In common, I suspect, with most members of my professional clan- specialists in Preventive Medicine- I have a respect bordering on reverence for the U.S. Preventive Services Task Force.

 The Task Force meticulously weighs up relevant evidence to establish letter-grade guidance for clinical preventive services such as cancer screening, the assessment of various biomarkers, clinical counseling, and immunization. Those recommendations carry the veritable weight of gospel for practitioners, translating more or less directly into the standard of care; and exert a powerful influence on policy makers and the prevailing standards of health insurance coverage as well. Task Force recommendations on a wide expanse of pertinent topics are at the disposal of any with Internet access.

Conceding all such due respect, I am nonetheless very much aware of important Task Force limitations. Three come immediately to mind; they are very far from inconsequential; and all three pertain to the recent guidance regarding nutrient supplements that – yet again- spawned commentary and media coverage far and wide. We will return to the matter of those supplements; first, though, those limitations.

To begin, the Task Force enjoys a luxury clinicians and patients lack: they can conclude not to conclude anything. In their letter grade system, this is an “I,” standing for “insufficient.” Reviewing evidence of only a particular type and quality (an emphasis is placed on randomized controlled trials, or RCTs), this conclusion is reached often and translates to:  “there is insufficient evidence in quality, quantity, or both to recommend for or against a given practice.”

Of- to my count- 87 Task Force recommendations on distinct topics, an “I” conclusion appears 46 times. Were clinicians to function accordingly- telling well over half of all patients “there is insufficient evidence to decide whether to treat you or not…” – clinical care would come completely undone. Functioning despite residual uncertainty is a critical element in clinical care and might be filed under “judgment,” or as preferred, the “art” of medicine.

Second, because of the inflexible requirements for a specific standard of evidence, the Task Force is ill-equipped to distinguish absence of evidence from evidence of absence (the former means evidence is lacking; the latter means evidence proves a lack of effect). Much of what is perfectly clear to the average human would evoke an “I” recommendation from the Task Force for want of randomized trials. That list would include: children running with scissors; looking both ways before crossing a busy street; and playing with fire, to name a few. I have looked and found no RCTs nor meta-analyses on any of these topics. Examples, all, of “absence of evidence,” i.e., a lack of studies; but surely not examples of “evidence of absence,” i.e., proof that running with flaming scissors into traffic is perfectly safe. 

Third, the Task Force generates answers only to the questions explicitly posed, and those questions can be bad. As there are no good answers to bad questions, this is a serious problem. The badness in question traverses a host of liabilities from a “one-size-fits-all” coarseness where nuance is required; failure to align a given question with the requisite sources of a reliable answer; failure to account for the influence and potency of countervailing forces; failure to distinguish real-world from randomized trial circumstances; and more. Each of these is a lengthy disquisition in its own right, so I am hoping I may invoke my prior work and ask you to take my word for it.

Of note, a bad question does not tend to get any better with repetition, nor does the likelihood of a good answer improve. With no disrespect to the Task Force, “Do nutrient supplements prevent cardiovascular disease or cancer, or reduce all-cause mortality?” is a rather bad question now, just as it was the last time they posed it.  I took care of patients for over 30 years and recall not a single instance where one of them took a multivitamin to prevent a heart attack, cancer, or death. As noted in the Task Force’s own paper on the topic, most people- including my patients- take a multivitamin “for overall health and wellness and to fill nutrient gaps in the diet.”

To begin the indictment, then, if not of the Task Force conclusion about nutrient supplements, at least of the widespread commentary over-interpreting it: the question the Task Force asked is largely irrelevant to prevailing practice. Select nutrients have been studied for the prevention of heart disease (e.g., vitamin E) or cancer (e.g., Beta-carotene), with negative results- although even here there are subtleties related to the spectrum of disease largely absent from the public discourse. But the blending of a judgment about this with commentary on nutrient supplementation in general for very different purposes is about as nuanced as the infamous failure to differentiate baby from bathwater.

All three liabilities noted above impact the Task Force position on nutrient supplementation. First, the conclusion regarding multivitamin use for prevention of major disease or death is an “I,” meaning insufficient evidence to conclude for or against. Since these are not the reasons people take a multivitamin, the lack of clarity is actually a surprising endorsement. Multivitamins might confer such a benefit, in other words, you just can’t count on it. 

If multivitamins do confer a mortality benefit, why is it so hard to prove? Consider that any such benefit, to qualify as “evidence,” would need to be adjusted for all other relevant factors: lifestyle behaviors, dietary pattern, toxic exposures, and medication use to name a few. Imagine how small the residual effect of a multivitamin on mortality risk would likely be when lifestyle and medication were reliably matched.

Let’s say, for instance, that in reality (to be clear, I am making this up)- multivitamins, even when added to state-of-the-art pharmacotherapy (and, ideally, lifestyle medicine, too) reduce mortality from heart attack by one case in 60,000 per year. In a clinical trial, if 60,000 people at risk not merely for a heart attack within a year, but dying of a heart attack within a year, were randomized to multi-vitamin, and another 60,000 were randomized to matching placebo, the mortality difference would be something like 1, and thus- statistically invisible. This trial of 120,000 very high-risk participants would not be nearly large enough, and a year not nearly long enough, to find meaningful “evidence.”

If this effect were real, however, since well over 600,000 people die of myocardial infarction each year in the United States alone, it would mean more than 10 lives saved annually- by means of a multivitamin. I am not saying this effect is real; I am simply pointing out that even if it were, it would be very hard to capture in any clinical trial.

So, what about nutrient supplements for health benefits? Well, in some cases the evidence has been so clear that what started as “supplementation” migrated to “fortification” of the food supply to ensure that all would benefit. Fortification is really just supplementation by alternate means, and the nutrients addressed that way should not be backed out of the dialogue. Grains are fortified with folate to prevent neural tube defects; many products are fortified with vitamin D to prevent the deficiency that would otherwise prevail. These and other exceptions suffice to prove the rule: nutrient supplements most certainly can confer a decisive and important health benefit.

As for multivitamins, their likely effect is subject to the paramount, and routinely neglected, consideration governing supplementation: supplemental to what? The very name denotes that all supplements are meant to be added to something else. That something else- the nutrients derived from food- is almost never known, due to historical challenges in assessing dietary intake. This stage of my career is devoted to popularizing a solution to that very problem, and making diet the vital sign it deserves to be. When nutrient intake from food can be reliably captured, the personalized case for or against any given nutrient supplement is far better informed. 

Stated bluntly, the likelihood of benefit rises when the “fix” pertains to something known to be broken (we do all learn in medical school: “if it ain’t broke, don’t fix it”). Boosting the level of a nutrient that is low (not at all uncommon for many nutrients in the context of a ‘typical’ American diet) is far more apt to confer benefit than providing more of what is already ample. By way of analogy, the very best antihypertensive medications would fail to show benefit if only ever tested in people with normal blood pressure at the start.

I believe nutrient supplementation should be personalized accordingly. Supplements should indeed be supplemental to, never substitutes for, the benefits of eating well. Knowledge of dietary intake to inform both should be universal; diet should be a vital sign.

As for the Task Force recommendation, it should be understood in context, and with due consideration of the question asked and answered (again). You should not rely on a multivitamin to prevent heart attack, cancer, or death. Benefits related to the more realistic motivations of energy, well-being, and so on - all fall outside the scope of the Task Force probe. 

Medicine, in practice, is both art and science for the good reason that much decision-making is required where data are sparse. Absence of evidence is not evidence of absence. Nutrient supplementation is a broad expanse with real estate enough for both baby and bathwater. Be judicious, and strive to distinguish between them.