So now that we’re on the cusp of a vaccine for COVID, this is the question everyone is asking. As you know if you’ve been reading my reports since March, I primarily report on known active cases, which I derive by modeling new cases and recovery rates. I do speculate from time to time on the “why” as well as the “what” from a mathematical viewpoint. A common theme of these reports is the observation that most geographical areas seem to hit a prevalence ceiling, where known active cases reach a limit (usually between 1/3 and ½ of one percent of the population, then recede. We’ve seen this in wave 1 (the Northeastern states), and again in wave 2 (the sunbelt states), and during the current wave 3 with a number of the midwestern and western states. I have hypothesized that this behavior occurs because a significant percentage of the population (varies by area) has some resistance to the disease. There have been a number of articles to back this up, finding an innate resistance to COVID from T-Cell cross-reactivity. In other words, if you’ve had other corona viruses (most notably the common cold), you may have some level of COVID resistance. Researchers vary on the percentage of people with innate resistance, but numbers range from 20 to 50% of the population (see my discussion from August 2nd near the end of this report).
If this is true (I think it is, as it explains the ceiling phenomenon), then herd immunity may be closer than many think. Let’s look at some numbers. There are about 330 million people in the United States. Sometime today, we crossed 15 million known cases domestically. Way back in the spring and early summer, antibody studies revealed a total prevalence to reported ratio of about 10 to 1. I can’t believe that this ratio is as high today with testing running nearly 2 million per day, so I’m going to ballpark the current number at ½ that, or 5 to 1. With that assumption, roughly 75 million Americans have or have had COVID. If we add to that the number of people with innate resistance (I’ll use the mid-range number of 35%) that’s another 116 million people, for a total of 191 million people with some level of immunity. That is about 58% of the population. The current run rate of new cases is about 1.4 million a week. If this continues for the next month (It could attenuate somewhat after the Thanksgiving bump), that’s another 5.6 million cases, which when hit with the prevalence multiplier is another 28 million people with immunity. So with that, we’d expect a total of about 219 million people with some level of resistance to the disease by early January. That brings us to 66% of the population, very close to herd immunity, if not already there.
Now, on to the vaccines. A lot will depend on the public policy in the various states as to who receives vaccinations first. If you vaccinate congregate living residents first, I think you’ll have the greatest impact on deaths, but not the greatest impact on cases, since this demographic is a high proportion of deaths, but a much smaller proportion of new cases. If you wanted to decrease the new case count the fastest, vaccinating young people would be most effective. In any event, let’s assume that vaccinations are uniformly distributed amongst those that already have COVID resistance and those that do not. It makes sense to me that those that have had COVID would not be vaccinated until the tail end, but I have no idea whether that will happen. So for every 100 people vaccinated, 66 will already have some immunity, and 34 will be those susceptible to COVID. With that in mind, it would only take 87 million vaccinated Americans to reach 75% resistance, which, depending on who you read, is in the ballpark of herd immunity. Now I’ve made a lot of assumptions here, and this is admittedly speculative, but in the absence of better data, I think this is a reasonable analysis.
In this nascent stage, the rollout numbers for vaccinations are all over the map, but it looks like we could reach herd immunity by the end of February. I’ll continue to report during this phase, as it will be interesting to see when meaningful decline takes place.
Now, on to the numbers…
We had quite a surge of cases after the Thanksgiving holiday, bigger than I expected. It’s already leveling, so hopefully it’s confined to the holiday fallout. It didn’t happen everywhere – of the states I track, the ones increasing rapidly are NY, CA, NC, VA, SC, MA, AZ, and OH. I model that known active cases are now about 1.43 million, or 0.43% of the population.
Here is the national daily death count. Increasing as expected with wave 3. The brown line is my projection of daily deaths, based on 2.75 deaths per 1,000 known active cases and a 21 day lag. I’m expecting daily deaths to crest at the end of December at more than 3,900 deaths per day. The IHME model increased their projection as of December 3rd, and is now projecting peak daily deaths on January 13th at 2,967 deaths per day. I’m still expecting an earlier, but higher peak.
Here are the daily deaths per 1,000 active cases with the 21 day lag. We’re hoping to see this decline, but so far it’s still quite stable, which gives me confidence in my daily death projections.
Daily reported test results are now in the 1.75 million per day range. The U.S. has now reported over 209 million tests.
Here is Arizona, continuing to increase rapidly, but leveling off today (just one day, though). At 0.56% of the population, this is higher than I thought it would go.
SC has seen quite the post-holiday surge. It’s now at 0.36% of the population.
Here is Florida – a bit of a post-holiday increase, then flat.
California looks scary – growing very fast for 9 days. They are now at 0.44% of the population.
Georgia had quite a post-holiday surge, but is still only at 0.28% of the population. A bit of a drop today, however.
Texas experienced growth after the holiday, but now declining again. The peak a few days ago was at 0.31%.
Both VA and NC have not fared well post-holiday. VA is at 0.29%, and NC is at 0.37%.
NC daily deaths are now increasing, as expected. I’m projecting a peak daily death count of just over 100 in the first week of January.
Washington has been roughly flat for 2 weeks. The peak on November 30th was at 0.28%, still a bit low.
Here are NY and NJ – NY continues rapid growth, while we see slower growth in NJ.
Here is Massachusetts. Rapid growth since November 30. Now at 0.36%.
…And here is Michigan, gradually declining over the past 2 ½ weeks. The high water mark was just a hair over 0.58%, so I think that was the top.
PA has seen rapid post-holiday growth, but flat for today. They are at 0.55% of the population, so I believe this should be the peak.
Here is Colorado – Down significantly until the holiday, then flat.
Here is Illinois, similar to Colorado — down significantly until the holiday weekend, then flat.
Here is Wisconsin, same pattern.
Here is Alabama, rapid post-holiday growth, then slowing over the past few days.
And Tennessee… Rapid decline over the second half of November, and now up again.
Here is Ohio – significant post-holiday bump.
Here is Indiana, up after Thanksgiving, but now flat to declining.
And finally, here is South Dakota, way down from the top and continuing to decline.
So that’s it for today. I’ll report again on Monday or Tuesday.
–Shane Chalke, FSA