Omicron: Covid’s Fury Unleashed

How high go the counts?

Today I want to address two topics, the trajectory of case counts and reports of Omicron causing milder disease.

In future posts I will discuss the following:

1. Omicron and immune escape: neutralizing antibodies and cellular immunity (T-cells)
2. The CDC says it is safe to stop my isolation after 5 days without a test, is this good advice? Can I also now use the top rung of a ladder as a step, make toast in the bathtub, or fly a kite in a thunderstorm?
3. Why don’t we have more tests? And why aren’t they free?
4. We are approaching the end of the beginning of this pandemic (3–9 months probably). Next comes the intermediate phase. What does history tell us happens in a pandemic’s intermediate phase?
5. A couple other things I can’t remember ’cause I’m posting this at 2AM.

The daily rates of infection are skyrocketing. How high can they go?

In Covid: Endgame , I described a number called S_end which tells us how many people ultimately will have been infected when the pandemic is all over, however long that may take. It does not tell us the time course of those infections, what shape they will take, or in what locations they will occur: just how many people will have been infected at one time or another over the course of the pandemic. There is another number, I_max, which tells us the peak number of simultaneously infected individuals. (don’t worry, I promise to keep the math simple.)

I_max is given by the super simple equation:

I_max = Everybody minus a portion of “almost everybody”.

Formally, it looks like this:

I_max = S₀ + I₀ — S₀/R₀ * [1 + ln (R₀)]

Here is the graph of the equation as a function of R₀ using 7.9 billion for the population of earth:

If R₀ is 10, it corresponds to 5.2 billion earthlings simultaneously infected!

If R₀ for Omicron is closer to 20, then that number becomes 6.3 billion!

Simultaneously.

Infected.

There is an important caveat. Humans do not behave like molecules in a well-stirred beaker. The equations above assume that that the “mixture” (the human population) is well-mixed. To relax that assumption, you have to introduce a spatial component to the equation. When you do that, the epidemic doesn’t just evolve uniformly in time. It also spreads out in space, racing outward from epicenters in a traveling wave.

The traveling wave has peaks and troughs, phases. This technique of modeling a pandemic is called S-I-R with spatial diffusion. Here you do not consider the whole population of the planet at once because not everyone is simultaneously experiencing the same phase of the pandemic. Instead, you consider only the part of the population that in phase with eachother, in particular, that are riding the crest of a wave together. In our current predicament it would be reasonable to consider the width of a band representing the crest of a wave to be roughly 100–200 miles. This may be applicable to something the size of an average US state: a 150 mile wide band moving across the country.

This is why it was so ridiculous when Florida Governor DeSantis so pridefully declared his appropriacy in choosing to allow the delta wave to ravage his state, boasting that because of his courageous leadership Floridians had finished with the pandemic and could get back to business as usual. All he really did was to ride the crest of a wave into a trough.

Look at Florida now.

As I wrote in a high school essay on the Titanic, “Nature, it seems, takes delight in making fools of arrogant men.”

So what can you expect for your city or state? This Omicron wave is going to go fast. Really fast. From the equation above and estimates of R₀ for Omicron, I think we can expect roughly 1/2 of all members of our communities to be simultaneously infected with Omicron at some point in this wave. Not all communities will peak simultaneously. If we estimate the duration of illness at roughly 10 days, then at the peak of Omicron in your town you will probably be seeing 5% of your community getting infected daily. If you live in a city of 100,000, that will be 5,000 new daily infections at the peak of Omicron.

But since not every community in the country will be experiencing its peak at the same time, the peak daily infections for the US as a whole will not be 16.5 million (5% of 330 million). Some areas’ peaks will coincide with other areas’ troughs. But it could be close to 2 million. As of today, we are already at almost 600,000 daily infections.

Another caveat. Because we do not have the testing capacity to run that many tests per day and because many tests are now being done privately at home and going unreported, we will not see such high numbers in any “official” reports. But if the official reports get to 1–2 million cases per day, then the actual number of new infections is much higher.

How many people are going to die? That is a hard question to answer. Let’s do a back-of-the-envelope calculation. First assume that Omicron infection is half as likely to cause death as the alpha strain of January 2021 (I address this estimate in the next section). Next assume that 4 times as many infections occur at peak in January 2022 wave as there were in January 2021. In that case you would expect twice as many daily deaths at peak in January-February 2022 as there were last January.

Last January, that peak occurred on January 11 with a peak 7 day average of 3,500 deaths. There are a ton of “if’s” built into that calculation. I think it is unlikely that the peak 7 day average daily deaths will be less than 3,500. It probably also won’t be higher than 10,000. Still, there is going to be a flu season this year superimposed on the Omicron wave (the flu is already here) and I don’t think our healthcare system will do well even at the low end of those numbers.

On Reports of Milder Omicron

There is no shortage of headlines about Omicron causing milder illness than Delta. More and more researchers and doctors are becoming convinced that this is the case. There are some notable holdouts among top covid researchers. To be clear, the two viewpoints at this point are: 1) Omicron is definitely milder; and 2) it’s too soon to say. No one at the moment is saying they think is worse. The voices saying it’s milder are getting all the airtime but the ones urging patience are worth of listening to (and have the stronger argument in my personal opinion).

There are a number of factors that may contribute to a variant causing milder illness. Broadly, these can be lumped into two categories: factors intrinsic to the virus and factors intrinsic to the host (such as some degree of existing immunity). The first mistake many people are making is assuming that milder disease in South Africa is due to something intrinsic to the virus. This assumption leads to the not-necessarily-correct conclusion that the experience in other parts of the world will be the same. If the milder disease seen in South Africa is due to host factors (such as a high degree of population immunity from vaccination or prior infection) the experience of one country may not predict what happens in any other. Also, do not forget the impact of seasonal effects — it is summer in South Africa.

The second mistake being made is a fundamental mathematical error. None of the research papers released so far has made of use of cross-correlation in its assessment of the milder/not milder question. Cross-correlation is a signal processing technique used when searching for one signal buried in another signal offset by some delay (such as the delay that occurs between diagnosis and death or between diagnosis and hospitalization). Cross correlation is used in radar and sonar to find the delay in signal return that represents the distance to the object being tracked. Those researchers who are unfamiliar with cross-correlation are the ones who have been on the “it’s milder” wagon for weeks. The ones who are proficient with cross-correlation are the ones advising restraint. After all, they argue, we should not be making ANY decisions based on any expectation of disease severity.

The third problem is a reporting problem. None of the studies getting attention has yet been peer reviewed. The reviewers are no doubt raising all of the questions above — and more — in the formal process of peer review. It is possible for all the studies now being actively consumed by the public to be rejected for publication.

When will the cross-correlation method give us an answer? The South Africa signal will emerge from the noise roughly on January 6. This date is determined for technical reasons I won’t bore you with unless someone asks. My guess at this point is that the data from South Africa will probably end up looking nontrivially milder (perhaps 45% or so reduced risk of death or hospitalization). But the noise amplitude is still so large that I can’t rule out the possibility that Omicron in South Africa will be the same severity as Delta (or trivially milder, say 5–10%).

What about other countries? The signal from Denmark will follow a few days later and the UK roughly January 14–16. The US data should be apparent sometime January 16–21.

Aside from just observing and computing what Omicron does to human populations is there any biological evidence that Omicron causes milder disease? There is only one study (not yet peer-reviewed) of Omicron’s intrinsic virulence. And it looks at virulence in mice and Syrian hamsters. It showed that relative to Delta, Omicron replicates 70 times more in upper airways and one-tenth as much in the lungs. Personally, I ascribe no value to that information whatsoever (it’s a good starting point for researchers to generate more hypotheses but nothing more).

What should we as individuals be doing now?

Get vaccinated. Get boosted. Get your kids vaccinated and boosted. This wave is gonna go really, really fast. It is going to be brutal on a population level even if it is mild from the individual perspective. It will coincide with a flu season and come at a time when the healthcare system is buckling financially and suffering from a deficit of morale and resources. For the next 4–6 weeks, if you can stay home, do so. When you have to go out for essentials — groceries, medical appointments, etc. — I recommend an N95 or KN95. Read about how to properly don and doff it. If you want to know what I am doing. I am staying home. I bought new glasses 2 weeks ago. They have been in for a week but I haven’t picked them up — I don’t want to sit face-to-face with someone while they scrutinize and adjust the glasses on my face. I’ll pick them up in February. I will be sending my daughter to school but I am asking her to wear an N95 instead of her usual cloth mask for the month of January. I am going to ask that she not have to participate in band. If boosters for her age group are approved next week, I will be taking her to get one.

What should our leaders be doing as a matter of policy?

There I will not tread. At least not now.