log in

I believe US could see rolling series of “Ireland events” over next 2 months

Hacker News - Wed Jan 13 19:20

Daily reported new Covid cases, Ireland, through Jan. 7, 2021

For the past year, I’ve been consumed with how Covid numbers are used/manipulated to create political narratives. From China to WHO to don’t-test-don’t-tell to Covid Trutherism in all its forms … that’s been the windmill I’ve tilted at for almost 12 months now.

Last week I became consumed by a new twist on all this – Covid numbers that were being largely ignored. Insane infection numbers coming out of UK and Ireland, apparently driven by a new virus strain, that we acknowledged over here but didn’t seem to be too mussed about.

It reminded me of the Covid numbers coming out of Italy last February. Was Europe once again our crystal ball? Were we once again going to ignore THAT?

And when I say “insane infection numbers” I mean a 30x spike in Covid cases in Ireland over the span of two weeks in late December, where the R number – the basic reproductive rate of the disease – went from something around 1.2 to something around 3. Where you suddenly went from a few hundred new Covid cases every day to more than six thousand cases every day. All in a country the size of Alabama (which, btw, currently has about 4 thousand cases every day).

So I’ve been trying to figure out what happened in Ireland, and whether it could happen here.

To do that I had to research this new UK-variant of the virus. I had to research the way in which Covid is explosively spreading in Ireland, and whether that was similar or different to US. I had to research what it MEANS to have an R-number go from 1.2 to 3.  And finally I had to dig into why this ‘Ireland Event’ was not being discussed by US Covid missionaries (to use an Epsilon Theory term) like Scott Gottlieb or Tony Fauci.

I’ll start with the conclusion.

I believe there is a non-trivial chance that the United States will experience a rolling series of “Ireland events” over the next 30-45 days, where the Covid effective reproductive number (Re not R0) reaches a value between 2.4 and 3.0 in states and regions where a) the more infectious UK-variant (or similar) Covid strain has been introduced, and b) Covid fatigue has led to deterioration in social distancing behaviors.

A single Ireland event is a disaster. A series of Ireland events on the scale of the United States is catastrophic. If this were to occur, I’d expect to see a doubling of new Covid cases/day from current levels in the aggregate (today’s 7-day average is 240k/day), peaking somewhere around 500,000 new daily cases before draconian economic shutdowns (more severe than anything we’ve seen to date) would occur in every impacted major metro area. Hospital systems across the country would be placed under enormous additional strain, leading to meaningfully higher case fatality ratios (CFRs) as medical care was rationed. Most critically, this new infection rate would far outpace our current vaccine distribution capacity and policy. Assuming that vaccines are preferentially administered to the elderly, aggregate infection fatality ratios (IFRs) should decrease, but the overall burden of severe outcomes (death, long-term health consequences) would shift to younger demographics.

Current US government policy rejects the possibility of an Ireland event, largely because of what I believe is a politically-motivated analysis by the CDC that models more than 100 million Americans already possessing Covid antibodies, prior to any vaccination effort. Using data from flu monitoring programs in prior years, the CDC models project that 70 MILLION Americans have already gotten sick with symptomatic Covid, but decided to just write it off as a bad cold and never got tested. I am not making this up. Add in another 10 million or so Americans who the CDC models as having already had asymptomatic Covid, add in the 23 million Americans who we know have had Covid, and voila! – per the CDC, one-third of the American population is already effectively immunized against getting Covid in the future. And obviously enough, if >30% of Americans are already effectively immunized against Covid because they’ve already gotten sick, then it’s very difficult to hit the Re numbers of 2.4 – 3.0 that Ireland is currently experiencing.

I think this model is wrong, and I think the CDC knows that it’s wrong.

I think it’s wrong because the 2021 behavior of someone who thinks they might have Covid is very different from the 2015 behavior of someone who thought they might have had the flu, but the CDC assumes it is the same in their models. You don’t ignore Covid. You don’t just brush it off. I’d say that no one just brushes off Covid symptoms the way they might have brushed off flu symptoms in the past, but of course that’s not true. I’m sure there are millions of Americans who have, in fact, had symptomatic Covid and ignored it, particularly in spring and early summer when our national testing capability was pathetic. But 70 MILLION Americans? Twenty percent of ALL Americans? More than three times the number of known Covid cases? C’mon, man.

I think the CDC knows this model is wrong because if it were true – if they actually thought that one-third of Americans were already effectively immunized by having Covid antibodies – this would be an ENORMOUS factor in determining vaccination policy. Otherwise, you are going to be wasting one-third of your precious supply of vaccines on people who don’t need it.

I think the CDC knows this model is wrong because if it were true, how do you make sense of Covid hospitalization rates?

Again, were there millions of undiagnosed and “brushed-off” Covid cases in the spring and early summer when Covid testing was ridiculously sparse? Absolutely. But unless you’re prepared to say that either the SARS-CoV-2 virus is much more dangerous today than it was in the spring or that hospital Covid admission policies are much more lenient today than they were in the spring, I think it is impossible to reconcile actual Covid hospitalization data on 23 million symptomatic-and-diagnosed Covid cases with a model of 70 million symptomatic-but-undiagnosed Covid cases.

So yes, I think this model is nuts. I think this was a politically-motivated Trump Administration exercise to “prove” that the US infection fatality rate (IFR) is really tiny and you’ve got nothing to worry about. One of many such politically-motivated efforts across many institutions to minimize the risk and impact of Covid-19.

But this CDC model is why prominent Covid missionaries like Scott Gottlieb and Tony Fauci have said that they expect daily case numbers to decline from here on out, not accelerate, and this is why I think a potential Ireland event is NOT priced into any mainstream market expectations or political expectations for 2021.

Unfortunately, once it becomes apparent that an Ireland event is occurring, it’s too late to stop it.

In our human-scale, linear world, we experience exponential growth like this: nothing, nothing, nothing … case, case, case … cluster, cluster, cluster … BOOM! But by the time we start to really pay attention to an exponential growth process – typically at the cluster stage – the process is already too entrenched to stop it, absent incredibly harsh social measures like you see China reinstating today in Shijiazhuang, a city of 11 million. No government in the West is prepared to even talk about these measures, much less implement them. So we’re always surprised by the BOOM. If an Ireland event occurs here, it will be no exception.

A full-blown Ireland event is driven by both the more virulent UK-strain AND a deterioration in social distancing behaviors. Either taken alone is bad enough. It’s the combination, though, that creates a regional superspreader event. Irish health authorities estimate that their starting point for Covid Re was something between 1.1 and 1.3 (meaning that, on average, one person infected with the SARS-CoV-2 virus would pass it along to 1.1 – 1.3 new people). They blame deteriorating masking/social distancing for the majority of their “event” (say, a 0.9 – 1.1 increase in the Re number), and the UK-variant for the balance (say, a 0.5 – 0.7 increase in Re). This is very much in line with the latest research from Public Health England, which estimates that the UK-variant Covid virus is approximately 40% more infectious than the baseline virus. Notably, the UK-variant shows an even greater increase in infectiousness for “close contacts” (not necessarily face-to-face, never touching and perhaps up to 2 meters apart) rather than “direct contacts”, meaning that the UK-variant virus is particularly successful at bridging the air gap between strangers or short-duration contacts in an indoor space. This is … ummm … troubling. As lax as we all have gotten with our mask wearing and our social distancing outside of the home, the UK-variant virus dramatically reduces the margin of error we have with mask wearing and social distancing outside of the home.

For the same reasons that we humans typically don’t recognize an exponential growth process prior to the cluster, cluster, cluster stage, we have an even harder time appreciating the impact of even a small increase in the effective reproduction rate of Covid. A 40% increase in Re has an enormous impact on how many people will be infected by Covid. For example, let’s assume that the current Re for the United States is something like 1.4 (I think it’s probably higher than that in areas like SoCal, and going up everywhere as Covid fatigue takes hold). With a 5-day infection cycle (assume you pass along the virus to 1.4 new people within 5 days of contracting the virus yourself, i.e. before you become symptomatic), a single Covid case will result in a grand total of 2,296 Covid infections over a 100-day period. Now let’s increase that Re by 40%, so that it’s not 1.4 but is 2.0 … now that single Covid case will result in more than 2 MILLION total Covid infections over a 100-day period.

This is the power of exponential growth. The numbers get silly … I mean, take that Re up to 3.0 (the high end of the current Ireland estimate), and a single Covid case will result in 5.2 BILLION total cases over a 100-day period, about 60% of the entire human population on the planet. Obviously our social behaviors around the disease would change dramatically well before we got to that point. But the real challenge of all this from a social behavior perspective is the nothing, nothing, nothing … case, case, case … cluster, cluster, cluster … BOOM! nature of any exponential growth process.

That Re of 2.0 that results in 2 million total infections from a single Covid case over 100 days? On Day 30 there are only 127 total cases. Not noticeable at all. On Day 50 there are just over 2,000 total cases. Barely noticeable. Let’s say you’re an elected political leader. Are you really going to take the steps that are necessary to stop this process – like shutting down domestic travel to and from an infected area, like physically quarantining entire cities – over a few hundred cases? Not a chance. Even if you’re right … even if you prevent a catastrophic outcome through your actions on Day 30 or Day 50 … your voters will never know that you were right. They will only experience the lockdown pain, and they will never credit you for the catastrophe averted.

I think we’re already at Day 30 in a dozen states. I suspect we’re already at Day 50 in a few.

So look, maybe I’m wrong about all this. Maybe we’re already well along the path to herd immunity, and one-third of Americans currently have Covid antibodies through prior exposure, just like the CDC models say. Maybe we’ll all rediscover that old-time religion when it comes to mask wearing and social distancing outside of the home. Maybe governors and the new Administration will focus on containing the UK-variant through domestic travel restrictions. Maybe we’ll wake up tomorrow with a new urgency about vaccine distribution.


But my spidey-sense is really tingling on this one.

To learn more about Epsilon Theory and be notified when we release new content sign up here. You’ll receive an email every week and your information will never be shared with anyone else.