California will be used to illustrate unless otherwise stated.
Early on, it became very apparent that COVID surges were linked to variants. This is pretty obvious when looking at proportions of variants in relation to hospitalizations.
Many sources of data have faded in and out, particularly good case data. That has necessitated other metrics to understand COVID in a particular geography. Positivity, wastewater, and percentage of emergency department visits have become other good early indicators and have strong concurrence, such as seen across the entire US in this view.
The coefficient of determination, or R-squared values, also support their tight relationships.
- 0.89 Wastewater and ED visits
- 0.86 Wastewater and Positvity
- 0.70 ED visits and positivity
This is a good explanation of r-squared if you want to dive a little deeper into statistics.
Here are the percentages of different variants with the plots of wastewater, positivity, and percentage ED visits on top of them. I also have the ED visits multiplied by 10 as a means to better see the curves. I’ll use the numbers and letters at the top to explain.
The vertical black lines represent the start of different surges, primarily using the wastewater data, but they correlate with the others as well.
- Line 1 – This is the start of what is commonly known as the delta wave. Here are the three variants isolated.
The three delta variants are closely related as shown below. 21A is the parent to 21I and 21J. That is part of the reason that they comprised a single surge. One can also conclude that 21J had a much higher r-naught, or reproduction value, and quickly dominated the other two.
- Line 2 marks the start of 21K, the original omicron variant. Notice how quickly that became dominant and caused a rapid surge like 21J
- Line 3 and 4 are where things get interesting because of some competition between 21L and one of its descendants, 22C, both of which are quickly overtaken by another subvariant of 21L known as 22B, so once again, a closely related family for that surge.
- Lines 5 and 6 follow the traditional single variant surge pattern, with 22E and 23A respectively, which have very different lineages.
- Lines 7 and 8 get messy again with 23B, 23D, and 23 F competing with each other, 23F being a subvariant of 23D, but all of the same lineage. The lower surge in wastewater values suggests short term protection among these different variants by infection with one of the others since they are similar. It’s also worth noting that none of these hit the 50% of sample threshold, which I will come back to later.
- Line 9 starts another big surge based on wastewater due to 24A.
- Line A begins the current surge which started with 24E in the late summer. You can see another surge superimposed on it at Line B, due to 24F. These are very distinct lineages, hence there is very little protection from one when infected by the other.
Hopefully this helps clarify what drives surges. It’s not seasons or elections. It strictly has to do with variant lineages and variant reproduction rates. It also supports my argument that aside from very different lineages, surges are related to about a 50% dominance of a variant.
This data also suggests why some surges might be smaller or larger simply based on how closely the predominant variants are related. If they are a close lineage, one will provide some immunity…FOR A SHORT TIME, but that still doesn’t guarantee protection. It’s insane to get infected with the notion that will provide “natural” immunity. Getting vaccinated with the current booster offers the best protection. I just wish these were updated every 3-4 months based on dominant strains and made available at that cadence as well.
Note: Much of the data and data visualization comes from https://covariants.org/.
