By March, This May All Be Over

People forget very quickly, which can be both a blessing and a curse.

As we enter into a period of vaccine triumphalism in the United States, at the same time that our national daily COVID-19 death tolls break and re-break records, it’s instructive to remember that public health approaches to stop a pandemic like this one have been well understood for more than a hundred years (perhaps even for more than five hundred, given that quarantine – as word and practice – derives from measures taken by Italian city states in the late 1340s to stop the spread of the plague).

But remembering is hard, and mostly, the public focus in the US has shifted to how soon we can forget – the pandemic, the year 2020, everything we should’ve learned. The question that dominates all attention now is: When will this all end?

Tl;dr on what follows – I point out the widespread failure to take into account the fact that a significant and rapidly growing fraction of the US population almost certainly already has immunity to COVID-19 through having contracted and survived the SARS-CoV-2 virus; conclude, through some back-of-the-envelope math, that there’s a reasonable case to be made that – through the combination of infection-based and vaccine-based immunity – herd immunity will be reached in the US sooner than expected (my best guess is sometime early in the spring); suggest, with no expectation my suggestion will have any impact, that a better vaccination strategy would involve not vaccinating anyone known to have already survived infection with SARS-CoV-2 because there is very strong empirical evidence that reinfection almost never happens; and reflect briefly on the injustice in global disparities in vaccine access. The math is all very simple, so please, don’t be daunted. Read on.

Let’s do some math. At the time of this writing (late morning EST on Saturday, December 19th), the New York Times shows ~17.5 million confirmed cases of COVID-19 in the United States and ~314,000 deaths. This, of course, makes no sense – though it’s no fault of the Times as the paper of record is just tracking (and beautifully visualizing!) the official numbers – because there is very strong evidence at this point that the case fatality rate (CFR) for COVID-19 is somewhere between 0.5% and 1%. Given that the virus seems to have spread more aggressively in recent months amongst younger people and that advances in knowledge about and treatment of COVID-19 have led to improved outcomes, I’m going to assume a CFR of 0.5% in the US, which implies that 1 out of 200 people who become infected with the virus end up dying from the disease.

Additionally, I’m going to assume that there have actually been 350,000 COVID-19 deaths to date in the US because there is good evidence that deaths have been consistently under-reported, and a 10% rate of under-reporting seems conservative but in the ballpark.

Now, combining those figures, we arrive at an estimate for the actual number of COVID-19 infections in the US (which is predicated on the fact that it is harder to miss deaths than infections, thus mathematically sound to extrapolate from deaths to total infections):

(350,000 deaths) x (200 infections/death) = 70 million COVID-19 cases in the US

That’s a lot more (~4x) than the official count; however, deaths are a lagging indicator of infections, so really, this math suggests that there had already been 70 million COVID-19 infections in the US three weeks ago (assuming, again conservatively, that the average time from infection to death for people who, tragically, die from what should have been a largely preventable disease is three weeks). Lately, ~3,000 people have been dying from COVID-19 each day in the US. Applying math similar to that above, we can estimate daily infections:

(3,000 deaths/day) x (200 cases/death) = 600,000 cases/day

Again, that’s a lot more (~3x) than the official figures, which have lately been above 200,000 confirmed cases per day. Given that these are very approximate numbers anyway, I’m going to round down to half a million cases per day in the spirit of conservatism and to make the math easier. Now, let’s say that there have been 500,000 new cases per day in the US for the past three weeks, and let’s round to 20 days, again for convenience. Then we’d expect – on top of the 70 million cases implied above by the current approximated actual US death toll (an estimated 350,000 deaths) – roughly the following number of additional US COVID-19 cases:

(500,000 cases/day) x (20 days) = 10 million

That suggests that there have been ~80 million total (70 million + 10 million) COVID-19 cases in the US so far. That’s roughly a quarter of our population of ~330 million people.

Okay, but aren’t people who have been infected with COVID-19 still susceptible to re-infection. The short answer appears to be: No. To quote the website of our beleaguered (because under political assault) CDC: “Cases of reinfection with COVID-19 have been reported, but remain rare​.​” From the standpoint of public health, all of us should continue to behave as if we are both at risk, and pose a potential a risk to others (certainly, that’s been my practice, although I’m almost certain that I had COVID-19 in the second half of March), but from the standpoint of empirical evidence and precedents derived from knowledge of other human coronaviruses (and human viral infections in general), there is a strong case to be made that the vast majority of people who become infected with SARS-CoV-2 are not susceptible to reinfection, at least in the near-term.

Strangely, this fact has not been incorporated into the discourse around vaccine rollouts, perhaps because it smacks of the discredited (and sadistic) herd immunity approach to COVID-19 response; however, while such a herd immunity strategy is idiotic as a public health measure, herd immunity is a thing.

Now, it has been very hard to get a handle on what the timeline of vaccine distribution will be in the US, but according to Business Insider, “Moncef Slaoui, the chief advisor to Operation Warp Speed, has estimated that 20 million Americans could get a COVID-19 shot before the end of 2020 [… and… ] that the US could immunize 100 million people by the end of February.”

There are roughly 70 days between now and the end of February. Let’s imagine that the torrid rate of spread of COVID-19 in the US continues unabated (not an unreasonable assumption given that many people in the US still seem to be planning to travel over the holidays and to host or attend indoor gatherings); that would lead to roughly the following number of additional cases during those ~70 days:

(500,000 cases/day) x (70 days) = 35 million

That’s a lot, and these are approximations, so let’s be more conservative again and guess that there will be only 30 million additional cases between now and the end of February. That would bring our total national case count to 110 million (80 million + 30 million) which, conveniently, is roughly one-third of the US population.

Okay, finally, let’s assume, perhaps wrongly, that past history of SARS-CoV-2 infection and vaccination against COVID-19 are independent events, such that, if 100 million people are actually vaccinated by the end of February, we’d expect one-third of those vaccinated to be people who had actually already been infected with SARS-CoV-2. That would then imply, obviously, that two-thirds of those vaccinated would not previously have been infected:

(100,000 million) x (2/3) ~ 67 million

I’m going to round that to ~70 million people vaccinated who had not previously been infected with SARS-CoV-2. If we add that 70 million to the 110 million people we’ve estimated will have already been infected by the end of February, we get 180 million people with meaningful immunity to COVID-19.

With respect to the percentage of the population that needs to have immunity to COVID-19 before herd immunity will be reached, estimates vary, but 70% seems to be a pretty widely used figure. Given a total US population of ~330 million people, that yields:

(330 million) * (70/100) ~ 230 million

Odds are that even a lower rate of population-level immunity would be enough to break epidemic community spread, although the utter failure of large portions of the US population to take even the most basic preventive measures calls that hopeful notion into question. Still, 180 million people (~55% of the US population) with some form of immunity by the end of February would put us, nationally, surprisingly close to the 70% figure posited to achieve herd immunity. Now imagine, instead that, only people who had not previously been infected with SARS-CoV-2 were immunized. Then you’d end up with ~110 million people with immunity post-infection and another ~100 million with immunity through vaccination for a total of ~210 million people (~64%) with some form of immunity. That’s even closer to the requisite 230 million. (More on this in the post-script below.)

Either way, what all this math suggests to me is that the pandemic in the US will likely be over by March or April, given the devastating amount of transmission that is already largely “baked in,” and so long as the vaccine rollout proceeds roughly as has been outlined above.

That will make for a long year for those of us who remember the start of March in New York City. This conclusion also offers a very strong reason for people, especially those at risk, to continue to take significant precautions to protect themselves (given that the Federal Government is doing next to nothing beyond Operation Warp Speed, to address the pandemic, and much of the population is caught in the throes of delusion and sociopathy): The end is in sight. It also points to the much-remarked-upon-but-still-profound injustice that the US population will enjoy near universal vaccine rollout before people in many other parts of the world even begin to have vaccine access.

Clearly, rich countries like the United States don’t plan on vaccinating their entire populations multiple times over, and, as Dr. Krishna Udayakumar explained on a recent episode of Democracy Now! (during which he also spoke about COVAX and the prospect for wide global distribution of vaccines developed in China), the governments of some rich countries have “made clear that if and when they have excess doses available, they envision making those available for low-income countries.” Vaccine hand-me-downism is not an excellent approach to ensuring public health globally, but it is among the prevailing paradigms for now.

All of this should be taken with a grain of salt. I’m neither an epidemiologist nor a mathematician, but the failure to account for the role of natural infection in driving immunity rates in the US is likely leading to an over-estimation of the time to herd immunity, just as an under-estimation of how bad the fall wave of the pandemic would be led some of us to over-estimate how long it would take to complete clinical trials and have vaccines ready under EUA from the FDA.

As we prepare to forget, I only hope we’ll also learn to remember, among other things, how little we know; how swiftly things can change; and how frequently inertia (or a belief that how things are is how things will be) can obscure the relationship between what has already been done and what is now very unlikely to be undone as the future unfolds. Such a line of thought obviously also has ramifications with respect to climate crisis.

Postscript: Unfortunately, our failed national testing strategy means we have very little information about who has and has not been infected with SARS-CoV-2, which complicates any plan to only vaccinate those who have not been infected (especially given the extent of asymptomatic infection and the unreliability of serological tests); however, even if only people with PCR-test confirmed cases of COVID-19 who were willing to forego/postpone vaccination voluntarily were skipped over in the vaccination queue, it would likely lead to millions, or even a low tens of millions of as-yet-uninfected-so-not-immune people getting the vaccine sooner than they otherwise would have.