Study Finds In-person School Safe, As Long As Hospitalizations From COVID-19 Are Already Rare

Jan 12, 2021

A new study from the National Center for Research on Education Access and Choice found no evidence that opening school buildings increased COVID-19 hospitalizations in places where hospitalizations from the virus were uncommon.

It’s the first study of its kind and includes data from nearly every U.S. county and parish. Education reporter Aubri Juhasz spoke with study co-author Douglas Harris. He’s a professor and chair of economics at Tulane University. The following is an edited transcript of their conversation.

Aubri Juhasz: Before schools reopened, there was speculation that they would serve as super spreader sites. Based on your research, did we see those fears validated?

Douglas Harris: Well, yes and no. I think we see that the spread of the virus from schools does perhaps seem to happen when there's a lot of the virus in the community to spread. In other words, when the number, in our case of hospitalization, seems very high, which indicates that there's a lot of COVID in the community, you know, then opening up schools may be risky. But when the number of hospitalizations is low, we actually don't see any evidence at all that opening up the schools is risky.

Do we have an idea at what point reopening schools or keeping schools open becomes more risky?

Yeah, so in our data it looks like 36 to 44 hospitalizations per 100,000 in the population per week. What does that mean? In New Orleans, we have a little less than 400,000 people. So we'd be looking at about 160 or so hospitalizations per week, new hospitalizations for COVID as kind of a benchmark. It's a rough guideline, so we don't want to make it sound like a magic number, but in the data it does seem to be a difference maker.

Note: Between Jan. 4 and Jan. 10, the Louisiana Department of Health reported 373 COVID-19 hospitalizations for the greater New Orleans area, which includes Orleans Parish, Jefferson Parish, Plaquemines Parish and St. Bernard Parish.

I appreciate the distinction because I think that’s something everybody has been wrestling with in terms of there are no magic numbers. We see schools setting all sorts of thresholds and then having to back off them because it’s far more complicated than just looking at one or two numbers.

I wanted to ask you about the idea of looking at hospitalizations rather than positive test rates or the raw number of coronavirus cases. Why did you decide to focus on hospitalizations?

Really, all the prior studies had focused on the COVID positivity rate and there are a few problems with that. One is that intuitively we know that very few people are actually tested, especially in schools. Very few children have been tested, few teachers have been tested. So the idea of looking at the positivity rate and thinking that we're really learning about spread in schools when almost nobody is being tested — it just didn’t seem like a very useful approach.

Another problem is that when schools reopen, they might actually change the amount of testing that happens. So if students go back to school, teachers are back in school, they're probably more likely to be tested. And that actually throws off the numbers then because you may increase the positivity rate or decrease the positivity rate just because you've changed the amount of testing, not because you've spread the virus.

A third reason for focusing on hospitalizations is that's really what we care about. We care about people actually getting sick. And so the positivity rate isn't useful because a very small percentage of those people actually get sick, whereas the hospitalizations were actually directly measuring people getting quite sick and having to go to the hospital.

Your research touches on the idea of tradeoffs and you make the argument that in some ways having schools open can improve aspects of public health and even help control the spread of the virus. Can you tell me a little bit more about that?

I think what we have to think about is what we're comparing opening schools to. So we think about schools being closed, students being at home, it could be that when students are home, that they're not as strict about their social distancing behavior. They might not worry about who they interact with, whether they have friends over and so on, because they're not going to be going to school where it might be risky and where they think they might spread it. They may actually be more lax when they're home.

Whereas when you send them to school, you know, the schools are very much on top of this. They're in touch with public health officials. There are strict rules around mask wearing, social distancing and so on. So in some cases, you're sending students into schools that are actually more vigilant about the policies. And the families may be more vigilant, too, because they know that if they're sending their child to school they're going to be in contact with others and that they want to make sure that they're treating their classmates and teachers safely.

How confident are you in the results of this study?

We’re very confident about the finding that there seems to be no spread from opening schools when the hospitalization rate is low. I say that because we tried it a lot of different ways. You know, we had two different data sets for hospitalizations. We had three different data sources for the school reopenings. And we had five different ways of actually estimating the effects. No matter what we do, we never see any evidence that opening the schools spreads the virus and increases hospitalizations when the rate of hospitalization is low to start with.

We have less confidence when we look at the high hospitalization rate area. Sometimes it looks like it increases the virus. Sometimes it actually looks like it decreases the amount of the virus and hospitalization. That's why we say in the paper that we have inconclusive results when we look at those high hospitalization counties.

Based on the level of confidence you have in your findings, what advice would you have for school leaders when they look at these results? How can they be using this to inform the decisions they’re making?

We don't come down strongly one way or the other about which schools should open and which schools should close and when. You know, that's really a job for policymakers. But I think it is an opportunity to think through the decision and frame the decision in a useful way.

One thing to consider are these public health effects, these direct effects that we've been studying, and hospitalizations. Another consideration is how schooling will actually look when you open up in-person. Because really what we mean when we say in-person now is some combination of in-person and remote, because when you open up in-person now, the schools are certainly going to give families the option of having their children learn remotely.

In that case, the school really has to operate in two different ways. At the same time, they have to both serve the students in-person and serve students remotely, which is really difficult. I've got a high schooler and I can see it myself, how hard that is. I'm an instructor at Tulane University and I can see it here, too, that it's just really hard to manage both of those at the same time. So we're not comparing going back to business as usual versus staying at home because that's not a viable option. We can't go back to the way it was until we've really eradicated the virus. So part of it depends on the ability of the schools to be able to manage all of that simultaneously, be able to manage the in-person and remote instruction at the same time.

Douglas Harris is a professor and chair of economics at Tulane University. He’s also director of the National Center for Research on Education Access and Choice.