By Toby Rogers at Brownstone dot org.

Introduction

In a previous article, I presented ten practical and material criticisms of Evidence-Based Medicine (EBM). But there are even larger metaphysical, ontological, and epistemological problems with EBM. Numerous authors make the case that EBM and evidence hierarchies elide important debates in the philosophy of medicine. In this article I will review seven philosophical debates in connection with EBM and evidence hierarchies including:

1. Hierarchies are not how causation in science is usually constructed;

2. Evidence and interpretation are two different things;

3. The inferential gap may be unbridgeable;

4. Bayesian statistics has long since proven superior to the frequentist statistics relied on by RCTs;

5. Science can never prove hypotheses, only refute them;

6. Actual medical practice is necessarily pragmatic and different from the objectivism of EBM; and

7. Medicine is a practice not a science per se.

Hierarchies Are Not How Causation in Science is Constructed

Several authors have noted that EBM tends to overlook and ignore the contributions of basic science (also called "bench" or "fundamental" science and I will use these three terms synonymously in this section). Bench research is defined as "any research done in a controlled laboratory setting using nonhuman subjects. The focus is on understanding cellular and molecular mechanisms that underlie a disease or disease process" ("Bench research", n.d.).

Merriam Webster's Dictionary defines basic science as, "any one of the sciences (such as anatomy, physiology, bacteriology, pathology, or biochemistry) fundamental to the study of medicine" ("basic science", n.d.).

The CEBM evidence hierarchy lists basic science as the fifth level of evidence, below the threshold suggested by Strauss et al. (2005) and others as even worth reading.

To be clear, the CEBM and other evidence hierarchies are not excluding bench science entirely from the study of medicine - they are proscribing the consideration of bench science by doctors when they make clinical decisions (presumably others, namely pharmaceutical companies and academic researchers would be free to continue with a more comprehensive approach).

Excluding basic science in this way is an odd choice because basic science has always been an essential component of establishing causation.

Bluhm (2005) writes,

[B]ench [laboratory] research and clinical (epidemiological) research are intimately related. The history of epidemiology shows that advances in one of these aspects of biomedical research often depends on advances in the other; this point is particularly clear in the case of infectious diseases but is equally important for understanding chronic disease (p. 538).

Bluhm (2005) argues that EBM should move from hierarchies of evidence to "networks of evidence" in which both epidemiology and lab-based biochemistry work hand in hand (p. 535). It is a fine point as far as it goes, but it strikes me that one could push this idea of networks of evidence even further - to include the subjective wisdom of both doctors and patients as well. I will elaborate on this point later in the article.

Rawlins (2008) writes:

Hierarchies attempt to replace judgment with an over-simplistic, pseudo-quantitative, assessment of the quality of the available evidence…. Hierarchies of evidence should be replaced by accepting - indeed embracing - a diversity of approaches (p. 586).

Goldenberg (2009) argues that the degradation of pathophysiology in evidence hierarchies is unwarranted "as pathophysiology often provides more fundamental understanding of causation and is in no way scientifically inferior" (p. 180).

La Caze (2011) voices alarm that evidence hierarchies overlook the contributions of basic science. As pointed out above, basic science is usually assigned to the lower tiers of evidence hierarchies. While the assignments to the different tiers are rationalized based on reference to "quality" in fact, "p...