Read on Twitter
In preprint w/ @KateGrabowski @khgrantz @andrewazman @CJEMetcalf @Lizstuartdc we look at #COVID-19 risk to household members from in-person schooling. A summary follows. Caution: this work has not yet been peer reviewed (1/8) https://www.medrxiv.org/content/10.1101/2021.02.27.21252597v1
This work uses data from @CmuDelphi COVID-19 Symptom Survey https://delphi.cmu.edu/covidcast/survey-results/?date=20210221. Based on responses throughout the country (A) we see huge differences frequency of in-person schooling (B), the amount full time (C) and the number of mitigation measures reported (D) (2/8)
After adjustment for reported case rates and other covariates we find a significant association between living w/ a child in inperson school and reporting #COVID-19 related. Risk increases w/ grade. We look at two measures of recent symptoms and +COVID test in last 14 days (3/8)
Odds of #COVID-19 outcomes decrease by ~8% w/ each mitigation measure. Symptom screening, teacher masking, and closing extracurriculars stand out Some that look less effective are implemented only with a lot of other stuff (possible saturation effects) (4/8)
When we stratify by number reported, it looks like ~7 school-based mitigation measures is enough to make household member risk the same as if they were living with kids not in in-person schooling. A lot probably depends on how well these are implemented. (5/8)
We couldn't drill in as much on teachers, but we did find that Pre K-12 teachers who reported "working for pay" outside home had an increase risk of #COVID-19 outcomes. The amount of this increase was similar to office workers. (Not sure what is up w/ food service folks) (6/8)
The type of survey and the topic raises lots of concerns about potential confounding and biases. We can't rule those out, but when we divided up respondents by attack rate (A), county urbanization (B) and propensity for in person schooling (C), associations hold. (7/8)
Overall, in-person schooling seems to be associated with #COVID-19 risk to household members, but this risk can be controlled by in-school mitigation measures. A lot still remains unknown, hopefully this study of ~600k people helps folks thinking about this thorny issue. (8/8)
