In a recent series of posts, I’ve started some analysis of COVID-19 cases in K-12 schools in Virginia. Having worked with data all my life, the first thing I’d like to know is how reliable the data sources are. So in this post, I first compare what my local school district reports to what the Commonwealth of Virginia reports, for what should be equivalent school-age populations. And then take it from there.
Who tracks what?
For Virginia as a whole, the only accounting of COVID-19 in K-12 that I have come across is the Virginia Department of Health’s count of self-reported “outbreaks“. In most cases, you can look up whether or not an “outbreak” was reported for a school of interest.
But outbreaks aren’t the same as cases. See my prior post for the definition of an outbreak. So far, there have been no outbreaks reported in my local school district, Fairfax County.
It appears that all responsibility for public reporting of known COVID-19 cases in Virginia K-12 schools rests with the individual school districts. For my neighbors, that would be Fairfax County Public Schools (FCPS).
(I believe that both City of Fairfax and City of Falls Church schools are included in the FCPS reporting.) Fairfax City students are in the Fairfax County data, but Falls Church runs its own independent school system. That’s not much of a data issue, as Falls Church only has about 15,000 residents (versus about a million in Fairfax County).
FCPS provides a daily count of cases reported in their schools, by school. You can find it at this FCPS link.
For reasons I cannot fathom, FCPS has made it virtually impossible to look up the data for a given school. (Which is nuts, because clearly that’s what parents are going to want to see.) You can, with effort, track a single school through their graphs.
(Or you can download the data, which, because they use the Tableau system, is always in the least-usable format possible. Tableau is a popular graphics system, but it’s more-or-less the roach motel of data. Numbers go in, but they never come out. I’ll address the unhelpful format of the FCPS data below.)
Contrast of FCPS counts and Virginia Department of Health counts.
The first question is, how do the FCPS counts of COVID-19 cases in school children square up with the Virginia Department of Health counts of cases in school-age children?
For all of Fairfax County, the FCPS lists 266 student cases in August, and 78 so far (to 9/6/2021) in September. Recall that FCPS schools opened on August 23, 2021. When I process their data file, I count 295 cases from August 23 to September 6. And, to match the Virginia Department of Health counts, I count 266 student cases reported in FCPS schools from August 23 to September 3.
But if I download the Commonwealth of Virginia counts of cases by age, date, and locality, I count 460 new COVID-19 cases ages 5 to 17, in Fairfax County, over the same period.
There are some factors that might explain part of the discrepancy. We have some private schools in this area, some home-schooling, and so on. But I’m pretty sure the number of such school-age children not in FCPS schools is trivial compare to FCPS enrollment. And the age match is not perfect (5 – 17). There may be some differences in the timing of when cases are reported. And so on.
But I think my conclusion is that the FCPS counts of school children with COVID-19 only capture about 60% of the known Fairfax County cases in school-age children. It looks like about 40% of the known cases among school-age children are not being reported either to or by the schools they are attending.
Still and all, for an ad-hoc system based on voluntary self-reporting, that’s really not that bad.
Sort the FCPS data by school
For now, I’ve just taken the FCPS data and sorted it by school and descending date. That allows you to see all the data for a school in one place.
That’s in the spreadsheet below. Note that each line of the file is a case, so if two students were identified on the same day there will be two lines. Also note that the data go all the way back to last school year. I’ve put in conditional formatting to highlight this year’s cases in red.
I’ll try to do something a little more user-friendly in a bit.
How does this year compare to last year?
Finally, because Fairfax has the data compiled since the start of the prior school year, I can compare the counts for equivalent time periods in last school year and this school year.
Source: Calculated from FCPS case counts dowloaded from this link.
You have to be a little cautious about that comparison. Any time you introduce a new data reporting system, there’s always the potential for a rough startup. So there’s a chance that last year’s numbers are low.
That said, the number of staff cases is roughly the same, so at least that portion of the reporting was working. The number of student cases, by contrast, is a different issue entirely. Either there really was no reporting going on at the start of the last school year, or this year is very different from last year.
We all expected that number to be higher this year. But I’m not sure anyone expected them to be that much higher. As you can see below, the level of disease circulating in the community is higher now than it was last year at this time. But only about four times higher.