Cheng decided to dig deeper. For months, he and colleagues pieced together the data from thousands of patients who were seen at his medical center. In results published last month, melatonin continued to stand out. People taking it had significantly lower odds of developing COVID-19, much less dying of it. Other researchers noticed similar patterns. In October, a study at Columbia University found that intubated patients had better rates of survival if they received melatonin. When President Donald Trump was flown to Walter Reed National Military Medical Center for COVID-19 treatment, his doctors prescribed—in addition to a plethora of other experimental therapies—melatonin.

Eight clinical trials are currently ongoing, around the world, to see if these melatonin correlations bear out. Few other treatments are receiving so much research attention. If melatonin actually proves to help people, it would be the cheapest and most readily accessible medicine to counter COVID-19. Unlike experimental drugs such as remdesivir and antibody cocktails, melatonin is widely available in the United States as an over-the-counter dietary supplement. People could start taking it immediately.

Yet Cheng emphasizes that he’s not recommending that. Like any substance capable of slowing the central nervous system, melatonin is not a trifling addition to the body’s chemistry. Its apparent benefit to COVID-19 patients could simply be a spurious correlation—or, perhaps, a signal alerting us to something else that is actually improving people’s outcomes. Cheng thinks that might be the case. He and others suggest that the real issue at play may not be melatonin at all, but the function it most famously controls: sleep.