Epidemiology: Misunderstood? Misused?

The plethora of The-Only-Diet Book-You-Will-Ever-Need publications widely available in sales outlets, each with its own version of the “correct” nutritional philosophy, is testimony to the confusion suffered by the general public. The reason is elegantly described by Taubes:

…600,000 articles – along with several tens of thousands of diet books are the noise generated by a dysfunctional research establishment. Because the nutrition research community has failed to establish reliable, unambiguous knowledge about the environmental triggers of obesity and diabetes, it has opened the door to a diversity of opinions on the subject, of hypotheses about cause, cure and prevention, many of which cannot be refuted by the existing evidence.

This indictment is harsh, but deserved. Witness to the fact is the massive human feeding experiment foisted on the American public decades ago. It has caused untold suffering and needless deaths from the chronic inflammatory diseases it created. This nutritional debacle was chronicled by Garry Taubes a decade ago and updated more recently by Nina Teicholz.

In brief, shortly after World War II, when the medical community became concerned about the rising attack rates of heart diseases, Ancel Keys came forth with an epic epidemiological study indicating that diets high in fat were the cause of the heart problems1. Keys’ thesis became known as the lipid hypothesis and his study became an almost immediate textbook example of how not to conduct an epidemiological study: it did not include all available pertinent data; it selected data that supported its thesis; and it claimed a cause-effect relationship where there was only a mere association2.

Well before the advent of Keys and his associates on the scene, epidemiology had developed as a well-matured discipline of inestimable value in the study and causation of infectious diseases. Despite Keys’ blatant misapplication of epidemiology, his peers supported him wholeheartedly. They validated Keys’ arguments with innumerable studies using equally flawed epidemiology. Only a few scientists called attention to his misuse of accepted epidemiologic methods. In less than two decades, Keys’ lipid hypothesis became official national policy.

The First Use of Epidemiology to Study of Nutritional Diseases

When medical science first discovered, more than 100 years ago, that certain foods could cause disease, epidemiology was essentially the only study tool available for investigating the relationship between a disease and its cause. Speculation about why the nutrition establishment has continued to use flawed epidemiology as the major methodology in nutrition research for the past half century is outside the scope of this post, but the fact that it has been kept as a basic study method is the reason for the failure of the nutritional establishment to produce “reliable, unambiguous knowledge” about diet and disease. The fault is not that of epidemiology itself but rather of its misuse, either deliberately or inadvertently.

What is Epidemiology3?

The definition of epidemiology depends on the dictionary one consults. It is variously defined; however, none of the definitions convey the true essence of the science of epidemiology. A general description of epidemiology is the study of the incidence, prevalence and movement of disease in population groups. In practice, epidemiology uses two branches of science. One is statistical methodology, usually termed biostatistics, to determine the incidence, prevalence, and movement of diseases. The second branch consists of common sense and any or all scientific methods, such as biochemistry and clinical trials, that are required to prove biostatistical associations are valid cause-effect relationships.

Where Did Epidemiology Come From?

The science of disease prevention was nonexistent before a London physician, John Snow, discovered the key to ending the cholera epidemics that were common in nine­teenth century England and in many other countries. In 1836, more than 50,000 people died of cholera in England. It was about that time that Snow began to suspect that filth was some­how connected with cholera; however, there was no general agreement about the cause, and there was no cure. This was the dawn of systematic study of infectious disease epidemics.

Dr. Snow, unable to help his patients who were dying from cholera, decided to interview the 93 families in his neighborhood that had been hit by cholera and compare the details of their eating and living habits with the eating and living habits of families in the same neighborhood that had no cholera (an important point). He made careful notes about each family.

Studying these notes, he found no common link with the type of food eaten or the ages of the cholera victims. But he did find that, unlike the general population, workers at a local brewery and inmates housed in a nearby jail had no cholera at all. Both of these estab­lish­ments used their own wells for drinking water. Very importantly, he noticed that cholera occurred only in families that got their drinking water from one well; this well had a hand pump set up over an open community well on Broad Street.

Dr. Snow took the time to walk around and look carefully at this well. He saw that hu­man waste was draining from a ditch in the street into the well. Remembering the awful diarrhea suffered by cholera victims, he thought that perhaps the human waste from the sick cholera victims draining into the Broad Street well was somehow connected with the spread of cholera. To prove or disprove his idea, he asked city officials to remove the handle from the Broad Street Pump so that local people would have to use a different source of water. In a short time, new cases of cholera stopped occurring among people who used the Broad Street Pump. Snow had rightly concluded that the water from this well was the cause of the disease4, p.412.

Three powerful lessons have come from Snow’s work. The first lesson is that common sense and good judgment are as important as highly sophisticated methodology in discovering the cause of a disease. The second lesson is that the suspected cause must be proven by some valid means to be labeled as the actual cause. For Snow this was removing the pump handle, preventing use of the water from the Broad Street well, and observing that new cases of cholera stopped occurring. The third lesson is that when the cause or the source is accurately known, the disease can be prevented absolutely by eliminating the cause or by cutting the connection between the source of the infection and the human population.

Formalization of Epidemiology

The model for a sound epidemiological study was established in 1876 by German physician Robert Koch3. In essence, Koch’s postulates set the criteria for deter­mining whether there was an association between an organism and a disease, and if there were an association, what specific steps were required to establish whether the association was causal. Thus, Koch’s postulates had the two essential parts of epidemiology: one, the descriptive associ­ation (might the organism have caused the disease?) and two, the analytic causation (could the organism actually trigger the disease?). Part two is proof analogous to Dr. Snow’s removal of the pump handle. Koch’s postulates became a requirement in the study of epidemics of infectious diseases. Had Koch’s postulates not been ignored during the pellagra epidemic, the fact that the disease was not of infectious origin would have been accepted much sooner.

Unfortunately, Koch’s postulates cannot be applied directly to the study of causes of nutritional diseases because of the differences between pathogens and nutrients in the manner in which they cause disease. However the postulates are useful for helping to illustrate the new and difficult problems faced by scientists who study causes of noninfectious diseases. A proposed translation of Koch’s postulates appropriate for application to research in causes of nutritional diseases was first published in 1984 and is still available for review5, pp195.

What Went Wrong?

There is more than ample material in America’s almost century-long history of nutritional research, regulation, and manipulation to write expository volumes that attempt to answer the question of what went wrong. An in-depth examination of the body of literature on which the government-supported diet/heart hypothesis was founded reveals many scientific missteps. The fact that nutritional science was in its infancy fifty years ago may excuse some of the errors apparent in many of the earlier research efforts, but such a defense cannot be used today for the scientific missteps that are still occurring in research into the relationship between diet and disease.

Innumerable errors occurred throughout the decades in all types of studies that form the data­base for the government sponsored heart-healthy diet. However, the media and popular press, with some justification, have now fixed the blame for the perceived failure of nutritional science on a dysfunctional science and its reliance on observational studies, which is popular jargon for epidemiology.

Although public belief that nutritional science is devoid of scientific knowledge is hyperbole, there is no doubt that continued dependence of nutritional research using only the first part of epidemiology, which can only suggest avenues of investigation and ignoring proof of cause is counterproductive.   Even more destructive to the credibility of epidemiological data is the development in recent years of a way of presenting these data as though they were scientific facts. An extreme example of such guile is what is termed New Epidemiology where associations are actually described as risk factors.

How is Epidemiology Misused?

Probably the most common failure of epidemiological studies is to forget or neglect to “remove the pump handle.” It has been repeated innumerable times in print, with little or no impact on the current status of observational reports, that no association can be claimed as cause without scientific proof. The general public seems totally unaware of this problem and accepts, for example, such warnings as eating red meat will do them in.

Incompetent Selection of Variables: Many of the diet/disease epidemiological studies, both from old studies and contemporary, contain serious defects in design one of which is not including all relevant variables. For example, in numerous studies, only fat consumption of partici­pants was recorded to the exclusion of all other diet components. To their credit, a few of the early day scientists reported that people who consumed high levels of saturated fat also consumed high levels of sugar. Unfortunately, their voices were not heard.

This matter of selection of variables underscores a major problem in the use of epidemiology in the study of nutritional diseases. The modern nutritional diseases are primarily the result of deficiencies, excesses or imbalances among dietary macronutrients and micronutrients rather than deficiency of a single micronutrient, as was the case with the classic nutritional diseases, such as beriberi and pellagra. Because of the great interdependence and complexity of interactions among macronutrients and micronutrients, determination of which and how much of each contributes to a disease process has been difficult; It is because of the need to evaluate a multiplicity of variables that makes the epidemiology of modern nu­tri­tional diseases so complicated and probably so unfruitful.

Faulty Premises:Another flaw in the design of epidemiological studies is the use of a faulty premise as a study objec­tive. A recent example is the Women’s Health Initiative (WHI) Controlled Dietary Mod­ification Trial. It was designed, in accordance with the prevailing opinions of international experts, to study the benefits of a low-fat diet on disease prevention. The study failed because the premise that low fat diets were a benefit to disease prevention was faulty. The study’s failure relegated it to the host of large, long-term prospective studies, conducted through the decades since World War II, that have failed to provide a statistical associ­ation between dietary fat and chronic disease, positively or negatively, despite a tremendous expenditure of time and money.

Confusion Between Cause and Effect: A common failing in many current epidemiology studies is that when an association is found the investigators do not seem to know which is the “cause” and which the “effect.” Even though such a failing does not occur in diet/disease research because of the obvious relationship between diet and disease it is important for health professionals to recognize its existence.

The claim that vacations make one healthy is an example of cause-effect confusion.  A number of years ago, a study that revealed obvious confusion in the minds of its investigators was publicized. The Wall Street Jour­nal, along with a number of other highly-regarded popular publications, reported the dire warning “for your health’s sake, do not cancel a vacation if you have a choice.”

Despite considerable anecdotal evidence that vacations are beneficial to one’s health and well being, such studies do not prove that regular vacationing prolongs life. They only show an association between health and recreational travel. But which is the cause and which the effect? Are people healthy because they take vacations or do they take vacations because they are healthy? People who are in ill health or who just do not feel well are not wont to go on cruises or engage in other vacation travel.

Another area in which there is often cause-effect confusion has to do with exercise and health. Are people healthy because they exercise or do they exercise because they are healthy?

Pseudoscience and Prejudice Still Prevail: Unfortunately, governmental and nutritional scientists charged with assuring the accuracy and adequacy of the official dietary guidelines not only have embraced the original flawed science on which the guidelines were based but also have continued to accept new data from studies that are equally defec­tive.

The tragic aspect of these new studies is not in their design failings, but rather that they are conceived and conducted by highly-regarded scientific institutions, published in scientific journals of unquestioned repute, and fully endorsed by nutritional scientists who have a major responsibility for the official dietary recommendations that are causing the epidemics of the modern nutritional diseases.

Even more egregious is the New Epidemiology, mentioned above, that these same scientists have devised. It is a clever way of absolving themselves from the need to prove that an association is causal, which is expensive, time consuming, and requires competence in the appropriate scientific discipline. This new statistical technique expresses epidemiological data not as the probability of the existence of an association but rather as the risk of the outcome predicted by the association actually occurring.

Computation of risk is only valid for studies in which a causal relationship has already been established. Obviously, if something cannot cause an event, there can be no risk that it will happen.

It is important to note that the wholesale adoption of bogus epidemiology by the nutrition community is the linchpin of the current so-called heart-healthy diet. If sound nutritional science is not able to dislodge the linchpin, the massive human feeding experiment foisted on the American public decades ago will continue to take its toll.

References

  1. Keys, A. Atherosclerosis. Journal of Mount Sinai Hospital. 1953; 20: 118-139.
  2. Yerushalmy J, Hilleboe HE. Fat in the diet and mortality from heart disease: a methodologic note. New York State Journal of Medicine. 1957; 57(14): 2343-2354.
  3. Adapted from Ottoboni A, Ottoboni F. The Modern Nutritional Diseases and How to Prevent Them, Second Edition. Fernley, NV: Vincente Books, 2013.
  4. Porter R. The Greatest Benefit to Mankind. New York, NY: W. W. Norton & Company, Inc., 1997.
  5. Frank P, Ottoboni MA. The Dose Makes the Poison: A Plain Language Guide to Toxicology, 3rd Edition. John Wiley & Sons, Hoboken, NJ. 2011

10 Responses to “Epidemiology: Misunderstood? Misused?”

  1. There are standards for when observational studies imply causality. I had listed a few in a blog post. Midway through writing it, it was pointed out that I was re-inventing the wheel fashioned by Bradford Hill who did the original study on tobacco and lung disease.
    http://wp.me/p16vK0-gz
    A minor point: risk is a probability. Relative risk, for example, between drug and placebo, is the ratio of the probability of the disease under the two conditions. (risk = events/total subjects). For the record, odds is slightly different = events/(total – events). They become roughly the same for low probability. The real problem with risk is it’s reporting as relative risk without clearly talking about the magnitude of the risk to begin with (Hill’s main point…and mine, but he was first…actually common sense was first…Hill said that two).

    Reply
    • Alice and Fred Ottoboni

      Many thanks, Dr. Feinman, for expanding on the subject of risk and for your excellent post on observational studies. You brought back pleasant memories of the Bradford Hill textbook everyone who studied biostatistics in the 60s used. The tobacco-lung cancer study by Doll and Hill was brilliant. Your reminder of the study was most appreciated.

      Sadly, our immediate question is how do prominent epidemiologists justify the conversion of data on associations to statements of risk in the absence of proof of cause.

      Reply
  2. We are all consumers of science now. The points you raise are important to everyone who is serious about their health. Well done.

    Reply
  3. One suspects that we have been, still are, in the “dark ages” of nutritional science. Ignorance, complicity, power/money are the root causes.

    Reply
  4. Rosemary Rinaldi

    What a great blog!!! You brought up so many excellent points. I appreciated your very clear and concise explanation of epidemiology, but more importantly the clever discussion of the errors when the scientists formulate theie conclusions of the studies. Unfortunately, we have developed into a culture that wants information on such a superficial level and short “sound bits” of data. The true scientific experiments and valid research is lost. I echo the points of power and money as the root causes!! Thank you for writing this, very much appreciated.

    Reply
  5. michael goroncy

    Elegantly written post..thank you!
    There are many analogies that are similar to your ‘vacation/health’ example.
    One that comes to mind is, ‘Sick people (arthritis) for example have difficulty exercising.

    I try to switch off from the frustration of Politics, Greed, and Scientific Jealousies/fraud that are a significant part of the Health scene.
    Also the Health Guru’s that mimic each other “8 glasses of water a day” …which somebodies Grandmother guessed at. Which brings to mind, the false trail of Stomach Ulcers being caused by stress…anyone for Tagament?

    State of mind (psychological problems) that come into play and addictions which in part are due to ‘the luck of the draw’ that circumstances have dealt you.
    In my own case….I have become Introspective since having a MI, 27 years ago, which causes me to feel every twinge and strange feeling as a negative, which in turn causes anxiety.

    I try to stop reading the dribble and find ‘serious and honest’ researches.

    Reply
    • Alice and Fred Ottoboni

      Good to hear from you again, Michael. Your attitude sounds positive. After all, you must be doing something right with a 27-year survival after an MI. Your body is in a continuous healing mode, so just keep providing it low carb and balanced omega-6/3s.

      We also think you are wise to direct your reading to serious and honest thinking. Have you read Hayek’s “Fatal Conceit? We each have read it four or five times, and each time we find some new thought to ponder. Hayek does not write about nutrition but about societal and personal interrelationships. His thoughts explain beautifully, among other things, why medical practice and public health functions find areas of conflict.

      Reply
  6. michael goroncy

    Yo! Alice and Fred
    Although I don’t read text books now-days, I have ordered “Fatal Conceit”.
    I basically only read fiction…humour (Alexander McCall Smith )…he has one character that specialises in “Abstract Portuguese Adverbs”….and Crime novels (English
    and Scandinavian).

    I am finally cured from watching TV mindlessly, and restrict it to very few select sports and docs.

    My parents use to watch the 5pm, 6pm, and 7pm News most of their lives, and were non the wiser.

    Good health and love to you both.

    Reply
  7. Trevor Turton

    A brilliant overview of epidemiology, what it is and isn’t. Also of its misuse in the form of New Epidemiology, which has reached epidemic proportions.

    Reply
  8. Thank you, Trevor, for your kind words. We both spent our professional lives as public health scientists; Alice in biochemistry-toxicology and Fred in prevention of diseases of occupations. Thus, we saw firsthand the great rewards of competent epidemiology and the terrible damage that bogus epidemiology can engender.

    The real problem is that the public has so little understanding of the fact that epidemiology can only validate associations. As a result, the public can be convinced so easily by unprincipled investigators that epidemiological findings prove cause and effect.

    Your comment tells us that we may have helped some people understand the problem. Again, many thanks.

    Reply

Leave a Reply