Media Alarmism Is Making It Difficult To Assess School Reopening
Kids do not catch or spread or suffer from coronavirus at the same rate as adults, no matter what your newspaper is telling you this week.
A headline Monday afternoon in The New York Times' "Coronavirus Briefing" email newsletter snapped my head to attention, and not just because of the old-timey classic rock pun: "The kids are not all right."
Like tens of millions of parents in this country, I have school-aged offspring and a non-insignificant amount of anxiety over just how they will be spending their time (and mine) a month from now. If, as the Times asserted in the first sentence of its write-up, "new research [is] reflecting just how vulnerable children can be to the coronavirus," that could significantly alter my family's decision-making process. I need to use this news.
Then I started clicking on the links. And lo! There was this troubling Wall Street Journal headline: "Latest Research Points to Children Carrying, Transmitting Coronavirus." Given that the preliminary research as of three weeks ago suggested that elementary school-aged children, such as my 5-year-old, rarely contract, get sick from, or transmit COVID-19—so rarely, in fact, that a leading British epidemiologist was quoted in the Times of London as saying there hadn't yet been a single case of a teacher catching the virus from a student anywhere in the world—then such a development would count as a major reversal.
"The latest research," the Journal went on, "indicates children may be carriers just as much as adults." Huge, if true.
Yet the only research presented in the article supporting the notion that children contract the virus "just as much" as adults came from a single overnight summer camp in Georgia, at which kids (though not staff) didn't wear masks, and participated in singing and shouting exercises in unventilated rooms. That is not a useful representative sample.
A more helpful survey of the available literature comes not from the world of journalism, but from a website called COVID Explained, run by Brown University economist Emily Oster, Harvard Medical School professor Galit Alter, and others. Their gloss, as of August 3:
There are a few [random sample] studies….An early one was conducted in Iceland. Researchers there tested about 13,000 random people, including 848 kids. Among the whole population, 0.8% of people (so, almost 1 percent) tested positive for COVID-19. Among children under 10, though, there were no positive cases. This difference was very unlikely to occur by chance.
Data from a single town in Italy which did very wide-spread screening leads us to the same conclusion: kids are much less likely than older people to be infected.
Another study uses mathematical models to estimate that people under age 20 are half as susceptible to infection as adults over 20….
Altogether this evidence suggests that kids are less likely to be infected with coronavirus than adults.
The Oster/Alter tentative interpretations may well be wrong, and the amount of random sampling is still maddeningly small, but their thoroughness of study-picking and cautiousness of conclusion inspire more confidence than 99 percent of the elite media I consume on this topic. There is an irritating tendency to elevate scary-sounding anecdotes over data, to collapse the definition of "children" into a single glob rather than as cohorts with importantly different characteristics, and to treat the science and the politics of this issue as national when it is actually a geographically variant and locally administered story.
Journalistic outlets have been pinning their school-reopening pessimism this week on a new American Academy of Pediatrics report that says the number of cumulative COVID-19 cases among children—the definition of which varies by state, from between 0 and 14 (Florida) to between 0 and 24 (Alabama)—grew by 97,000 in the last two weeks of July, from 242,000 to 339,000. That is an increase of 40 percent over a short period of time.
Since mid-April, the share of coronavirus cases accounted for by children has marched steadily higher in each two-week reporting period, from 2 percent to the current level of 8.8 percent. What explains the steady increase? Many suspect the recent re-opening of day care facilities and summer camps (on which more below). It is also true that as testing has become more widespread, it has detected more asymptomatic carriers (who previously would not have been tested), which is one reason why the case fatality rate has continued to fall.
Children's share of coronavirus-related hospitalizations has remained disproportionately low, going from 0.8 percent in late May to 1.4 percent now. (The overall hospitalization rate of kids who test positive has been cut nearly in half, from 3.8 percent to 2 percent.) Happiest of all, just 86 people classified as "children" by the states have died, which means kids account for just 0.06 percent of all coronavirus deaths, and their case fatality rate is just 0.03 percent.
Before your eyes glaze over at all those single-digit percentages, remember this: Children are 23.1 percent of the United States population. That means that they account for 1 out of every 4.3 people, 1 out of every 11.4 coronavirus cases, 1 out of every 50 hospitalizations, and 1 out of every 3,333 deaths.
So it is flatly untrue that "children may be carriers just as much as adults." What's more, treating children as an undifferentiated mass is particularly unhelpful when making parental decisions and public policy, particularly as it relates to the pressing question of whether and how to reopen schools.
The Centers for Disease Control breaks out national coronavirus statistics by several different age cohorts, producing this startling number: Only 20 kids between the ages of 5 and 14 (from kindergarten through middle school) have died from coronavirus since February 1. This group was two and a half times more likely to die from the flu, and four times more likely to die from pneumonia. Among the more droplet-swapping demo of 15-to-18-year-olds, only 31 have died. An additional 25 kids under 5 years old have also perished.
All of which means children so far have not been measurably "vulnerable" (to use the New York Times adjective) to coronavirus. But what about their ability to transmit the potentially deadly disease even when asymptomatic?
Here's how The New Yorker treated this urgent scientific question: "Some studies have found that young children are not as likely as adults to spread COVID-19. But summer camps on Long Island and in Georgia, and schools internationally, have had virus outbreaks." That's one way of assessing things, I suppose.
The Wall Street Journal, whose headline suggests that the "latest research points to children…transmitting coronavirus," cites just one study outside of the Georgia sleepaway camp example: a big contract tracing effort in South Korea, which found (in the WSJ's wording) that "children between 10 and 19 years old transmitted the virus within their own households at the same rate as adults of certain ages," but also that "children under the age of 10 didn't spread the virus as much."
Here's how COVID Explained parses that same study's data (the site also examines two smaller surveys):
It appears that outside of household contacts, transmission from kids is really low (note that in South Korea they wear masks and are socially distancing, which contributes to the low non-household contact transmissions. Wearing a mask is really important!). Among young children, transmission even within a household is low: if your little kid has COVID-19, there seems to be only about a 5% chance you'll get it. Older kids do seem to transmit the virus to household contacts as efficiently as adults yet are still less likely to have the virus and less likely to transmit to non-household contacts.
Sorting through the available (and rapidly growing, if still woefully insufficient) underlying science is the critical precursor to having a rational discussion about school reopenings. (Which as of this writing is already off the table in 17 of the country's 20 largest school districts.)
But we also have two other very relevant, if not always referenced, datasets: school reopenings worldwide, and day care reopenings at home. What do they tell us?
The University of Washington's Department of Global Health five weeks ago surveyed 15 countries that have reopened schools, and concluded: "So far, countries that reopened schools after reducing infection levels—and imposed requirements like physical distancing and limits on class sizes—have not seen a surge in coronavirus cases….There have not yet been rigorous scientific studies on the potential for school-based spread, but a smattering of case reports, most of them not yet peer-reviewed, bolster the notion that it is not inevitably a high risk."
The report does highlight a case that comes up frequently in these conversations: Israel. "But there have been school-based outbreaks in countries with higher community infection levels and countries that apparently eased safety guidelines too soon," the study cautioned. "In Israel, the virus infected more than 200 students and staff after schools reopened in early May and lifted limits on class size a few weeks later."
The European Centre for Disease Control and Prevention has a more recent rundown of European Union (E.U.) school reopenings, concluding, among other things, that "If appropriate physical distancing and hygiene measures are applied, schools are unlikely to be more effective propagating environments than other occupational or leisure settings with similar densities of people."
Those in favor of reopening schools should not skim over the conditionality in both studies' conclusions. The "if"s are critical.
When it comes to day cares and summer camps, data are considerably harder to find, which increases journalistic reliance on negative anecdotes. "The media coverage," Emily Oster wrote on July 30, "can give one the impression that it is impossible to operate a child care setting without outbreaks. These data contradict this."
What data is that? Well, Oster and her team generated a voluntary, crowdsourced self-reporting database among child care centers, then also examined media reporting on cases, and sifted through state-reported surveys. The results are admittedly uneven and incomplete, and Oster herself warns that "we should all draw our own conclusions." Nevertheless, this is hers:
On the bad news side, the fact that we see some outbreaks here tells me that it is not realistic to expect no school outbreaks at all. In some cases, the large outbreaks cited above seem clearly linked to behaviors we'd hope schools would avoid, but it is simply not realistic to expect no clusters of cases to emerge. Some of these could be large.
But on the positive side: many child care centers and camps, even large ones and even in high prevalence areas, are operating without significant outbreaks. They are dealing with cases without having them turn into clusters….
It's easy to forget the denominator, but estimates suggest in the range of 5,000 summer camps are operating this summer, including perhaps 1,500 overnight camps. The number of child care centers in the U.S. is in the hundreds of thousands. Yes, it is concerning that there are 14 child care locations in North Carolina with clusters, but this is out of a total of about 6,500 locations.
I am grateful that Oster has gone to such lengths to provide at least the beginning of some context to the individual reports of kids contracting and spreading coronavirus. But I am irritated at how rare it is to see journalists treating this important question with similar care.
For what my opinions are worth, I remain in favor of at least re-opening elementary schools—fully, not part-time, as New York City is planning—in places where the positive-test rate has been consistently below 3 percent. (That's lower than New York state's threshold of 5 percent.) And while I'm worried about the higher transmission rates of teens (my eldest daughter is 12, and I've seen how these people act), I'm leaning toward being in favor of at least partial reopening of middle schools and high schools in lower-threshold places as well. Perhaps having a gradualist approach there will make it easier to process new information and adapt to unwelcome shocks.
But that's where both the discourse and the policy, I fear, are leaving us ill-prepared. Because of poor decision-making and management at all levels of government, coronavirus tests are taking more time, not less, to turn around. This is unacceptable and should have been the focus of federal policy in particular since before day one. Testing delays, more than any other factor, could grind school reopenings to a halt, just as soon as the first cluster appears.
Meanwhile, the way we talk about this stuff matters, too. It is literally impossible in this brain-damaged political moment to enter into a public conversation about possible school reopenings without people leaping into bad-faith accusations about wanting to kill children, or tip elections, or God knows what else. Journalism, particularly in places where science abuts policy, is difficult and prone to error (mine included) in the best of circumstances; a deadly pandemic during an unusually polarized presidential election is about the worst imaginable backdrop. But this is precisely where we need the best-faith attempts at acquiring and transmitting knowledge, including admitting where the available knowledge is inadequate, rather than chasing after cheap culture-war dopamine hits.
My plea, as a parent, to journalists more scientifically literate than I: Please do better.
