How quickly we “reopen” society depends partly on the day cares, schools and summer camps that parents rely on for child care. Nearly all schools in the United States are closed until August or September; some child-care centers have remained open, and others are scheduled to start back up. The fate of summer camps is ambiguous. Parents, meanwhile, can’t go fully back to work until their children are looked after.

Anxious parents are divided on what these institutions should do. On one side are those who worry that at any large gatherings of children and adults, the coronavirus will spread uncontrolled. On the other are those who are concerned about the mental, physical and academic tolls that school closures are taking on children (and adults); they contend that schools are a unique case, because so few young people appear to get covid-19, or at least its worst symptoms. Perhaps opening schools and child-care centers would be safer than opening adults-only workplaces.

Such a choice certainly involves trade-offs. And the only way to make a good decision is to understand those trade-offs. At present, the data suggests that schools might be one of the least risky kinds of institutions to reopen — and that doing so would have tremendous benefits. And we’re lucky it’s only May: Summer-camp season offers the chance to test the idea that bringing young people together may not spur a significant spread of the virus. School administrators could draw on the resulting data as they contemplate a broader reopening.

Opening up the economy in an unconstrained way — as President Trump and some governors seem eager to do — will increase cases of covid-19, possibly by a lot. When adults begin working closely together again, many of them will get infected; see, for example, the extremely high infection rate at meatpacking plants. But there is increasing data suggesting that children are relatively unaffected by the virus. This doesn’t mean they cannot get sick, and there have been a few reports of kids becoming very ill with an inflammatory disorder that may be linked to covid-19. But the incidence of serious or fatal illness among children in the United States and elsewhere is far less than that from, for example, the seasonal flu.

Children, it is true, could be spreaders of the coronavirus even if they don’t show symptoms. But the preliminary evidence on kids’ role in viral spread is, frankly, encouraging. One compelling set of data comes from Australia. Although it’s not yet peer-reviewed, the study, by the staff of that country’s National Center for Immunization Research and Surveillance, analyzes data from 15 primary and high schools in the state of New South Wales from March to mid-April. Across these schools, nine staff members and nine students had covid-19 and were in close contact with hundreds of other people (specifically, 735 students and 128 staffers). Yet according to the researchers, no staff members or teachers developed covid-19. One primary-school student and one high school student may have contacted covid-19 from these initial cases — a very low spread rate. 

In general, data from Italy, Iceland and China shows low infection rates in kids. In one Italian town that performed extensive testing, researchers found evidence of coronavirus infection in 2.6 percent of 2,812 people, but none of the 217 children under 10 were infected, and only 1.2 percent of people age 11 to 20 (three out of 250). If kids are infected at low rates, then spread in schools is likely to be limited. 

Of course, schools do not contain only children. Adults work there, and they are more susceptible. Even if kids do not spread the coronavirus — the rosiest interpretation of the Australia study — adults may infect one another. There is a lot we can do to curb the spread (limiting interactions among teachers, for example), but there will undeniably be some covid-19 transmission in schools when they open.

These conclusions are based on small-scale early data. There are also some analyses that use epidemiological models to try to evaluate the role of schools. One, published in April in the Lancet, uses data from other epidemics (notably SARS and the seasonal flu) and from the coronavirus outbreak to evaluate the possible role of school closures in curbing covid-19. It concludes that they have a relatively small impact, with very high societal costs. The model suggested that school closures alone prevent only about 2 to 4 percent of total deaths, much less than other social distancing interventions, such as forcing nonessential workers to stay home. Given those findings, the authors floated the idea of opening schools at least for the children of health-care and other essential workers.

A more recent analysis in Science uses current and past survey data on the frequency of contacts among people in Wuhan and Shanghai, China. The researchers combine this data with a model that takes into account lower infection rates among children — but assumes similar rates across age groups. They then conclude that schools may be important: that closures could reduce the peak incidence of infections by 40 to 60 percent and delay the pandemic. The assumption that children and adults are equally infectious, however, does not seem supported by current data on the coronavirus, as other scholars have pointed out. More generally, when we have direct data on infectivity, as in the Australian study, it doesn’t make sense to give equal weight to models that do not draw on such data. And so far, the data is reassuring.  

On the other side of the ledger, there are compelling economic and even public health reasons to consider opening child care. For essential workers, child care will help them keep their jobs. It is also crucial for lower-income parents who cannot afford to forgo work outside the home. Parents who are now telecommuting are finding it’s increasingly untenable to balance child care, tutoring and professional duties.

Lots of kids rely on schools and camps for meals: Thirty million children in the United States get free or reduced-price school meals each weekday, and governmental and charitable replacements, which require families to travel to distribution sites (and are not available every day), are inadequate substitutes. New research suggests that hunger is on the rise.

Lengthy social isolation harms mental health, too, and the burdens on both adults and children are mounting. Specialists foresee a rise in mental health disorders among children and teenagers who don’t attend school; children in abusive households face especially grave risks.

When kids fall behind academically or drop out, there can be long-term consequences for work opportunities and for health, among other things. We know that academic “summer melt” is an issue — kids forget what they’ve learned during the break — especially for low-income students and students of color. Summer this year started in April. More-educated adults have fewer health problems than less-educated adults, meaning we may see long-term health effects of the pandemic that have nothing to do with covid-19.

It is also inevitable that, without better alternatives, people will seek their own child-care solutions. Unsanctioned neighborhood summer camps will crop up. It will be difficult to police this, and such gatherings may be more hazardous than official camps and schools that have to adhere to covid-related safety regulations, such as frequent cleaning.

Trade-offs of this sort are difficult to discuss. If we reopen schools, some people will get sick, and there may be deaths. Mehmet Oz, the celebrity doctor, got in trouble recently for saying on Fox News that he’d be willing to make such a deal. He somewhat incorrectly summarized the Lancet study by saying that “the opening of schools may only cost us 2 to 3 percent in terms of total mortality” and that it “might be a trade-off some folks would consider.” (He later apologized, saying he “misspoke.”) His tone and choice of words were awkward — “We need our mojo back,” he said — but he wasn’t entirely wrong to recognize the need to balance harms.

So what should we do? First, we should stop arguing in black-and-white terms: After looking at the evidence, reasonable people will disagree about whether to open camps and schools — just as, if camps open, some people will choose to send their children and some will not. Those who want to keep schools closed are not, by definition, taking covid-19 more seriously than those who come to the opposite conclusion.

Second, we might consider using summer camp as a “soft open” for schools in the fall. (The University of Pennsylvania oncologist and ethicist Ezekiel Emanuel proposed a version of this idea last month.) Camps are more limited in their scope than schools and use younger people as instructors and staffers. They’re also easier to shut down than schools if the virus spreads too widely. We could open camps with restrictions on group sizes and with extensive health monitoring. Kids with any cold symptoms should not attend. There should be repeated testing of counselors and other adults throughout the session. In some places, there is a shortage of coronavirus tests, but given the economic and social stakes of opening schools, spending federal money to procure more for this experiment is an obvious call.

If we can do this, we solve some of the summer problem parents are dreading and we get a road map for how we can open schools safely. We’d get a better sense of the risks. If we run a month of summer camps and see limited virus spread, that will tell us something very important.

Twitter: @ProfEmilyOster