Don't people who've been infected count toward the herd immunity. That ~30% of Americans. I know there's some overlap but if we can get to 70% of people vaccinated I think we'd be really close, no?
The flaw with these kinds of statements is that it assumes the 70% is uniformly distributed throughout a population. Obviously, if an entire town decides to avoid vaccination, it'll be vulnerable. Less obvious is if certain social circles don't get vaccinated such that things look geographically well mixed, but in terms of actual interaction patterns, not so much.
Reinfection and new variants surviving in networks like this imply COVID will be around for a very long time.
Israel has 60% vaccinated and they have controlled the pandemic through that. The US probably has more natural immunity than them so overall I’d say we are in a decent position. The issue will be localized outbreaks in places with less vaccine coverage.
There have been high-profile cases of re-infection in India, and the experience from Manaus tells the same story. Expecting herd immunity from letting the disease spread is wishful thinking.
It's important to understand why reinfection might not be such a big problem.
First, the virus can only spread broadly with an R0 above 1. Even if some portion of the population is susceptible to reinfection (vaccinated or natural immunity), so far there's been little evidence that it's widespread. In a fully vaccinated population, the R0 will only reach above 1 if the virus is able to break through vaccines wholesale.
Second, the vaccines seem to be highly effective at preventing serious cases, hospitalizations, and deaths, even in reinfected individuals.
There has been 70 cases and 3 deaths of reinfection worldwide [0]. If this source is not false, reinfection is negligible, and I am yet unconvinced herd immunity will not materialise
No matter what you'd see greatly reduced spread. Ideally everyone would be vaccinated, but it would be unwise to ignore the substantial overlap between antivax people and antimask people. Everything pushing R beneath 1 is a good thing that should be counted.
