I’m an epidemiologist and a dad. Here’s why I think schools should reopen.
Covid-19 is upending our lives and forcing us to make complex decisions with little information and conflicting guidance from authorities. Summer, typically the season of staying up late and popsicles in the park, offers no escape. Many of us are already turning to the fall, and the fate of schools.
What will we do with our kids? Can we really send them back to school? If we keep them at home, will they forget how to read? If we send them to school, what might be the consequence? We are living a nightmare, but this is where we are. The choices are high-stakes and plagued by uncertainty. Even thinking about them makes me sweat.
I am the father of three girls ages 16, 13, and 10, and like every parent in America, I am worried about the fall.
I’m also an infectious disease doctor and epidemiologist, and have spent the past four months drinking from the fire hose of Covid-19 science, designing infection control policies for my hospital, and caring for patients on the front line. I serve on the reopening committee for my synagogue and for my school district. I consult for businesses as they reopen.
I have a first-row seat to the coronavirus pandemic, both as a parent and as a professional. In both of those roles, I hear the same questions, repeated with mounting urgency: “Are our kids going to be safe?” “Are our teachers going to be safe?” “Will kids bring Covid-19 home to our family?” “Will opening schools lead to a second wave and lockdown?” “What are the risks of not reopening?”
I have spent time reviewing the data and seeking answers to the challenging questions we face. Having the knowledge to make your own assessment, however, need not be a position of professional privilege. With this short primer, I hope you can add your voice to the debate and advocate for yourself, your family, and your community. The good news is, we can hope to send kids back to school in the fall, but there is a lot of work to do.
Are our kids going to be safe?
If any of us is ever going to send our kids to school again, we need a clear answer. Fortunately, I think we have one, at least for the children. Children are less likely than adults to be infected with Covid-19. There are multiple ways to study this question, and all the approaches arrive at this same conclusion.
First, when we look at public health reporting, children under the age of 18 make up only 2 percent of cases in the US, even though they represent 22 percent of the total population. Similar studies in Chicago and Massachusetts found that children make up fewer Covid-19 cases than expected, as have studies in Italy, South Korea, and Iceland. For me, that is a lot of similar results for this to be a fluke. When one study in one location produces a finding, it is notable. When five studies from five different settings find the same thing, it is compelling.
One reason case counts may be lower among kids than would be expected is that we did close our schools in March. Maybe we protected our kids by keeping them out of harm’s way. But if we send them back to school this fall, will they still enjoy protected status from the coronavirus?
One way to study this question is to estimate the “attack rate” of the disease — that is, the proportion of people exposed who become infected. Multiple studies from China investigated the attack rate among people living in a house with someone who is infected. They found that only about 4 to 5 percent of kids developed an active infection. In comparison, about 17 to 20 percent of adults became infected after exposure.
To be fair, data in the US is more concerning. In New York state, 57 percent of people living with a Covid-infected person developed an infection. It is hard to take reassurance from that fact. But even with such a high attack rate, children were still less likely to develop an infection and there was a gradient over ages, a sort of “dose effect” for age.
Finally, even in the worst-case scenario, in which a child does contract Covid-19, the outcomes of the disease are less severe in younger people than among older adults. In one analysis of more than 550 confirmed cases among children under age 18 in China, Italy, and Spain, only nine people (1.6 percent) had severe or critical disease. In another study, approximately 5 percent (one out of 20) developed symptoms that required hospitalization, but only 0.6 percent required intensive care. In comparison, a recent Centers for Disease Control and Prevention report indicates that among those ages 60 to 69 who have the coronavirus, 22 percent require hospitalization and 4 percent require intensive care.
Are teachers going to be safe?
There is far less data specifically on teachers and staff than on kids. One study in France is reassuring. In that investigation of 541 students and 46 teachers, there were no documented transmission events from students to teachers. However, while many of us immediately think of the risk to teachers from exposure in the classroom, we may not consider the additional risk that teachers face in break rooms and staff meetings.
Working in the hospital, I have personally seen that staff have a difficult time maintaining personal protection at all times. Doctors and nurses tend to let down their guard when they are away from patients and during breaks. Masks come down, people eat snacks in potentially unsafe spaces, and social distancing lessens.
The same will likely be true in schools. The potential risk to teachers, therefore, goes beyond the classroom. Staff risk in schools likely looks similar to the risk of any adult working in a crowded indoor environment during the pandemic. School opening plans must consider teacher safety in addition to the well-being of students.
Will my kids bring Covid-19 home to our family?
For most parents, the next question after the safety of their kids will be their own safety and that of loved ones in the house. Even if the kids are all right, could they bring the coronavirus home?
Here, again, the data appears reassuring. One large review of over 700 scientific publications found that children accounted for only a small fraction of Covid-19 cases, and that they were rarely the first case in a cluster of infections in a household. For example, in China, only 5 percent of household clusters were found to have a child as the index case. Similarly, in Switzerland and Holland, children accounted for only 8 percent of household transmission clusters.
Unfortunately, the US numbers make me a little less certain. In a Chicago study of 15 households with available data, 73 percent of infected children contracted the virus from an adult. However, that means that 27 percent of infections were child-to-child, which is substantially more than 5 to 8 percent.
Still, the Chicago study only examined 15 households, and adult-to-child transmission remained far more common than child-to-child or child-to-adult.
Will someone in America contract Covid-19 from their sick child? Yes. Should I structure my life around such a rare occurrence? I do not think so.
Will opening schools lead to a second wave and more lockdowns?
We have reached the most challenging question to answer and one that is a holy grail for Covid-19 epidemiologists. I want to give you the plain answer here — we do not know.
An objective summary of the evidence in hand suggests that schools will play little role in sustaining the pandemic. A recent review of 210 transmission clusters around the world found that only eight of them (3.8 percent) involved school transmission. Case studies of outbreak investigations in Ireland, France, and Australia demonstrate almost zero cases of in-school transmission.
Modeling studies demonstrate no clear role of in-school transmission in explaining current Covid-19 epidemiology. All of this data tells us that despite our gut instincts and parental anxiety, schools will likely be okay this fall.
But the story does not end there. First, there are examples of in-school outbreaks that did force a second shutdown. Israel is an example.
Israel reopened schools with limited class sizes in early May and lifted class size restrictions on May 17. By June 3, they had to reclose after a major outbreak. The largest outbreak was 116 students and 14 teachers at one school. Per NPR, one child tested positive without symptoms and the school decided to quarantine the grade. Next, a child in a different grade tested positive and they closed the school.
At that time, they discovered that they already had more than 100 cases. It is not certain that all of those children were infected in the school, but the story is concerning and it raises the bar on monitoring our schools.
The data that’s available is mixed. If a person (or school district) wants to tell you that schools play little role in transmission, then ask them how their district is different from Israel’s. Why can an outbreak happen in that setting but not yours? Perhaps there is a reason, but until someone can give you a good one, be skeptical.
What are the risks of not reopening?
A discussion of school closures that focuses only on Covid-19 and not at all on education is incomplete. There are real risks to keeping our children at home. In fact, the risks of staying home are in many ways clearer than the risks of returning to school.
One study using statistical models projects major losses in math performance if we continue with remote learning until 2021. Perhaps more compelling than statistics, however, is some simple common sense.
On any given day, it is hard to point to the loss of learning from home. At the same time, we all agree that education is essential. If we keep our kids home for another school year, they will have missed 12 percent of their total education. I cannot identify the specific losses from that much absence, but I am confident there is a cost to missing that much school. For perspective, missing 12 percent of school time is the same as missing 22 days of school in a single year.
Further, the losses will not be equal. The “Covid-19 slide” will likely be greatest among the socially vulnerable, such as children with learning disabilities and those whose situation at home is not conducive to homeschooling.
We must also acknowledge that the losses will hit people of color much harder than those who are white. Further, school officials account for approximately 20 percent of formal reports of child abuse and domestic violence. Without school-based counselors and social workers, these concerns may not be investigated.
All of these harms weighed on the American Academy of Pediatrics’ guidance that school reopening plans start with the goal of having students be physically present in schools.
What should we do?
A great mentor of mine, Milton Weinstein at Harvard, is generally credited as being the person who introduced the field of medicine to the concept of rigorous decision-science. The central question to all decision-science is: “What should we do, given that we have imperfect information?”
Milt is fond of the expression “a decision has to be made.” His wisdom has never been more pertinent than it is today. We have to make a decision. There is no choice to do nothing, because either way — go to school or learn remotely — we are making a decision.
Unfortunately for all of us, we are making a decision with significant uncertainty about all the risks involved. Fortunately, this is not the first time that people have been forced to make decisions with uncertainty. There are approaches to making uncertain decisions in a way that maximizes the chances of a good outcome and minimizes the harm if the outcome is poor.
You’ve likely heard of one of them: hedging your bets. When multibillion-dollar investment funds make a choice to invest, they recognize that they could be wrong. They do not make all-in versus out decisions. Instead, they hedge their bets. They may think that the newest beach toy is destined for greatness, but just in case of a rainy summer, they also invest in umbrellas.
When I look across all the data, I see an uncertain decision. First, I propose that the balance of data we have now suggests that we need to try to open schools in the fall. The risks of reopening are uncertain; the harm of staying home is clear.
If your school district cites the data above to you that “schools are safe,” ask your school board: What is the plan beyond reopening? What if we are wrong? How will your district know that things are going well (or not well)? Don’t let the conversation stop at “data suggests that schools are safe.” Don’t let the plan stop with “symptomatic people should call their doctor.”
If we are going to open safely in the fall, we must have the capacity to know — quickly — when an outbreak occurs. Israel is an important cautionary tale. When Israel closed down its schools again, it had only identified two school-based cases, yet in the end it discovered that more than 100 students had been infected.
To do this well, and to do it safely, we must have school-based Covid-19 symptom screening, testing, contact tracing, and isolation. “School-based testing” does not mean that the test themselves must occur in school buildings. “School-based testing” means that students and teachers can easily access a test by contacting the school, and that the results of those tests are sent directly to the school district in real time.
That seems straightforward, but it is not. The community does not yet have adequate testing, contact tracing, or isolation. Schools currently have nothing.
It requires building new capacity in schools for testing and contact tracing. It requires a budget. It requires a formal plan. Ideally, that budget should come from the federal government and be directed to states and ultimately school districts, as part of a national Covid-19 testing strategy. Realistically, given the lack of any such national plan, the funds need to come from individual states.
Building such infrastructure comes at a cost and many districts are already facing budget shortfalls. Districts that rely only on their existing testing infrastructure will not have the real-time information they need to make good decisions. Imagine a child has a fever and cough in October and is told by the school to call the doctor for a Covid-19 test. Results are typically returned in two days to the doctor’s office. After another day (or two), the data might make it to the school district. So it will take at least four to five days for the district to have any information.
We need testing within the school system to shorten the delay at every step of the process and reduce the turnaround time for the test to only a day. With that kind of time resolution, we can increase awareness of the situation at our schools, along with the ability to react appropriately. Without it, we are flying blind and gambling with the health of our children, teachers, and community.
Ultimately, when I look at the decision about school as both a father and a scientist, I see a difficult decision that must be made despite uncertainty. The risks of opening are uncertain, but the benefits are clear. We need to try to reopen.
We have been wrong before about Covid-19. In March, the epidemiology world was quite confident that transmission could not occur before a person develops symptoms. Three months later, there is consensus that asymptomatic people were likely one of the main drivers of the pandemic. In March, the CDC and the US surgeon general told the public that masks play no role in controlling the spread of the disease. Now we see masks as a central component of our reopening strategies.
We could be wrong about schools, but we cannot afford to wait to find out for certain. We need school-based Covid-19 symptom screening, testing, contact tracing, and isolation. Opening without a plan to test is irresponsible and a gamble with our children’s health.
Benjamin P. Linas is an associate professor of epidemiology and an infectious disease physician at Boston University School of Medicine. Find him on Twitter @BenjaminLinas.
