In a hastily-called press conference Sunday evening, President Donald Trump and the heads of the Food and Drug Administration and Department of Health and Human Services announced what the president had touted as a "breakthrough" in treating the pandemic disease Covid-19. That breakthrough turned out to be a change in the bureaucratic status of convalescent plasma, a treatment made from the blood of people who have recovered from the disease-one that most people can already get, that at least 70,000 people have already received (and probably more like 100,000), and that despite the president's assurances currently lacks solid evidence for its effectiveness.

"It is proven to reduce mortality by 35 percent. It's a tremendous number," President Trump said at the press conference. "Based on this data, the FDA has made the determination the treatment is safe and very effective." That last part isn't actually true yet-without large-scale randomized controlled trials, which haven't happened yet, no one actually knows that for sure. Yet the agency has issued what's called an Emergency Use Authorization, not quite an approval of the treatment but still giving the all-clear to physicians who want to administer the frothy, yellow liquid.

Based on the president's comments, and those of FDA head Stephen Hahn and HHS director Alex Azar, the data that tipped plasma into getting this Emergency Use Authorization came from a nationwide program to give people wider access to convalescent plasma, led by researchers at the Mayo Clinic. That program was not a large, randomized, controlled clinical trial, which is typically the standard for determining whether a treatment is safe and effective. That "expanded access program" did publish some outcomes in a preprint two weeks ago. It still hasn't been peer-reviewed, but the researchers did keep track of the levels of antibodies, disease-fighting molecules, in the plasma administered to 35,000 people. It turned out that those people who received plasma earlier, and with the highest levels of antibodies, did have an increased survival rate. But the 35 percent number wasn't for everyone who got plasma, just a specific statistical slice of the tens of thousands who received plasma sooner and with higher antibody levels. And even though the people who set up the expanded access program wanted to run concurrent randomized, controlled trials, that's not what this was.

The timing of the announcement makes the whole thing even weirder. The president, who during this week's Republican Party National Convention will presumably accept his party's nomination to run for a second term, is under increasing pressure to show some progress in the fight against Covid-19-a disease from which nearly 175,000 Americans have died, with little sign of abatement. The only randomized trials that have published data on plasma so far have been small and, while promising, inconclusive.

Technically, that's fine for an EUA. The criteria for issuing one really just demand reasonable safety and the possibility of usefulness. The agency still maintains that it'll take actual randomized trials to make convalescent plasma into standard care. And it's not clear that the authorization does anything that the expanded-access program wasn't already doing, making sure anyone who wanted plasma could get it.

"It's sufficient for an Emergency Use Authorization in the sense that most EUAs make no claim of benefit. They simply authorize utilization during the emergency, and when the emergency is over they are automatically cancelled," says Peter Bach, director of the Center for Health Policy and Outcomes at Memorial Sloan Kettering Cancer Center. "That said, this is one of the more frustrating examples, where we have had every opportunity to do proper randomized trials on convalescent plasma. The sole excuse offered back in February and March was, 'We don't have time,' and it's now nearly September."