Physicians have recommended vitamin D supplements to their patients for a decade, with good reason: dozens of studies have shown a correlation between high intake of vitamin D—far higher than most people would get in a typical diet and from exposure to the sun—and lower rates of chronic diseases, such as cancer and type 1 diabetes. So when the Institute of Medicine, which advises the government on health policy, concluded in November that vitamin D supplements were unnecessary for most Americans and potentially harmful, patients were understandably confused.

The issue exposes a rift among experts over what constitutes valid proof when it comes to nutrition and could affect medical advice on many other supplements. On the one hand are scientists who insist that the only acceptable standard is the randomized clinical trial, which often compares the effects of a medical intervention, such as high intake of vitamin D, with those of a placebo. The scientists who reviewed the vitamin D findings fall heavily into this camp: trials “typically provide the highest level of scientific evidence relevant for dietary reference intake development,” they wrote. Their report set intake levels based only on clinical trial data.

The institute panel, however, discarded a raft of observational studies, in which researchers compare the health of populations who take vitamin D supplements with those who do not. In theory, such epidemiological studies are inferior to clinical studies because they rely on observations out in the real world, where it is impossible to control for the variables scientists seek to understand. Researchers compensate for the lack of control by using large sample sizes—some vitamin D studies track 50,000 people—and applying statistical techniques. According to these studies, high levels of vitamin D are generally beneficial.

In the aftermath of the institute report, some physicians are now taking potshots at clinical studies. In nutrition, they say, true placebo groups are hard to maintain—how do you prevent people in a control group from, say, picking up extra vitamin D from sunlight and food, which can lead to underestimating the vitamin’s benefits? It is also tough to single out the effect of one vitamin or mineral from others, because many work in tandem. “It is wrong-headed thinking that the only kind of evidence that is reliable is a randomized controlled trial,” says Jeffrey Blumberg, a Tufts University pharmacologist.

The next chapter in this debate may come in the spring, when the Endocrine Society releases its own vitamin D guidelines. The organization now recommends higher blood levels of the vitamin than the institute suggested—30 nanograms per milliliter as opposed to 20—which would require supplements. Stay tuned.