A group of researchers, spearheaded by Brown University Professor Emily Oster, have created and made available the most comprehensive databaseon schools and Covid case rates for students and staff since the pandemic started. Her data—covering almost 200,000 kids across 47 states from the last two weeks of September—showed a Covid-19 case rate of 0.13% among students and 0.24% among staff. That’s a shockingly and wonderfully low number. By comparison, the current overall U.S. case rate is 2.6%, an order of magnitude higher.
Other research has shown that hospitalization and fatality rates for school-age children are also extremely low. People 19 and younger account for only 1.2% of Covid-19 hospitalizations in the U.S. during the peak of the pandemic. The Centers for Disease Control and Prevention report that of all Covid-19 deaths up to Oct. 10, only 74 were of children under age 15. During the 2019-20 flu season, the CDC estimates, 434 children under 18 died of the flu. Yet we don’t shut down schools over the flu.
This is from David R. Henderson and Ryan Sullivan, “End the School Shutdown,” Wall Street Journal, October 20 (print edition: October 21).
30 days from now, which is November 20 (the day before my 70th birthday), Â I’ll post the whole thing.
A friend on Facebook asked me about the issue of compulsory schooling. He knows I oppose compulsion. I don’t know my co-author’s view on that and I wanted to stick with issues we agree on. So I didn’t raise it. But my view is that any parents who want to keep their children out of school should be able to do so. I predict that this will be under 10 percent of parents.
READER COMMENTS
Jon Murphy
Oct 21 2020 at 12:04pm
Wow. That’s amazing. One of my big concerns with schools opening up would be that, while young people are quite protected from the disease, older people are less so. Many teachers/professors are older. I was afraid of transmission from student to teacher. But that doesn’t appear to be a major worry.
PJC
Oct 21 2020 at 12:27pm
Two concerns about this editorial.
First, “Her data—covering almost 200,000 kids across 47 states from the last two weeks of September” – because Covid can have an incubation period of up to 14 days, a two-week sample is not compelling evidence.
Second, while you mentioned the risk of spread to teachers, you did not mention spread to family members, which is a major concern about opening schools.
JFA
Oct 22 2020 at 2:38pm
Regarding your first point, most schools in the US start early to mid-August, so using data from the end of September gives enough time to show how the virus is spreading in schools.
Dylan
Oct 22 2020 at 3:22pm
Is that a recent thing? I never went to a school that started before Labor Day (and I lived in multiple states)
robc
Oct 22 2020 at 3:33pm
I started 1st grade in the mid 70s, so if you consider that recent, then yes.
It seems to depend if your area relies heavily on tourism and high school summer employees. If so, they are more likely to start back after Labor Day Weekend.
JFA
Oct 22 2020 at 3:34pm
My statement might have been too broad. Depends on where you are. Most (all?) southern states start in August. Ohio seems to start early. The northeast seems to start a bit later. California is all over the place.
Dylan
Oct 23 2020 at 12:45pm
Fascinating. I was mostly in the PNW and Hawaii, which is I guess one of the later start regions. Belated realization that all the back to school sales that started the week I got out of school weren’t all run by no fun killjoys after all!
Alan Goldhammer
Oct 21 2020 at 2:53pm
Oster’s data is not as robust as it appears with many large school districts not represented. I would note that Boston is shutting their schools today as COVID-19 cases have been rising over the past two weeks when the opening guidelines require them to be level or dropping.
I wonder if given the opportunity how many readers of this blog would volunteer to be teachers right now.
David R. Henderson
Oct 21 2020 at 3:06pm
You wrote:
Do you have a cite for the data?
Vivian Darkbloom
Oct 21 2020 at 3:47pm
I just read this article about schools going remote in Boston. Â However, please note that the guidelines refer to a positivity rate for *the city* and has nothing to do with the Covid positive rates *in schools*. Â The “problem” here seems more likely to be with the city policy rather than the prevalence of the virus in the schools. Â This does nothing to challenge Oster’s data.
https://www.nbcboston.com/news/education/boston-public-schools-shift-to-all-remote-learning-due-to-spike-in-covid-cases/2215661/
Alan Goldhammer
Oct 21 2020 at 5:35pm
Vivian is correct that the schools are going by the community infection rate. It is a difficult call to make in this regard and the major groups that have established guidelines: National Academy of Sciences and the American Academy of Pediatrics both called for decreasing rates of infection. What makes it hard is that kids in K-6 grades do get infections but they are mild. Based on data that I’ve seen, they are carriers of the virus and can infect others, including family members as well as teachers. This can be ameliorated by requiring masks in schools and having HEPA air purifiers in the class room.
Vivian Darkbloom
Oct 22 2020 at 2:55am
Regarding HEPA air filters, I’ve also been thinking lately that this would be a sensible approach to reducing the spread of the virus in certain settings.  But, my first  priority would be in nursing and assisted care homes.  I’m surprised there has not been more discussion (and action) regarding this.  And, I suspect the advantages would go beyond the corona virus.  I understand there is a program in Germany, which seems to have done very well and early on in protecting residents of those homes, to do just that.
As a small aside, your comment reminded me of a discussion I’ve had on Pierre Lemieux’s recent post “Imports as a ‘Drag on the Economy'”. Â I could easily have taken those three sentences of yours and jumped to the conclusion that you made an “obvious error” in linking the school closure to an increase in cases in schools. Â But, I read your comment carefully more than once and concluded that you didn’t say that and that it would be unfair to simply imply that is what you meant.
Vivian Darkbloom
Oct 22 2020 at 8:45am
Sorry for the afterthought, but this (HEPA filters, etc) is, I think, exactly the sort of targeted action that the authors of the Great Barrington Declaration likely had in mind. Â The benefits of these types of actions to the most vulnerable are large, but the negative effects on the economy and the costs to the rest of the population are minimal to non-existent.
robc
Oct 21 2020 at 6:07pm
Volunteer? No.
With a pay raise over my current job? Absolutely.
JFA
Oct 22 2020 at 2:43pm
Volunteer to be a teacher… definitely not… my own kids are just enough for me to handle. Also, I know several teachers in my district, and they are eager to get back into the classroom. Also, I find in-person schooling to be a much more vital service than I do eating in a restaurant, getting a haircut, or going to the gym. There seem to be plenty of people “volunteering” to work in those places.
Bill.D
Oct 21 2020 at 3:57pm
Agree with the theme of the article. One point missed: the main group of schools closed, at least in my area, are the government-run schools. The private and parochial schools are mostly open. It’s amazing how incentives work. At my kids’ private school the administration and teachers worked their collective butts off to come up with a plan to open. At the city schools the administration and teachers union butt heads and just come up with excuses not to get things going.
Kevin Dick
Oct 21 2020 at 4:26pm
When you say “order of magnitude higher”, I think you may be comparing daily case rates to cumulative case rates
From the Oster database today:
Student daily case rate 9/28-10/11: 10 per 100K
Teacher daily case rate 9/28-10/11: 26 per 100K
Our World In Data says the the rolling 7 day average of cases per day in the US as of 10/11 was: 48,000. That works out to 15 per 100K.
Airman Spry Shark
Oct 21 2020 at 4:34pm
COVID-19 infection appears to entail a high risk of serious long-term health effects after recovery, even in asymptomatic cases; the flu does not.
Comparing the two without addressing this distinction is naïve at best.
Alan Goldhammer
Oct 21 2020 at 5:37pm
Yes, this is correct. The weird thing about this is that women seem more prone to ‘long-haul’ symptoms than men whereas the mortality rate for men is higher. Lots of things we don’t know about this virus.
JFA
Oct 22 2020 at 6:34pm
Covid is probably worse than the flu for long-term complications, but your statement “COVID-19 infection appears to entail a high risk of serious long-term health effects after recovery, even in asymptomatic cases; the flu does not” is probably not true.
https://www.nejm.org/doi/full/10.1056/NEJMoa1702090 (risk of heart attack increases substantially in the first 7 days of flu)
https://academic.oup.com/eurheartj/article/29/1/96/2398118 (risk of stroke and heart attack are elevated for 4 weeks after onset of flu) (this suggests the same thing with less direct data: https://www.nejm.org/doi/full/10.1056/nejmoa041747).
Just like Covid, if flu has those measurable impacts, it probably also has more subtle long lasting impacts that are harder to measure.
Alan Goldhammer
Oct 22 2020 at 8:43am
This paper just came over the server today. A pair of Simon Fraser Univ researchers found, “Early evidence indicated that schools were low risk and children were unlikely to be very infectious, but it is becoming clear that children and youth can acquire and transmit COVID-19 in school settings and that transmission clusters and outbreaks can be large.”
This is a useful paper to read as it relates directly to David’s op-ed and previous two blog posts on the Oster data. As I’ve noted in my earlier comments, this issue is complex and there is not an easy answer.
J Mann
Oct 22 2020 at 10:15am
David, one technical question.
When Professor Oster compares the infection rates in schools to the larger population, I wonder if there’s a screening problem that makes the comparison difficult.
In other words, the 2.6% number comes from the people we test nationally, and due to the limited supply, a large number of those are people who have symptoms that indicate they might have Covid. We don’t actually think that one person in 40 walking down the street today currently has Covid, do we?
By contrast, I wonder if schools are doing random testing, or requiring testing to start school, etc.
I’m not sure if that’s an issue – hopefully Prof. Oster addresses it in her work.
Thanks!
Vivian Darkbloom
Oct 22 2020 at 10:33am
Very good question.  I’m curious as well.  When evaluating the data, it is very important to consider what  the composition of the pool is.  I think it should be pretty clear that between a sample group with symptoms and a random group, the latter is always going to have a lower infection rate.
Alan Goldhammer
Oct 22 2020 at 10:56am
I have not seen much that indicates school districts are doing much testing on a routine basis. There are a lot of colleges and universities that are doing this right now and I just read a paper this morning at the JAMA Network site on this topic. They provide a useful table and rating system.
I suspect the Oster data on infections comes from those who have been diagnosed and tested positive. The problem with children is that many of them have slight symptoms and can be positive for SARS-CoV-2 and capable of viral transmission. Those individuals would only be picked up via a random test protocol which is likely not in place.
Thomas Hutcheson
Oct 22 2020 at 11:23am
It is an error if not disingenuous to discuss school opening mainly in terms of risks to students. That would be a very myopic cost benefit analysis. The issue is total community costs to balance against the cost to students and their families of not opening physically. This might be part of the difference between pubic and non-public schools.
Comments are closed.