Alethia Eleutheroo

Clinical Trials – Heterogeneity Problem

The fundamental methodological problem of RCT for drug testing is that of the evident and clear variability in human metabolism operating within an approach that depends on homogeneity of test subject population.

Anatomists have long known that there is little homogeneity in the shape of organs even in apparently healthy persons, and that the drawn shape of organs is an artificial ideal or representation of reality, not reality itself. More recently, biochemists and physiologists have discovered that each person has a range of metabolic functions and markers that differ, sometimes widely, from those of others. Apparently “normal” persons manifest very different metabolic patterns.

While healthy young men of the same height and weight may resemble one another in their overall oxygen consumption, specific chemical reactions may take place, under basal conditions, five or ten times as fast in one individual as in another. Perhaps the most direct extensive evidence on this point is based upon differences in enzyme levels and enzyme efficiencies. Perhaps next to this in importance is the fact that there are wide individual differences among “normals” with respect to several endocrine activities, and there is also anatomical evidence of substantial differences in endocrine patterns … The overall conclusion seems clear that, while the body chemistry of each individual is subject to some change with environmental conditions, each individual would, if subjected to the same stress and given the same food, exhibit a highly distinctive metabolic pattern. This pattern is genetically determined and undoubtedly correlated with his distinctive set of organ weights and activities. (Roger Williams, quoted from Coulter, The Controlled Clinical Trial).

Williams further claims that likely less than 3% of persons deemed “normal” need the presumed “minimum daily requirments” of nutrients. There is actually no normal in an absolute sense; instead, normal is defined effectively as the average of individual variability, meaning that the person who actually fits the artificially created “normal” regarding biochemistry is less common than those who are apparently healthy and don’t fit the statistically defined norm; indeed, Williams claims that “we have never found [a subject for clinical trials] who exhibited a pattern which was free from distinctive variations from the average.”

This is the problem of human individuality. It extends from fingerprints, to biochemistry to thought processes. Human heterogeneity is greater than that of animals species. Metabolism also varies with age, sex, diet, seasons, dirunal cycles, emotions, etc.

A critical heterogeneity for RCT relates to “immune globulins.” These are involved in reacting to any poisonous materials, including drugs, which are essentially poisons (RCTs were not so much intended by health authorities to measure efficacy, but safety, thus, in essence, whether any efficacy outweighs adverse effects, which is deemed to constitute “safety”).  “Even the reaction of the same individual to the same antigen can vary with circumstances” and the number of antibodies one individual can generate seems to be endless. (Coulter, The Controlled Clinical Trial)

As Coulter concludes, while there are clearly things in common, “human variety is the fundamental issue” in RCTs, and such a procedure will only be valid “if it can accomodate and allow for, the heterogeneity of human sickness.” For the very basis of the RCT is a presumed “homogeniety” that must be there, or at least produced, screening for the evident and widespread heterogeneity:

…that there are significant similarities among the diseased patients which we can recognize and group into diagnostic categories [which] are meaningful [such that] patients presenting the same medical condition will respond to the drug in a similar fashion.” (Theodore Greiner, MD, 1965, quoted in Coulter)

And yet, the statistics show that most drugs tested only work for less than half of the patients they are prescribed for, as revealed by the head of research at one of the largest pharnaceutical companies. (https://www.rejuvenation-science.com/n_drugs_performance.html)

Thus, the very nature of human difference is what undermines the effectiveness and validity of the RCT.

The conflict is demonstrated in the fact that to the extent that a RCT takes into account the heterogeneity as much as possible, it will produce more valid knowledge of human biology of a given disease condition, but at the cost of reducing the validity of the findings for the efficacy of treatment. Another is that the more the patient group is narrowly defined to account for commonalities, the smaller the number of patients will be found for the trial, as expressed in the so-called “Lasagna’s Law” named after the medical educator, which states that “it is worldwide experience that the supply of case material is in inverse proportion to the facilities for studying it.” (quoted in Coulter) One observer stated that sample sizes are too small to reach a conclusion in most studies. (Curtis Meinert of the John Hopkins School of Hygiene and Public Health, quoted in Coulter).

Despite the best efforts to make the process filter out and account for the differences, in order to find any commonalities, there are still severe methodological problems that cannot be overcome. Part of this is due to human nature itself (physician and patient biases) and part due to significant epistemological issues involving a test which seeks to remove human difference to test a drug which then must be applied to humans who will react as much if not more on their individual biochemistry as on what they might have in common. The low success of drug treatment, and the fact that they work only slightly better than placebo, and the very problem of modern medicine – so-called side-effects, or rather unwanted poisoning effects – is itself a function of this innate and deep human individuality.

Another major problem is reproducibility, which is one of the hallmarks of the physical scientific method, as clinical trials are often not reproducible or even comparable with one another.

If the clinical analyses of one doctor and another, of one medical center and another, or of patient groups within the same medical center are to be comparable, then the populations must be identified and divided according to their pertinent clinical properties … Without such identifications, subgroups of patients cannot validly be compared. Without the identifications, unreproducible clinical investigations are perpetuated and increased… Neither the experience of many clinicians, the reports and surveys of the medical literature, nor the data now being enthusiastically stored in computer programs have been arranged with consistent uniform classifications for differentiating precisely among clinical subgroups. Whether stored in clinician, literature, or computer, the data of one system or source often cannot be compared with those of another; physicians in one location may find they cannot rely on interpretations made elsewhere; statistical and computational analyses, therefore, yield precise but useless generalities, often inaccurate, and often valueless in application to individual patients. (Alvan Feinstein, M.D., 1976) (quoted in Coulter)

Epistemologically, it would make more sense to design a test that worked with and on the basis of human heterogeneity than a test – RCT – that works despite and against this heterogeneity, like trying to put a square peg into a round hole.

June 13, 2010 - Posted by pleromic | Science