4 ms·

These are very good points. I'd add that herd immunity isn't a clearly defined threshold, either. It's the point at which the effective R in a population is < 1

These are very good points. I'd add that herd immunity isn't a clearly defined threshold, either. It's the point at which the effective R in a population is < 1 -- AKA exponential decay. With a circulating virus, depending on how much below 1 you get, exponential decay might take a few months or even a few years to get down to elimination. The current dip in the UK is trending toward the former. But it's further complicated by the fact that levels of immunity in the population will continue to oscillate up and down. (Nevermind that once schools open in September, 20% of the population with relatively sparse vaccination will be mixing in large cohorts). Another factor in herd immunity is vaccine efficacy - which is somewhere between 86% and 95% for the mRNA vaccines (and will continue to change). A herd immunity threshold of 90% presupposes a vaccine with 100% efficacy - when your vaccines are not that effective (not to mention, losing effectiveness over time) it bumps the threshold even higher. Right now, much of the population has some level of "sterilising immunity" from the first round of vaccines and previous infection (i.e. having enough circulating antibodies to potentially block symptomatic infection and transmission). But everything we know about existing respiratory viruses (and indications so far this year) suggest antibodies will wane. Of course, indications are most people will come away with robust cell-based immunity. (Those plucky T-cells and B-cells). But eradication for the UK looks *very* close to impossible given the near-total-abandonment of public health measures recently.

nicoburns 5y ago
> eradication for the UK looks very close to impossible given the near-total-abandonment of public health measures recently. I don't support the recent abandonment of restrictions, but could that policy, followed up by stricter restrictions in a few months time actually be helpful for eradication? Eradication is likely a hopeless prospect in the UK anyway, because there is no political will for closing the borders. But it's interesting to conjecture.
- XorNot 5y ago
  The ship has sailed on eradication. We're ALL going to get Delta at some point now, it's just a question under what circumstances (and whether the antivaxxers will have sufficient numbers to denial of service the healthcare system as they die).
  - dragonwriter 5y ago
    > We're ALL going to get Delta at some point now, That's by no means certain; even ignoring the possibility of effective countermeasures, a more transmissible variant could displace it (if we’re lucky, it could also be one producing less severe disease, the way it was when Epsilon variant for a while displaced most others in California, before Delta, which is more transmissible than Epsilon, took hold.)
    - manwe150 5y ago
      Tangential question: what is the mechanism by which we observe ‘displacement’ of earlier strains? Is it just the exponentially higher transmission factor of the variant that is making it relatively disappear? Or are they somehow also get out-competed directly?
      - XorNot 5y ago
        The former: once you have COVID you're immune to it. Unless a strain totally evades the antibody response of recovered individuals (very unlikely), then they're no longer capable of spreading it. So if there's overall less replicants of one strain then the other, then the population will effectively select for the more infectious one. Actual geography of course is also key here though: if a carrier flies over to an unexposed area with any strain, it's going to take over if the other one doesn't make it there.
lamontcg 5y ago
> But everything we know about existing respiratory viruses (and indications so far this year) suggest antibodies will wane. Not necessarily the case. Infections with HCoV-229E produces durable immunity and reinfection is due to antigenic drift: https://www.biorxiv.org/content/10.1101/2020.12.17.423313v1 https://www.biorxiv.org/content/10.1101/2020.12.17.423313v1 That it at odds with prior studies that found ~12 month reinfections with coronaviruses, but it is with considerably better methodology. And it is simultaneously: not clear that vaccines produce sterilizing immunity, not clear it is necessary to do that at all, and not clear if how much measurements of NAbs correlates with sterilizing immunity. I can tell you a story about how delta reaches 90-100% thresholds of herd immunity in this wave and burns itself out and we don't see it much again for 3-5 years in countries like the US/UK. Or I can tell you a story about how it keeps coming back every year with 2 or 3 waves. Anyone who tells you they know which of those outcomes must be correct probably knows a lot less about virology than they think they do (even if they have a degree). I'm reasonably hopeful its the former, but I never predicted delta and thought we'd be virus-free this summer and was wrong. Even in the worst case though the vaccinated/recovered are never going back to 2019 levels of susceptibility to severe outcomes, so we're looking at a steady drip-drip-drip of people killing themselves to avoid vaccination up until we get bored with that and they pretty much disappear (that's where T-cells and B-cells and cross reactive partial immunity to all future variants comes in).
- sharken 5y ago
  Setting up two likely scenarios is a very reasonable way of approaching future development, both scenarios sound likely to me. We could look at the flu for clues, in that case every year will have one or more variants that are the most prominent. I'd argue that in the age category with the most infected (10-29 years old) in the UK, that these people are not killing themselves due to their young age. And the elderly have already gotten the vaccine and are thus up to eight times more likely to not get infected.
  - lamontcg 5y ago
    > We could look at the flu for clues Like the 1918 H1N1 second wave mutation that caused ARDS in 20 year olds? We still have no idea how that happened, but with this coronavirus evolving to be more transmissible/virulent that is still on the table. If it can evolve to become more transmissible but at the cost of triggering ARDS like that, then it will face evolutionary pressure to do so. And hospitalization rates of 20-29 year olds aren't anywhere near as good as the fatality rates. While the death rate drops by a factor of 2 every 8 years, the hospitalization rate drops by a factor of 2 every ~15 years. And delta has 2x the hospitalization rate of alpha across age/sex/deprevation/comorbidity matched controls. It should still be about a 1-in-50 chance of hospitalization for 20-29 year olds. There's no proof that number can't continue to get worse.

nicoburns 5y ago

> eradication for the UK looks very close to impossible given the near-total-abandonment of public health measures recently. I don't support the recent abandonment of restrictions, but could that policy, followed up by stricter restrictions in a few months time actually be helpful for eradication? Eradication is likely a hopeless prospect in the UK anyway, because there is no political will for closing the borders. But it's interesting to conjecture.

XorNot 5y ago
The ship has sailed on eradication. We're ALL going to get Delta at some point now, it's just a question under what circumstances (and whether the antivaxxers will have sufficient numbers to denial of service the healthcare system as they die).
- dragonwriter 5y ago
  > We're ALL going to get Delta at some point now, That's by no means certain; even ignoring the possibility of effective countermeasures, a more transmissible variant could displace it (if we’re lucky, it could also be one producing less severe disease, the way it was when Epsilon variant for a while displaced most others in California, before Delta, which is more transmissible than Epsilon, took hold.)
  - manwe150 5y ago
    Tangential question: what is the mechanism by which we observe ‘displacement’ of earlier strains? Is it just the exponentially higher transmission factor of the variant that is making it relatively disappear? Or are they somehow also get out-competed directly?
    - XorNot 5y ago
      The former: once you have COVID you're immune to it. Unless a strain totally evades the antibody response of recovered individuals (very unlikely), then they're no longer capable of spreading it. So if there's overall less replicants of one strain then the other, then the population will effectively select for the more infectious one. Actual geography of course is also key here though: if a carrier flies over to an unexposed area with any strain, it's going to take over if the other one doesn't make it there.

lamontcg 5y ago

> But everything we know about existing respiratory viruses (and indications so far this year) suggest antibodies will wane. Not necessarily the case. Infections with HCoV-229E produces durable immunity and reinfection is due to antigenic drift: https://www.biorxiv.org/content/10.1101/2020.12.17.423313v1 https://www.biorxiv.org/content/10.1101/2020.12.17.423313v1 That it at odds with prior studies that found ~12 month reinfections with coronaviruses, but it is with considerably better methodology. And it is simultaneously: not clear that vaccines produce sterilizing immunity, not clear it is necessary to do that at all, and not clear if how much measurements of NAbs correlates with sterilizing immunity. I can tell you a story about how delta reaches 90-100% thresholds of herd immunity in this wave and burns itself out and we don't see it much again for 3-5 years in countries like the US/UK. Or I can tell you a story about how it keeps coming back every year with 2 or 3 waves. Anyone who tells you they know which of those outcomes must be correct probably knows a lot less about virology than they think they do (even if they have a degree). I'm reasonably hopeful its the former, but I never predicted delta and thought we'd be virus-free this summer and was wrong. Even in the worst case though the vaccinated/recovered are never going back to 2019 levels of susceptibility to severe outcomes, so we're looking at a steady drip-drip-drip of people killing themselves to avoid vaccination up until we get bored with that and they pretty much disappear (that's where T-cells and B-cells and cross reactive partial immunity to all future variants comes in).

sharken 5y ago
Setting up two likely scenarios is a very reasonable way of approaching future development, both scenarios sound likely to me. We could look at the flu for clues, in that case every year will have one or more variants that are the most prominent. I'd argue that in the age category with the most infected (10-29 years old) in the UK, that these people are not killing themselves due to their young age. And the elderly have already gotten the vaccine and are thus up to eight times more likely to not get infected.
- lamontcg 5y ago
  > We could look at the flu for clues Like the 1918 H1N1 second wave mutation that caused ARDS in 20 year olds? We still have no idea how that happened, but with this coronavirus evolving to be more transmissible/virulent that is still on the table. If it can evolve to become more transmissible but at the cost of triggering ARDS like that, then it will face evolutionary pressure to do so. And hospitalization rates of 20-29 year olds aren't anywhere near as good as the fatality rates. While the death rate drops by a factor of 2 every 8 years, the hospitalization rate drops by a factor of 2 every ~15 years. And delta has 2x the hospitalization rate of alpha across age/sex/deprevation/comorbidity matched controls. It should still be about a 1-in-50 chance of hospitalization for 20-29 year olds. There's no proof that number can't continue to get worse.