Some Historical Evidence Against Evidence-based Medicine.

Elliott B. Martin, MD        

University of Connecticut, School of Medicine, Farmington, CT

This article, a reflection upon the earliest records of empirical versus speculative thought, as compared with the more recent trend toward an exclusive empiricism in all scientific proceedings, presents the historical case for the allowance of renewed speculation in all scientific endeavors.

Keywords: Ancient medicine, Babylonian medicine, Egyptian medicine, empiricism, evidence-based medicine

For full text of this article, including supplemental notes please download the PDF at the end of the page.

Share on Facebook.

Some Historical Evidence Against Evidence-Based Medicine

The idea for this paper occurred to me one night as I found myself absently wondering how an ancient mind – and not a C+, but an A-worthy intellect at that – might interpret a Randomized Control Trial. To the pre-Greek mind, that is, the nearly perfectly empirical mind, the ends far outweighed the means. And more, the very idea of clinical detachment would be incomprehensible. Should a sick man swallow a pill and become healthy once more, then there is a ‘magic’ in that pill; should another sick man, with the exact same manifestations of disease, swallow a pill in every way like the first man’s, and not become healthy, then there is no magic in that second man’s pill. To the Babylonian or Egyptian mind there would be no issue here. The gods favored one man, and not another. In fact, if one were to believe that a little pill swallowed whole were all that were necessary to cure a man of illness, then one might indeed be considered mad in a world where the evidence before one’s very eyes changed daily.

But let’s say an especially observant ancient physician noticed over time that half of those who swallowed that same pill recovered their former health, while the other half did not. By current standards, statistically speaking, this would be considered a pretty effective treatment. By ancient standards, statistically speaking, this would be considered exactly what it is: the equivalent of a coin flip.

And make no mistake. The Ancient Mesopotamians were exceptional classifiers, maintaining inventories of ‘Lists’ of every conceivable category. Much like early and late Enlightenment ‘naturalists’, such as Linnaeus and Darwin catalogued the natural world as far as they could see, so did ancient Mesopotamian and Egyptian scribes meticulously catalogue the ‘natural’ world all around them. Was this ‘science’? If science, at its core, relies on empirical observations of the natural world, then there is no argument legitimately to be made against it. Where the Enlightened and the Ancients part ways, however, is in the cataloguing of the ‘supernatural’. The Mesopotamians especially, living in a more volatile environment than their Egyptian brethren, along with their lists of fauna, flora, gods, kings, words, historical events, and even laws, maintained extensive catalogues of omens, these based on standardized interpretations of dreams, physiognomy, star alignments, oil-in-water patterns, miscarriages, and internal organs of both animals and humans.

Beyond the Enlightenment how might one then conceive of a ‘scientific’ study to test the reliability and validity of such phenomena? The Mesopotamians themselves essentially performed prospective cohort studies. Standard mantic texts such as the “Dream Book” and the “Diagnostic Manual” were compiled not overnight but over centuries. These were conclusions based on centuries of observations, centuries of observations clearly indicating that if a man dreamed such and such, then such and such, more often than not, was likely to happen, that if a sacrificed goat’s entrails displayed such and such, then such and such, more often than not, was the likeliest outcome. This is not religion, nor is it superstition. This is empiricism at its purest. And if the diviners were correct even one-third of the time, would that not validate their practice? At least by current standards?

So what conclusion is to be drawn from this? Are we to return to the practices of oneiromancy, extispicy, tocomancy, and teratomancy? Have we deviated at all from those paths? What is medicine after all if not a prophetic practice? When we treat a patient with a medication, for example, because that medication has been shown to be effective in half of all other patients, are we not predicting the future? Do we not habitually reassure patients in this regard: ‘(Based on my clinical observation) this medicine will have you feeling better in a week.’ Or, if the ‘evidence’ is equivocal: ‘If this medicine does not have you feeling better in a week, we can try something else.’ And of course, when all else fails, the one prediction that is always correct? Death.

Death, however, posed an interesting philosophical conundrum to the Ancient mind, a conundrum that has arguably been overlooked since. For the Mesopotamians distinguished at least two different types of death, that from natural causes and that from violence. This distinction was based on divinity: though the gods were exempt from natural death, they, too, were not exempt from violent death. Ancient physicians wanted nothing to do with the dying patient when the etiology of death was disease. Yet all manner of medicinal contrivance was employed on the battlefield. In fact much of the divination art is devoted to the prediction of death by causes other than the natural. This, however, should not be dismissed as evidence of primitive superstition; on the contrary, given the ancient obsession with classification and observation, this search for predictors of death is properly to be understood, as is our current search for predictors of death, as the ultimate attempt at the imposition of order. After all, any physician can predict death as the outcome of a fatal illness. The physician, however, who can predict death from among seeming randomness has certainly acquired a superior level of insight.

And a brief word about the texts themselves. The mantic texts were created as study guides for practitioners. And the fact that these practitioners were also the inventors of writing must not be overlooked. Whether writing in cuneiform or hieroglyphs, those few in possession of that very special knowledge were in the very process of creating abstractions from concreteness. For example, the pictogram of a hand is a concrete symbol, originally meaning ‘hand’. However, with pressures to expand the semantic range of script, such concrete representations soon took on more abstract meanings as well, such as “to give”, “to strike”, “to help”. This of course, depending on context, allows for a fair amount of ambiguity, a fair amount of individual interpretation of a text. Thus, writing itself was developed as an empirical tool, at its heart based in a deep-rooted realism, or phenomenology. What followed was, and is, the construct of abstraction, of theorizing, of the scientific process. To those more familiar with the much later Greek rationalism, early script may be likened to Plato’s ‘Forms’, the origins, or prototypes from which all else derive.

This brings us back around to my original question as to how an Ancient mind might interpret a Randomized Control Trial. I probably have asked the wrong question. Statistics is a game of logic, and the Ancients were extremely competent logicians. But the predictive value of logic is minimal in a world governed by confounders. And the ancient mind simply could not divorce itself from the empirical world. What is relevant for us, however, is that, as among the great religions that followed, there existed among these creators of the written record the need to record and transmit their accumulated knowledge for future generations. To put this in perspective, for nearly five thousand years science had advanced human understanding by relying on speculative thought as applied to the ‘natural’ world. For about twenty some odd years now, science has relied on statistical analysis as applied to a ‘controlled’ world. As a once and future student of the humanities, I wish only to point out that our knowledge has remained unbroken over these millennia, such that a 21st century physician should have little difficulty in overstepping his or her own historical bounds. The world, after all, is still a confounding place.


Elliott B. Martin, Jr. is a first-year resident in Psychiatry at the University of Connecticut Health Center. He also has a Master’s degree in Near Eastern Languages and Cultures (NELC) from UCLA, and is formally advanced to doctoral candidacy. A former instructor in both Classics and NELC, and a former secondary school teacher, medicine is a second career for Dr. Martin. He can be reached at:

Historia Medicinae Editor,
Dec 21, 2009, 8:57 PM