First, my tweet thread from August 26:
First, let’s look at the pediatric hospitalization metrics from the CDC:
Pediatric admissions increased, but I don’t know how much stock to put in this data because I’ve read conjecture that with the RSV surge coupled with Delta “choosing” them as the biggest population of unvaccinated, more kids are being admitted and have positive COVID tests, but are not actually being admitted for COVID in and of itself. I need to do more digging into this. But note that the pediatric admissions are still a fraction of those for the rest of the population.
So that’s the hospitalization part. Now let’s see how I did on the case counts, as per the St. Louis County dashboard of confirmed cases.
Everything trended slightly downwards from that point… because I was singularly focused on the infectivity of Delta, and forgetting about seasonality.
Theory: seasonality is what governs the coming and going of COVID, just as it has every coronavirus. NPIs and the like only change the curve marginally at best. Edgar Hope-Simpson observed this in the early 1980s, that respiratory viruses come and go based on a distinct seasonal pattern depending on what climate you’re in.
(I’m actually kind of thankful that I’m wrong, because I suspect that increased cases along with schools opening would have led to a lot more closures in our area.)
Let’s look at this graph again, focusing on the time period between April and September:
See how the shape of the curve is basically the same? The magnitude of it changes a bit due to increased testing in 2021, but it heads steadily downwards between April and June, heads upward in the beginning of July, then plateaus with a slight downward trajectory until the end of September.
Or, let’s think about this in terms of the climate. There’s a lull in cases between May and July while people are spending more time outside and getting sunshine to build their vitamin D reserves. Then August steps up due to outdoor heat forcing everyone indoors. Then a bit of a dip again, then cases start escalating in November-December as everyone heads inside again and sunshine exposure dips.
So here’s my hypothesis:
As we saw last year, cases will begin to rise throughout October, reaching a peak before Thanksgiving.
If my understanding of seasonality is correct, then we should be reaching 500 cases a day by Halloween. This will fly in the face of the mask mandate just passed in St. Louis County, and will lead to lectures to act more responsibly. But it’s nothing to do with people acting badly or not listening to their leaders. It will be the virus doing what the virus does, in the time and manner it chooses to do so.
I’ll revisit this in a couple of weeks and see how my hypothesis is holding up.