It’s a simple and critical fact of medical science that women’s bodies differ from men’s, and those differences can affect everything from diagnosis to treatment. For instance, women experiencing a heart attack often present different symptoms than men. Because doctors considered men’s symptoms the norm, they for years misdiagnosed heart conditions in women.

In recent years, women’s health advocates have rightly pushed for greater attention to such issues in an attempt to oppose sexist misperceptions. But the rush to identify such distinctions may also be leading researchers astray in some contexts, potentially putting patients at risk.

There’s nowhere that these challenges are clearer than with the sleep drug zolpidem, the basis of the popular sleep aid Ambien. In 2013, the Food and Drug Administration claimed that women should receive a lower dose of zolpidem than men because they clear the drug from their system more slowly and are more impaired by it at the 8-hour mark when people wake up and begin their day. Advocates cheered this dictate, and zolpidem became a flagship example of how sex differences can influence health, as featured on Freakonomics and in the New York Times.

Even at the time, however, there was reason for skepticism. Critics, including ourselves, cautioned that science did not clearly support sex-specific dosing for zolpidem. Now, a new review of the science by Tufts University’s David Greenblatt, a leading expert in how zolpidem affects the body, argues that the FDA’s “recommended dosage in women is not based on actual studies showing any safety benefit.” Reduced dosage for women, he warns, “may in fact increase the public health risk due to undertreatment of insomnia in women, and the consequent impaired automobile operation as a result of sleep deprivation.”

Greenblatt’s review of the evidence finds no difference between men and women in levels of zolpidem still in the body or impairment after eight hours. While women clear zolpidem from their system more slowly than men, part of this is related to body weight — rather than sex — and the remaining difference is not enough to matter at the critical eight-hour mark. Furthermore, while women on zolpidem sometimes rate themselves as more impaired than men, brain wave measurements show no differences in impairment between men and women. Among individuals, there is no consistent relationship between zolpidem dose and impairment. The two most impaired people in Greenblatt’s study were a high-dose woman and a low-dose man.

In short, there isn’t any scientific evidence supporting sex-based dosage for zolpidem. The drug clears quickly in both men and women and does not lead to greater impairment in women.

Zolpidem offers a valuable angle for thinking about sex difference science and women’s health. When the National Institutes of Health (NIH) introduced a mandate in 2016 requiring sex difference analysis in all research, whether in cells, zebra fish or mice, it cited the case of zolpidem as Exhibit A: “Drugs such as zolpidem, used to treat insomnia, require different dosing in women and men,” stated NIH director James Collins and NIH Office of Women’s Health Research director Janine Clayton, justifying the policy in the journal Nature. The aim of the new policy is to advance women’s and men’s heath equity by considering sex differences at all levels of preclinical research. Specifically, NIH contends that studying sex differences in laboratory models will help address higher rates of adverse drug outcomes in women.

But NIH’s call for a broad, non-hypothesis-driven dragnet for sex differences in biomedical research can lead to findings that, even if statistically significant, have few if any relationships with actual men and women’s health. In the case of zolpidem, a rat model showing enzymatic differences between males and females in metabolizing zolpidem helped convince the medical community that the sex difference in humans was real. As Greenblatt shows, it turns out that this model does not apply to humans, which use different enzymes to metabolize this class of drugs.

It is overwhelmingly clear that the FDA must now reverse its sex-specific dosage guidelines for zolpidem, and the sooner the better for American women, a quarter of whom suffer from sleep disorders. But the implications are broader. With NIH now requiring analysis of sex in all $30 billion of the medical research it funds each year, we need rigorous standards for documentation of clinically relevant sex differences, including taking into account intersecting variables like body size and composition. And while considering sex can be important in basic biology, sex difference studies in cells and rodents should not be applied to women without validation in human studies. Ignoring sex differences relevant to drug response can be a public health peril, but zolpidem reminds that so, too, is overstating them.