John Beigel’s problem was that not enough Covid-19 people died. Not that he wanted them to, of course. It’s just that, as dire as the pandemic has been, it hasn’t killed as many people as it seemed. And Beigel, a researcher at the National Institute of Allergy and Infectious Diseases, was designing a large study to see if an experimental antiviral drug called remdesivir would work against the disease. He and his team needed the right “end point”: what they could count, convert to data, and analyze.
Mortality is a great end point. It’s right there in the name.
“But when we were thinking about the endpoints, we thought that a mortality study would need 3,000 or 4,000 people,” says Beigel, associate director of clinical research in the NIAID Division of Microbiology and Infectious Diseases. Those high numbers would give his team enough “events,” meaning deaths, to get a statistically significant measure of the drug at work. And they didn’t have enough time to enroll as many people in their study in the United States. “We thought it was important to do the study, get a clinically meaningful endpoint, without taking the time we would have to do a mortality study,” he says.
This was in February, before anyone really knew anything about Covid-19. So Beigel’s team tried a different tactic, one familiar to scientists and regulatory agencies like the Food and Drug Administration (which issued the Emergency Use Authorization under which people are studying the drug). Participants in the study would obtain scores, every day, calculated on what is called an ordinal scale. Healthy and discharged from the hospital, you get a score of 1. Dead, you get an 8. The other numbers were for everything else, like if the person has to be admitted to the hospital or needs oxygen or has to go to a hospital . mechanical fan.
Then they found another problem. “When we wrote the protocol the first time, the chosen end point was the ordinal score on day 15,” says Beigel. “That’s something we’ve used before for influenza studies, so we knew the FDA would agree to that, and it’s something that really matters to the subject.” That’s a good end point: it is not just a statistical entity. It is long enough to show an effect on many diseases, and has clear clinical relevance.
“In March we started hearing reports that the course of the disease could be much longer and that there were people in the hospital for three weeks, up to four weeks,” says Beigel. “What happens if the recovery is much later than the 15th? You could actually make a significant difference, but you wouldn’t show it. “
So, Beigel’s team changed their end point: ordinal score on day 28. They hadn’t seen their data yet when they made the change. That would have been unethical, pursuing statistical significance by falsifying its methods. But get ahead of the data? Kosher, but they knew it would be controversial. “It raised suspicions for our study,” says Beigel. “If this were something well known, like the flu, and we switched midway through the study, it would be really suspicious. For Covid, we have not seen something like this. “
In a study published in late May in The New England Journal of Medicine (and previously seen during a press conference at the White House), the team concluded that the patients who took remdesivir recovered in an average time of 11 days, while the people who received a placebo took a median of 15. It was enough To add remdesivir to the US standard of care, soon: the first drug identified as having a beneficial effect on the new disease.
.
