Take homes:
We will see maximum spikes in the US of 1/3 the Omicron peak
The first counties have peaked and are coming down. NYC will peak in about a week and then will be down in two to three weeks.
These are being attributed by the lay press to BA.2.12 and BA.2.12.1 but this does not look clear to me in the data.
Contrary to many in the media being certain that it would happen, the US as a whole did not see a BA.2 surge as I forecasted a couple of weeks ago.
However, there is a small uptick in cases in the N. Eastern US emanating out from regions around the Canadian border. Something is also happening in coastal Alaska but I’m not going to focus on that for now.
Fine citizens on social media have been asking me how high of a surge we will have, when it will peak, and when it will be over. So I did some investigating. I will walk you through my logic. Most of my predictions currently use ratios and timing to understand what is happening and what will happen from first principles. I could automate this and turn it into math and visuals but haven’t had time yet to get this going again. So instead I will talk you through the back of the envelope logic that basically anyone could do.
Step 1. Where did cases go up first? Cases went up first in Oswego County, NY around March 22. I checked the other hotspots in VT, ME, Ontario, and Quebec and all of these locations went up five or more days after Oswego. Cases appear to be spreading from there so it’s a good bet that whatever variant is causing spikes was birthed in Oswego.
I did some sleuthing to see if there was anything about Oswego that might make it prone to birthing a variant but didn’t come up with much. Located on Lake Ontario it promotes itself as "The Port City of Central New York." It only has 18,000 residents and was until recently amongst the less vaccinated NY counties. This recent case uptick is apparently convincing more people to get the jab. Other than that it seems to be known for a criminal scandal with its former mayor:
“In September 2005, Oswego Mayor John Gosek was arrested by the FBI and state police in a sting operation outside of a motel near Syracuse after allegedly giving a woman $250 to arrange a sexual encounter with two 14-year-old girls.”
But none of this probably relates to the recent uptick in COVID cases. So onto…
Step 2: What variants are where, in what percentages, and does the timeline match the case rise? So contrary to what this Forbes article and many other articles have said in the last 5 days, I don’t think we can attribute the rise in cases specifically to BA.2.12 and BA.2.12.1. The talented (soon to be Dr.) Meng Liu pulled the variant data in NY from the international variant repository, GISAID, and we took a look. So over the time period March 20th- April 6th when we saw cases rise in NY, BA.2 went from 22% to 55% of sequences, BA.2.9 (which no one is saying anything about for some reason) went from 15% to 35%, BA.2.12.1 went from 10% to 20%, and BA.2.12 went from 10% to 9%. So which of these is causing the case rise? I think we would need to do a state by state analysis formally but it’s not super clear. My money is actually more on BA.2.9 with the most dramatic non BA.2 rise. There are also about 10 other variants in the mix so, shrug. Media loves to declare a new variant, I know that much.
Step 3: How high did the cases go in the first place to go up? How long did it take? Can we infer the other case peak heights and timing from this? So back to Oswego County we go. Whatever happened there we can roughly guess will happen similarly in the next place whatever variant is causing this goes.
You can see below that there is a roughly 3:1 ratio between the winter Omicron peak and the peak from BA.2.? in Oswego. So we can expect the same in other places. There might be a little variation due to demographics and vaccination level but roughly I would expect this to hold in the US.
So what about the timing? When will cases peak? Well in Oswego the time from trough to peak was 21 days.
So add 21 days to when cases started to go up in your county and that is when you should expect cases to peak. Also waves are roughly symmetrical so you can expect cases to be fully down 21 days after the peak. We can take the example of NYC below.
Cases started to go up in NYC around March 23rd, which was only a couple of days after Oswego so maybe NYC will peak with even less cases in a couple of days. Possibly it will start to go up more dramatically if whatever variant is causing this hasn’t gotten there yet but, meh. At most we would see that 1/3 of Omicron peak level and people are a lot more vaccinated in NYC than in Oswego so this could be it. This little bloop that the news had everyone flipping out about for three months.
Anyway, it would be better if we had variant data that was on a county level and was not weeks delayed. That is once again my top wish for this pandemic- please US government, invest in better data so that we can have better predictions. We can only do as good as the data we have.
I hope this was helpful to understand what is going on, is a little bit reassuring, and perhaps instructive on how to figure out case predictions just by using some basic logic. If you want to support this SubStack, please consider subscribing. You can chat with me in the comments or find me on Twitter @DrGlorioso.
All of this beautiful data tracking (except for variants) can be found at the New York Times Covid Tracker .