting with other drugs, hormones and cytokines is almost unpredictable to say the least. Moreover, this unpredictability is directly correlated with the severity of the disease and the extent of organ dysfunctions. The studies responsible for raising the hype around the possible effectiveness of Hydroxychloroquine (HCQ) against Coronavirus  were performed in vitro, in animals and in small clinical studies 9,10 . Clinical trials with appropriate methodology and sample size are still in progress [11][12][13][14] .

The fragile biological plausibility
Another fundamental wariness concerns the biological plausibility of the large effect size advertised by some HCQ enthusiasts. Claims of up to zero deaths and zero hospitalizations have been made on the early anecdotal descriptions of the HCQ effect 15 . Since there is no treatment one-hundred percent effective in medicine, discredit on the supposedly found benefits is justified. Even disregarding these enthusiastic hyperboles, the probability of a game-changing effect remains low. Although drug repurposing is a wellestablished research strategy, the success rate is poor 16 . Specificity is one of Hill's causality criteria, which also give some analogical guidance on the evaluation of potential therapies. Unless based on the collateral effects, non-specific therapies indicated for different diseases due to distinct pathophysiological processes hardly present large effect sizes for the repurposed target. By contrast, highly specific therapies tend to be highly beneficial when proven effective (e.g. antibody, molecular markers and gene based therapies) 17 .
There is a recurring argument emphasizing the HCQ perceived safety and low cost 18-20 . Nevertheless, when a treatment is unequivocally effective, we do use it despite adverse effects and despite the associated cost.  22 . At the same time, the disseminated impression of absence of a treatment for Covid-19 (which is somewhat questionable, since we do know a lot on how to manage severe viral infections) conflicts with the generalized physician mindset that urges: "we must do something" -the active physician mindset.
The ground for the adoption of unproven therapies is now settled.
Hyperbolic and fanciful rumors lead to higher broadcasting than wary and realistic reports. Thus, premature claims of scientific discoveries leading to possible new treatments overflow the media and are intensified by social media bubbles. This predisposes to the availability heuristic, a cognitive bias through which the frequency or probability of an event being true is judged by the number of instances of it that can readily be recalled 23 . "If everybody is talking about the benefit of HCQ for Covid-19, then it must be true". Once accepted this belief, evidences that might confirm it are overvalued, while evidences that might disconfirm or refute it are ignored -the confirmation bias 24 . The likelihood of new, opposite facts or logical reasoning changing one's belief is disappointing due to the uncomfortable state of mental conflict caused by the disagreement between the belief and the new evidence. This cognitive dissonance bias is especially painful in the setting of a commitment bias related to past ideas and public behavior 25 .

False dilemma as the basis for a new, potential case of medical reversal
Many people, including politicians and health care workers, have argued that given the pandemic emergency there is no time available to pursue perfect clinical trials and, thus, the adoption of HCQ is justified. This is a false dilemma/dichotomy. Although it is true that large, multicenter, randomized, controlled clinical trials (RCT) well-powered for all severity strata and subgroups within the Covid-19 spectrum may not be feasible in a short period of time, this does not translate into the acceptance of the results of extremely flawed studies. Small, proof-of-concept RCT or well-designed observational, non-interventional studies (either cohorts or case-controls) can also provide good guidance 26 . Unfortunately, the scientific community demand for minimum standards before the wide adoption of HCQ or other therapies has even been pejoratively classified as scientism.
The current scenario is propitious to a new case of medical reversal: the phenomenon of a new trialsuperior to predecessor because of better design, increased power, or more appropriate controlscontradicting current clinical practice 27 . Reversing widespread medical practices involves time, resistance and costs (financial and, sometimes, lives), in addition to societal loss of credibility in the medical community 28 .

The individualist ethical fallacy on the indiscriminate compassionate use of HCQ for Covid-19
The argument of using HCQ as a hope can be understood as an individualist ethical fallacy, which is the misconception of confusing two distinct and very different notions: the view that the dictates of the individual conscience are always right; and the view that no individual should act against the dictates of conscience 29 . Action and belief must be treated separately. Compassionate use would be justified when the manifestation of the desire to use HCQ comes from the patient or its family, especially for severe Covid-19 cases. In such a situation, coupled with a weighted regard to the patient's autonomy, the benefit of the doubt is in favor of the unproven therapy, given the potential risks were properly disclosed. However, it is a completely different situation to, based on the doctor's belief, indiscriminately use of a medication that lacks minimum scientific evidence background. To obscure the circumstance even further, we should not overlook that informed consent and shared decision is commonly an utopia, given the high disproportion on hierarchical position and technical knowledge between the patient and his doctor. To indiscriminately use HCQ for Covid-19 is a trivialization and disrespect to the hope of patients and family members. A realistic therapeutic plan tied to empathetic prognosis disclosure would be more valuable than a fallacious hope.
In conclusion, there is a need to rescue the principle of non-maleficence -not only applied for the individual patient, but also for the community. There is no hope outside science and ethics.

Competing interests
No financial, legal or political competing interests with third parties (government, commercial, private foundation, etc.) were disclosed for any aspect of the submitted work (including but not limited to grants, data monitoring board, study design, manuscript preparation, statistical analysis etc.).