Bad link in original post, fixed now.
If you just stop and think carefully about what you are observing and what the meaning might be, you can come up with good insights on any subject. I have tried to do that with this epidemic from the start. One of my biggest concerns about the extreme lockdowns was that they were ordered rashly, with no real appreciation of what the epidemic would be like, notwithstanding the clues that were there from the start. One concern I had about the modeling in particular, and the assumptions about the course of the epidemic, was that it just didn’t appear that everyone got infected, even after close and heavy exposure. The Diamond Princess cruise ship was one of the earliest signs of this. (And don’t give me the crap about quarantining passengers on the ship. By the time that happened everyone had to have multiple exposures.) That meant that models that assume some random sampling of the population by the virus were going to be erroneous. No model that I am aware uses a dose or susceptibility variable in its contact or other formulas. It appeared to me that the epidemic would be front-loaded with very susceptible individuals, so that early infection, hospitalization and death rates should not and could not be extrapolated to the entire population. But everyone did the opposite and panicked and literally destroyed the world economy, inflicting incalculable human misery.
A new paper validates this theory. (Medrxiv Paper) The authors, a multi-country, prominent research team, titled the paper in a manner that tells the story: “Individual variation in susceptibility or exposure to SARS-CoV-2 lowers the herd immunity threshhold.” (I don’t like the term “herd” immunity; I know we are acting like sheep, but we are human after all. I use population immunity.) Currently people think 60% to 70% of the population must be exposed to reach population immunity. These authors project a lower number, potentially as low as 10%, depending on the variability in susceptibility. They used a SEIR model (modeling progression from susceptible to exposed to infected to recovered) but incorporated a variation in susceptibility term. They tested their models on the course of the epidemic to date in Italy and Austria to explore what variability that susceptibility term might have. They tested values between 0 and 4. Amazingly to me, although this concept of variation in individual susceptibility to infectious disease has been explored in the literature for other agents, I have not seen it discussed in regard to the current coronavirus epidemic, and as I said, it has not been incorporated into the models.
It is truly hard to overstate the importance of this paper and the need for modelers and policymakers to address its implications.