Where Does Science End and Policy Begin?
Vitamin D
4 Jul 2008 in Regulatory Science, Regulatory Policy
Washington Post staff writer Rob Stein reports on a controversy over whether the federal government's recommendations for Vitamin D intake are adequate. The story illustrates the perils of delegating policy decisions to scientists.
In a Page One story, Stein says new data suggest that the government's recommended intake of Vitamin D is too low.
A flurry of recent research indicating that Vitamin D may have a dizzying array of health benefits has reignited an intense debate over whether federal guidelines for the "sunshine vitamin" are outdated, leaving millions unnecessarily vulnerable to cancer, heart disease, diabetes and other ailments.
The studies have produced evidence that low levels of Vitamin D make men more likely to have heart attacks, breast and colon cancer victims less likely to survive, kidney disease victims more likely to die, and children more likely to develop diabetes. Two other studies suggested that higher Vitamin D levels reduce the risk of dying prematurely from any cause.
The studies have produced evidence that low levels of Vitamin D make men more likely to have heart attacks, breast and colon cancer victims less likely to survive, kidney disease victims more likely to die, and children more likely to develop diabetes. Two other studies suggested that higher Vitamin D levels reduce the risk of dying prematurely from any cause.
Both inside and outside the government, there is resistance to drawing the inference that its recommended intake ought to be increased.
But many leading experts caution that it remains premature for people to start taking large doses of Vitamin D. While the new research is provocative, experts argue that the benefits remain far from proven. Vitamin D can be toxic at high doses, and some studies suggest it could increase the risk for some health problems, experts say. No one knows what consequences might emerge from exposing millions of people to megadoses of the vitamin for long periods.
"The data are intriguing and serve as, no pun intended, food for further fruitful research," said Mary Frances Picciano, at the Office of Dietary Supplements of the National Institutes of Health. "But beyond that, the data are just not solid enough to make any new recommendations. We have to be cautious."
The problem for the government has three aspects: uncertainty and variability, the one-size-fits-all nature of any set of federal guidelines, and the challenge of balancing costs and benefits when any federally recommended value will be wrong for more people than for whom it is right."The data are intriguing and serve as, no pun intended, food for further fruitful research," said Mary Frances Picciano, at the Office of Dietary Supplements of the National Institutes of Health. "But beyond that, the data are just not solid enough to make any new recommendations. We have to be cautious."
UNCERTAINTY AND VARIABILITY
Uncertainty is the term used to describe scientific or technical ignorance. If a "correct" value exists, we will never know what it is. The best we can do is estimate it. It is generally agreed that our estimates should be unbiased -- that is, we should not rig the game so that we knowingly under- or overestimate the true (but unknown) value. The idea is that after producing an unbiased estimate of the "correct" value, we can then decide how much precaution is appropriate in deciding what action to take. If the risks of underdosing on Vitamin D are equivalent to the risks of overdosing, then we are probably best off choosing the unbiased estimate as our level of Vitamin D intake. However, if overdosing is much more dangerous, then it is both reasonable and appropriate to err on the side of too little Vitamin D. (This perspective is not universally shared. The US Environmental Protection Agency, for example, has a longstanding policy and practice of deliberately overestimating risks. If Vitamin D were a chemical within its jurisdiction, EPA would almost certainly ignore the new data Stein mentions and focus solely on its potential adverse effects.)
Variability is the term used to describe the fact that people are not identical, and that the effects of any substance (such as Vitamin D) are not the same for everyone. Population variability can be substantial, and this means that the "correct" amount of Vitamin D for you may be very incorrect for me.
FEDERAL RECOMMENDATIONS ARE ONE-SIZE-FITS-ALL
In general, when the federal government (or any other authoritative body) sets a recommendation -- whether it is a recommended minimum daily requirement for a nutrient or a "safe" level of exposure to an environmental contaminant -- it is explicitly or implicitly saying to the public that this single value is "correct" for everyone. In short, it is ignoring population variability. The more variability there is in the population, the greater will be the proportion of people for whom the recommended value is wrong.
The public does not know that these one-size-fits-all recommendations are wrong for most people because the authority issuing the recommendation typically chooses not to disclose it. If this detail is disclosed in the fine print, the disclosure is dropped in subsequent communications. The Food and Drug Administration's regulation requiring Nutrition Facts labels is perhaps the example to which people are most familiar.This label, which must be included on virtually every food package, tells the consumer exactly what proportion of "good things" (e.g., vitamins and minerals) and "bad things" (e.g., fat, cholesterol, sodium, and carbohydrates) is contained in the package. Notice that FDA's label assumes everyone in the population is the same.
BALANCING COSTS AND BENEFITS
Sometimes, the government makes an effort to balance costs and benefits when making one-size-fits-all recommendations. To do this, one would need to know (or estimate) the risks associated with under- and overdosing. These risks may be symmetrical in magnitude, in which case unbiased estimates of the "correct" amount may minimize the sum of risks of error. But they might be very asymmetrical. The risks of getting too little of a nutrient or too much of a contaminant might be greater. Benefit-cost analysis allows the government to be explicit and careful about its decisions rather than blithely intuitive or ideological.
Based on Stein's report, a crusading ideology appears to drive some of the competing scientists in the Vitamin D war:
The current clash is the latest in a long, often unusually bitter debate. Some skeptics question whether funding by the tanning, milk and vitamin industries is biasing some advocates. Frustrated proponents accuse skeptics of clinging to outdated medical dogma.
"It feels kind of ridiculous working in this field sometimes," said Reinhold Vieth, a professor of nutritional sciences and pathobiology at the University of Toronto. "Every week, I get interviewed about the next important publication about Vitamin D. But this field remains mired in the muck."
Vieth is one of a small but vocal cadre of researchers pushing doctors and patients to stop waiting for new official guidelines. Physicians should routinely test their patients for Vitamin D deficiencies, and more people -- especially African Americans -- should take supplements and increase their exposure to the sun, they say.
"It feels kind of ridiculous working in this field sometimes," said Reinhold Vieth, a professor of nutritional sciences and pathobiology at the University of Toronto. "Every week, I get interviewed about the next important publication about Vitamin D. But this field remains mired in the muck."
Vieth is one of a small but vocal cadre of researchers pushing doctors and patients to stop waiting for new official guidelines. Physicians should routinely test their patients for Vitamin D deficiencies, and more people -- especially African Americans -- should take supplements and increase their exposure to the sun, they say.
Some proponents of doubling Vitamin D intake make dramatic predictions that are testable, but only if population-wide changes in behavior occur:
"The first thing we'd see is a reduction by 80 percent in the incidence of Type 1 diabetes," said Cedric Garland, a professor of family and preventive medicine at the University of California at San Diego. "The next thing we'd see is a reduction by about 75 percent of all invasive cancers combined, as well as similar reductions in colon cancer and breast cancer, and probably about a 25 percent reduction in ovarian cancer."
Achieving those changes requires allowing scientists to make policy and providing them the means to coerce people into making the required behavior changes. These scientists are not uncertain, they do not appear to take population variability seriously, and they are convinced they know what one-size-fits-all prescription is correct. Of course, if the health benefits they predict are real, benefit-cost analysis would be a sufficient tool to demonstrate it. Thus, there is no need to give scientists the authority to make policy decisions.
