For the second straight day, a bright light shines through the gloom of a darkening COVID outlook this fall. Yesterday the good news was that fears of schools becoming mass incubators of coronavirus appear to be overblown. Positivity rates across a sample of tens of thousands of students and staff in NYC are minuscule. Maybe we can have safe in-class instruction this fall.

Today’s news is even better. As cases explode in Europe and climb towards a new peak in the U.S., a tantalizing mystery has emerged: Why aren’t we seeing the same crushing death tolls in hot spots that we saw in the spring? Deaths *are* rising in Europe from their summer lows (upwards of 150 daily on bad days in the UK and France recently) but they’re nowhere near the gory levels of April even though both countries are setting records in the number of daily cases. The trends in cases and deaths here in the U.S. also diverged awhile ago:


The pessimistic theory of why deaths in western countries haven’t approached their spring peaks is that the number of infections during the spring was much, much greater than the meager testing available at the time was able to detect. Deaths still track cases under this theory; the spring death toll was a simple function of there having been a gigantic “hidden” epidemic at the time.

The more optimistic theory is that science has simply caught up to some degree. People who would have died if they were infected in March are surviving now because doctors have gotten better, and better equipped, to treat the disease. According to NPR, two new studies suggest that the optimistic theory is correct. COVID has become less lethal.

[One] study, which was of a single health system, finds that mortality has dropped among hospitalized patients by nearly 18 points since the pandemic began. Patients in the study had a 25.6% chance of dying at the start of the pandemic; they now have a 7.6% chance

So have death rates dropped because of improvements in treatments? Or is it because of the change in who’s getting sick? To find out, Horwitz and her colleagues looked at over 5,000 hospitalizations in the NYU Langone Health system between March and August. They adjusted for factors including age and other diseases, like diabetes, to rule out the possibility that the numbers had dropped only because younger, healthier people were getting diagnosed. They found that death rates dropped for all groups, even older patients by 18 percentage points on average

“I would classify this as a silver lining to what has been quite a hard time for many people,” says Bilal Mateen, a data science fellow at the Alan Turing Institute in the U.K. He has conducted his own research of 21,000 hospitalized cases in England, which also found a similarly sharp drop in the death rate. The work, which will soon appear in the journal Critical Care Medicine and was released earlier in preprint, shows an unadjusted drop in death rates among hospitalized patients of around 20 percentage points since the worst days of the pandemic.

Mateen says drops are clear across ages, underlying conditions and racial groups.

There’s no silver-bullet explanation for the decline, researchers suspect. Partly it’s a matter of doctors getting better at anticipating dangerous complications, like blood clots. Partly it’s a matter of younger, hardier people composing a larger percentage of patients. Partly it’s a function of mask-wearing, one researcher speculated, believing that masks might filter out some particles before they’re inhaled, leading to smaller viral loads and less severe cases. Doctors did stress, though, that even a 7.6 percent chance of death is unusually high for a disease, and certainly higher than the flu. The fact that so many COVID survivors suffer complications afterward, sometimes long-term, also distinguishes it from more familiar respiratory diseases. Coronavirus may be less likely to kill you now but it’s still capable of making life miserable for weeks or months.

“But what about antibody treatments?” you say. Regeneron’s treatment saved Trump, didn’t it? Well … maybe not. The data from two new randomized clinical trials, one from Italy and other from France, found no major benefit from the antibody treatment tocilizumab. And even if Regeneron’s and Eli Lilly’s treatments do help, they’ll be in short supply initially and pose logistical difficulties. The drugs are administered intravenously, so you’d need to get to a clinic or ER to be treated. And because state agencies that are consumed with other COVID business are being tasked to coordinate distribution, it’s an open question as to how efficient the dispersal will be.

Just like it’s also an open question whether Regeneron’s drug did anything for the president. He keeps talking about how great he felt after getting it but that could easily be explained by the dexamethasone he was getting, not the antibody drug. Steroids (in)famously can induce feelings of strength and even invincibility.

While we’re busy tempering expectations about the efficacy of antibody drugs, we should probably take a moment for a reality check on the vaccine too. Lord knows it’ll help reduce infections, but the dream scenario where we all get the shot, whip off our masks, and have a big indoor kegger where we’re all breathing in each other’s faces probably isn’t happening:

If most of the infections are in the placebo group—say 26 out of 32—that would suggest the vaccine is at least 76 percent effective. That’d be pretty good. But scientists have cautioned that a COVID-19 vaccine might be less effective than we’d like, based on how vaccines against respiratory viruses tend to work. The FDA has set a bar of at least 50 percent efficacy for a COVID-19 vaccine. It’ll take longer and more cases for trials to reach a conclusion if vaccine efficacy is on the lower side. So if the first interim results are a little disappointing, that “doesn’t mean this is a failed vaccine,” Lowe says. “We’re just going to keep on rolling.” We’ll have a better idea of efficacy once we’ve seen how the vaccine performs in more people.

Conversely, it shouldn’t come as a shock if some of these vaccine candidates do turn out to be ineffective. The development process from Phase 1 to 2 to 3 has gone very smoothly so far. But, in general, more than 90 percent of drugs and treatments fail, and close to 50 percent of them fail in Phase 3. Lowe says he expects COVID-19 vaccine candidates to do much better because scientists are building on research into MERS and SARS, two related coronaviruses.

I haven’t yet digested the possibility that masks could become a regular part of American life for the next several years or longer, especially in winter, because the vaccine just isn’t effective enough to put COVID down for the count. *Maybe* the combination of the vaccine, masks, and improved treatments will get us to the point relatively quickly where the disease isn’t much worse than the flu even if you do get it, but we may be playing the waiting game for awhile. Which wouldn’t be an entirely bad thing: Even if only, say, a third of Americans made a point of wearing masks semi-regularly in winter, that would break many chains of transmission both for coronavirus and for lesser respiratory diseases like influenza. We might see an abiding perennial drop in flu hospitalizations.

Here’s Scott Gottlieb warning that we may be only a week or so behind Europe in experiencing rampant community spread here in the U.S. Europe’s present was our near-future back in the spring; Gottlieb’s speculating that that will be true in the fall as well, especially with case counts in the U.S. already climbing. If you’re thinking that’s not a big deal given the lower mortality rates now, note this bit from the NPR piece: “Mateen says that his data strongly suggest that keeping hospitals below their maximum capacity also helps to increase survival rates. When cases surge and hospitals fill up, ‘staff are stretched, mistakes are made, it’s no one’s fault — it’s that the system isn’t built to operate near 100%,’ he says.” The more infected people there are, the more hospital beds will be occupied, and thus the more preventable deaths are likely to result.