Thread by @douglasharris99

I’m pleased to say @REACHCenterEd has released: “The Effects of School Reopenings on COVID-19 Hospitalizations.” We think it’s the 1st US study involving COVID health *outcomes* & 1st w data on whole country. Thread: reachcentered.org/publications/t…

Conclusion: We find no effect of school reopenings on #COVID hospitalizations for counties where pre-opening new hosp rates <=36-44 per 100k population per week. As of mid-Dec, 58% of counties across the country have rates this low.

Above 36-44 per 100k, the results vary by data/methods so we can’t draw clear concl. (BTW, we’re giving a range of numbers here to avoid implying a false sense of precision. It's not suddenly risky to open when you get to 45. Just a broad guideline.)

The fact that it seems safe in some places but perhaps not in others isn’t surprising. Schools should spread the virus less in places where there is less of it to spread.

One key advance is using #COVID hospitalizations. It seems all prior studies to date are based on COVID positivity rate. This is a problem: positivity rate is based on infrequent & unsystematic COVID testing.

In fact, opening schools itself can trigger more #COVID tests. Ex: students & schl staff test more often when they return to schl buildings to ensure they aren’t spreading virus. IOW, schls can change the pos rate w/o spreading virus. Hosp data don’t have this prob

Also, ultimately, we’re most concerned about whether school reopenings make people quite sick. Hosp get at that directly. (We understand that not everyone goes to the hospital, but that’s true in all counties & should be proportional to actual hosp over time within counties.)

How we did the study: We linked 3 data sets on school reopenings for all trad public schools & 2 data sets on hospitalizations from @Change_HC & @HHSGov. We incl essentially all school districts & all hospitalizations in the country.

We compare hosp in areas where schools reopened more in-person w/ areas opening less in-person. We also control for state reopening policies, college reopening, & (indirectly) things like political orientation of counties.

Interpretation: We are NOT saying that all schools below the threshold should open in-person. These decisions are the jobs of policymakers not researchers. We are just providing useful facts & necessary context.

We're also NOT saying that schools above the threshold should remain closed. Health conseq of reopening must be weighed against conseq of cont schl closures: child mental illness, abuse, malnutrition, & long-term health conseq. Parents also suffering these indirect effects.

Key to interp is that we are comparing schools that adopt policies that *allow* students to come back in-person. Many families are choosing to opt out & stay remote even when they can go in-person. Also, many school staff in risk categories are working remotely.

This means schools that did reopen in-person are doing it w more social distancing than would be poss if all students & staff actually returned. This is one reason we're NOT saying all students & staff in the low-baseline-hosp counties, should necessarily return to in-person.

What we are saying is this: It appears safe to adopt a policy of optional in-person school attendance in counties w relatively low levels of hospitalizations. Schools should work w/ local public health agencies for updated numbers. Hosp rates are changing

Another key issue in reopening decisions: Since most schools give remote options & some students are always in quarantine—“in-person” schls have to teach BOTH in-person & remote. It’s hard to do both, so reopening in-person could reduce instructional quality in some ways.

Please also look at the great study from @caldercenter & @EPICedpolicy. They study COVID pos rates, but otherwise use similar methods in 2 states (MI & WA) & come to similar conclusion—generally little inc spread due to schl opening, except where community pos rates are high

Thanks to our incredible team, which also included co-authors Engy Ziedan, a great health economist, & Susan Hassig, an expert in infectious disease @Tulane. We worked straight thru holidays to get this done.

Thanks again to our many partners. First, @REACHCenterEd, espec Olivia Carr, Dan Oliver, Sara Slaughter. Also: @Change_HC @HHSGov @educationweek @BurbioCalendar @mchdata & many more listed in the acknowledgements.

We went through a great deal of peer review in a short time. Thanks to @SarahCohodes, Austin Denteh, @CEDR_US @DrMikeHansen, Bob Kaestner, @imbernomics, Wei Long, @JonValant. Also whole @REACHCenterEd leadership: @joshcowenMSU @julieamarsh @amyschwartz @KatharineStrunk

Finally, thanks to school leaders for doing what's best for students & staff. We know how difficult this year has been & appreciate your efforts. Let’s all work together to minimize harm, help everyone rebound from this difficult situation, & bring back schls better than ever.