Worried Sick: A Prescription for Health in an Overtreated America

Four: You Are Not What You Eat

In 2005 it was considered reprehensible to feed your child butter; today, it’s reprehensible to feed your child margarine. You must be thin, but not too thin. Bran is fast fading as a salutary “must” that lowers your cholesterol and burnishes your colon. Now eating some fish will save your life thanks to omega-3 content, unless the mercury content does you in first. A diet low in carbohydrate is good for you, or not. A diet low in fat is good for you, or not. Red meats are bad, except maybe a little is okay; white meats are good unless toxic substances lurk in the juices. The print and broadcast media are committed to dishing up the scare of the week and the cure of the weekend. The public is insatiable. Nearly all of this has an authoritative source: some governmental agency, a disease-specific, notfor -profit association or foundation, the public relations department of some medical school, or an earnest and smiling scientist. Major medical journals (JAMA, New England Journal of Medicine, and others) have public relations departments that forewarn the press of an upcoming newsworthy article, many of which are displayed online before they are in print. Nearly all of this advantages some producer or purveyor and spells harder times for some other. And we, the gullible, swallow any drivel or babble that we are told relates to health. We are far less gullible when it comes to the pronouncements by our politicians, or by those hawking other wares. We listen, with a prepared mind, for nuanced duplicity when it comes to our pocketbook or our sense of community. We accept only the pronouncements that pass this filter, and we do so tentatively. Caveat emptor is our way of life. We are all too aware that the pronouncements of politicians are qualitative and value laden, if not as corrupted as the doublespeak of “used-car salesmen.” Mention our health and we are patsies. You Are Not What You Eat chapter four 58 | you are not what you eat Literature such as we reviewed in the previous chapter provides fodder for my cynicism. Epidemiology often seems to have lost its way. In fact, epidemiology may lose its credibility given its tendency to declare last month’s “good” this month’s “evil.” Epidemiology has evolved from a science of the epidemic to its contemporary mission of probing for associations between less-dramatic exposures and less-cataclysmic health effects. To wit, modern epidemiology has enlightened us as to the association between smoking and lung cancer, inhalation of certain types of asbestos fibers and mesathelioma, alcohol consumption and motor vehicle accidents. Giddy with such successes, modern epidemiology is probing for associations between exposures that are difficult to define and health effects that clearly have multiple causal associations. Epidemiology has acquired near religious beliefs that statistical modeling is a match for any degree of variability and uncertainty in the definition of exposure , and probability theory and its imperious p-values are a match for discerning tiny health effects. In chapter 3 we examined the UKPDS and West of Scotland pravastatin trials as examples of the pitfalls in defining exposure (Who has type 2 diabetes? Who’s cholesterol is high?) and both for issues in whether the outcome was meaningful. These were randomized controlled trials in which the investigators had some control over the exposure (the drugs), and still the results are equivocal and the authors’ interpretations unconvincing. Nonetheless, these randomized controlled trials stand head and shoulders above observational studies. Epidemiology falls flat on its face when it tries to tease minor exposures and minor health effects by observing all the glorious variability that is humanity. Biases and confounders lie in wait for any epidemiologist trying to tease minor events and influences out of the complexity of life. Condemnation lies in waiting for any epidemiologist who is willing to analyze data with the goal of supporting a preconceived notion (“data dredging ”) rather than testing a hypothesis. Most disconcerting is the fact that much of this epidemiology is generated in trials that are supported by companies that are purveying, or wish to purvey, the pharmaceutical under study. Recent analyses of the published results of such industry-sponsored trials document systematic biases that favor the drug or the foodstuff when compared to trials of the same drug sponsored by other sources. The reader of Worried Sick will take little for granted in the future. [18.218.234.83] Project MUSE (2024-04-25 21...