This page updates my analysis of Covid-19 in Fairfax County, Virginia to include all of 2023. The data from the Virginia Department of Health (VDH) and my model from October 2022 are shown in the graph below. The basic results are the following: In the 365 days of 2023, there were 24,580 reported positive PCR tests, which, given my model to compensate for test undercounting of cases (see below), implies approximately 98,000 actual cases. There were 304 hospitalizations and 117 deaths reported by VDH for Fairfax County in 2023. My model tracks fairly closely the reported data, given inevitable statistical variations. This gives credence to the model assumptions (see below). If I had known a year ago in January 2023 that there would be 98,000 cases in 2023, I would have predicted that we would see approximately 280 hospitalizations and 98 deaths in 2023. These numbers are not far from what was reported (304 H and 117 D), with actual hospitalizations and deaths being respectively 9 percent and 19 percent higher than the model predictions. While hospitalizations track the data well across all of 2023, there was a spike in reported deaths in December 2023 over ten times my model predictions. This is far more than expected based on cases in November. I strongly suspect that the reported increase is NOT due to any new virulence of the disease, but is rather more likely due to VDH catching up on old unresolved deaths that were recently closed before the end of the year as being due to Covid-19.
Fortunately, cases are running at a lower rate than in prior years and hospitalization and death rates remain quite low for a county of 1.1 million people. However, nearly 1 in every 10 residents of Fairfax County will have had Covid-19 in the last year (100,000 out of a million), but only around 1 in 4000 went to the hospital with it, and only 1 in 10,000 died. Consequently, its overall impact is minor.
This gives the model assumptions: the ratio of H to C and D to C are assumed to be constant at the values established in late 2022, namely, 1 hospitalization for every 350 cases and 1 death for every 1000 cases, occurring approximately 5 weeks after case onset. VDH only reports cases tested by PCR tests and thus reports an undercount, since many cases are untested. To conform to reported H to C and D to C ratios before the widespread use of home tests, I assume that the ACTUAL number of cases is 4 times the reported number of PCR tests. The solid lines in the graph represent the 7-day average of daily C (blue), H (green), and D (red) events from the VDH data. The dashed lines represent the predictions of my model for H and D, based on the reported cases C.
Word and Fire 