ryankemper
commented
4 years ago Thanks Jerry.
I appreciate the feedback and agree with basically all your points. Only thing is I might not include the Stanford studies since there was a lot of controversy around the false positive rate / calibration techniques used - basically I think the studies have validity but they're more disputable than the other serological studies.
Your 3 areas of improvement
(1) I agree the pareto mitigation section needs to be fleshed out more.
I'm about to start the next iteration right now, so I'll try to implement as much of this as I can.
(2) That's a good point, perhaps it could simply be described as the intersection of https://en.wikipedia.org/wiki/Loss_aversion and basic game theory.
(3) Yeah, one of the goals for this next iteration that I've already got slated is to do a better job characterizing the "containment until vaccine" approach. So that would logically fit there.
There is one more thing I need to account for. Even ignoring the vaccination / improved treatments (I'm not super bullish on treatment - I think we'll get small wins like with Remdesevir but won't get a true game changer), the area under the curve is not the same if the alternative is full containment. The reason for this is the phenomenom known as "overshoot"; pareto mitigation will require going a bit above the herd immunity threshold before the epidemic essentially stops. (And just to be clear I believe SARS-CoV-2 will always be endemic and will join the regular cast of characters like the various Influenza strains, Adenoviruses, etc that we already deal with each year). Which reminds me - I should probably explicitly mention that eradication is completely infeasible because a lot of people seem to think it's possible. (To me it seemed so obvious that a zoonotic-originated global pandemic is impossible to eradicate but I realize most people don't understand that)
Your 2 omissions
(1) Yeah, I'd like to work in Sweden eventually. I don't currently have a ton of knowledge in this area to be able to do it justice, but will definitely be watching what unfolds in Sweden. My assumption is that their timeline will be shifted forward - i.e. they'll accumulate infections/deaths sooner - but end up far ahead once accounting for the lack of shutdowns (their economy has taken a massive demand hit but they didn't shoot themselves on top of that demand hit the way the US did)
(2) Totally agreed, this absolutely needs to be included. It's worth mentioning that for those practicing containment, PCR tests are still very valuable, even with the window. But logistically it's still super infeasible.
The benefit of pareto mitigation / natural herd immunity is that we really don't care about PCR testing except on a diagnostic basis to help doctors make decisions, except insofar as we need to use PCR testing to protect our nursing homes, etc. i.e. we should understand that they have a limited window of reliability, but they're still better than nothing when it comes to a population who has a >= 12.5% chance of dying if they're exposed.
--
This is great feedback. Thanks again @JerryLopatin!
I think this is a fantastic article. If I were still hiring software engineers I'd hire you in an instant for your clear thinking and communication skills.
I have 2 areas that i think could be improved:
Section 4, Pareto Mitigation, is the most important going forward. You want to substantially expand on this, tying it to IFR and the Dutch / Italian fatality distributions. Speaking of IFR, you also want to reference the Stanford and USC studies showing both community exposure and IFR, since that's US data. The New York data would also strengthen your arguments.
Section 5, you'll want to expand on why public officials act the way they have. I don't think it's Hanlon's razor, but the other point you allude to, that the penalty for not appearing to act aggressively is high, while the penalty for over-reacting is almost non-existent, particularly among public health officers. They are actually acting in their political self interest, not necessarily the public's.
Section 3, you say that the area under the curve remains the same. Yes, that is true until either an effective mass treatment is discovered, or an effective vaccine is deployed. I agree with you that both are long shots in the near term, but you need to at least mention this in this section.
There are 2 omissions that I'm surprised you left out:
The example of Sweden, where they have accelerated towards herd immunity, accepting that there will be higher death rates now.
The testing fallacy - that we need a lot more PCR tests to re-open. No one has even tried to adequately explain how mass testing changes the situation, given (1) PCR tests have a 20-30% false negative rate, (2) a person may be most contagious when they are asymptomatic, in the 1st couple of days after infection, and that's when PCR tests often give a false negative result, (3) a person's infection state can change daily, so that logically, everybody would need to be tested daily, and (4) there is no logistical scenario for 10 million PCR tests a day in the US within a couple of years, much less 100+ million. PCR testing is good for sampling, something that we're not even doing well because we mainly test symptomatic or scared people.
Again, great work!