This is a very good write up, where he goes through how he develops the estimates for real infections, deaths, etc. with the available data. He's had the highest quality model in this whole thing, so it is very interesting to read through his approach.
https://covid19-projections.com/estimating-true-infections/

Some sections that people will find interesting here.

Quote:

---

Lower IIFR Over Time
The IIFR in the US decreased from over 1% in March to 0.25% in July. Below, we present a few explanations to why the IIFR in the US has decreased significantly since March/April.

• Lower median age of infection (see section above)
• Better protection of vulnerable populations (nearly half of COVID-19 deaths in March/April were in care homes)
• Improved treatment (new drugs, better allocation of resources, more experience among staff, etc)
• Earlier detection

The above are explanations that would explain a true decrease in IFR. We believe the lower median age of infection and better protection of high-risk populations are the primary drivers behind the decrease in IIFR. Below are some reasons that could skew the IIFR lower, but not change the true IFR:

• More comprehensive reporting of confirmed cases
• Changes in the distribution of age groups tested (e.g. more younger people getting tested would skew IIFR down)
• Inflation of the test positivity rate (e.g. double-counting positives, not reporting negatives, etc)
• Longer lag in death reporting
• Underreporting of deaths

---

Quote:

---

Herd Immunity Threshold
Looking at the data, we see that transmissions in many severely-impacted states began to slow down in July, despite no clear policy interventions. This is especially notable in states like Arizona, Florida, and Texas. While we believe that changes in human behavior and changes in policy (such as mask mandates and closing of bars/nightclubs) certainly contributed to the decrease in transmission, it seems unlikely that these were the primary drivers behind the decrease. We believe that many regions obtained a certain degree of herd immunity after reaching 10-30% prevalence.
A widely-accepted method to calculate the herd immunity threshold (HIT) is to use the basic reproduction number, R0:
HIT = 1 - 1/R0
. Back in March/April, we estimate R0 in the US to be around 2.3. This corresponds to a HIT of
1-1/2.3 = ~0.6
, or 60%. But the effective reproduction number, Rt, has decreased dramatically since then due to a variety of reasons such as greater population awareness, mask-wearing, reduced larger gatherings, and implementation of social distancing guidelines. The Rt in most regions around the US where there are outbreaks is now between 1.1-1.6. This corresponds to an effective HIT of 10-35%. As a result, it makes intuitive sense that we are seeing a decline in transmission after those regions reach a 10-35% prevalence.
One thing to note is that original definition of the herd immunity threshold is derived from the basic reproduction number, R0, and assumes no intervention/social distancing. Hence, by definition, the HIT of the SARS-CoV-2 virus remains unchanged over time, between 50-80%. But the effective HIT does change over time as the effective reproduction number, Rt, decreases due to society adjusting to the virus. That's why we are seeing an effective HIT of 10-35%.
Also note that reaching the herd immunity threshold does not stop transmission - it simply slows down further transmission.

---

Interesting! Thanks for posting.

I think many may gloss over this key fact:

Also note that reaching the herd immunity threshold does not stop transmission - it simply slows down further transmission.

Sometimes it seems people use "herd immunity" to indicate "the point at which it will stop spreading". I could be misinterpreting, but that's the tone I think I hear.

When the new case curve drops off one thing seems guaranteed - any threat to overwhelm a healthcare system is gone. While still spreading, it is slow and the system can easily handle it.

Also

Quote:

---

But the effective reproduction number, Rt, has decreased dramatically since then due to a variety of reasons such as greater population awareness, mask-wearing, reduced larger gatherings, and implementation of social distancing guidelines.

---

So to have a HIT of 10-30%, you have to maintain all of these things like mask wearing, social distancing, reduced social gathers, etc.



Ever step you take to open it up, pushes the HIT higher and higher until you get back to 60%

I think he also has it right on the herd immunity. The herd immunity assuming no intervention of any kind is ~60%. However, the implied herd immunity WITH interventions is going to be lower. That's probably why people are seeing the 10-20% "burnout". However, it's the interventions that are lowering the R_t from 2.3 to something closer to 1, even if they aren't mandated by governments -- working from home, not shaking hands, social distancing, no large events, larger awareness, testing, etc. If you were to remove those interventions, then the implied herd immunity threshhold would go up again.

I was hopeful that there was some cross immunity or other thing making a large percentage of the population not susceptible. However, as some epidemiologists have pointed out:

• If half the population is not susceptible, then the R_0 is actually twice as high
• If R_0 is twice as high, then the herd immunity threshhold amongst the susceptible population is higher
• So, in the case with 50% not susceptible, the herd immunity required would actually be 40%, not 30%

His explanation also better explains areas around the country that have blasted through the 10-20% "burnout", such as Northern Italy. It makes sense that an area that had it's breakout with much higher R_t due to lower awareness and mitigations would have gotten to higher population prevalence -- 57% in Bergamo, Italy for instance.

Agreed. For me personally, I first noted NYC cases and deaths dropping off to very low numbers and then that antibody study came out that said 20-25% of NYC may have antibodies. It was an observation (very few new cases and 20-25% have some immunity), wasn't sure why, but did note something similar in some other regions - small countries or states. But I wasn't sure if it was due to changing behaviors (distancing, masks, fear) or due to some immunities we didn't yet understand - certainly hoped for the latter. And when some studies have later come out indicating people may be able to fight this off due to their previous exposures to coronaviruses, or for other reasons, my hopes went up. But I still come back to the fact that many behaviors have changed and I can't buy the "viruses gonna virus" statement, which I actually find quite dumb (shingles, HIV, ebola, rabies , the common cold are all viruses and they all don't act the same). The most likely explanation for dropping cases would be a combination of factors, I think - yes, some have had the virus and now are not susceptible, yes some people are able to fight it off with no effect, some have it with no symptoms and don't get caught as a confirmed case, and also many people - certainly not all, I know - have changed their behaviors.

Whatever the case may be, dropping to 1% positivity rates, like in NYC, should allow things to reopen, and schools to teach kids in person.

Imo there where clear policy changes in July in Texas, we implemented mask ordinances, closed bars etc

No, that's not the case.

Sweden does not employ masks and has bars and restaurants open and their HIT was in the ~20% range.

The 60% calculation assumes a homogeneous social graph and a truly naive population, neither of which is true.

Take it up with the author of the article.

Quote:

---

Changes in human behavior and policy interventions such as mask mandates also contribute to a slowing of the spread. If current interventions and social distancing are relaxed, the effective herd immunity threshold would go back up.

---

Quote:

---

I think he also has it right on the herd immunity. The herd immunity assuming no intervention of any kind is ~60%.

---

I don't think that's correct at all.

Again, the simplistic 60% calculation assumes homogeneity and no pre-existing immunity.


Do you think the social graph is homogeneous?

Do you think there is no pre-existing immunity?

Keegan99 said:

---

No, that's not the case.

Sweden does not employ masks and has bars and restaurants open and their HIT was in the ~20% range.

The 60% calculation assumes a homogeneous social graph and a truly naive population, neither of which is true.

---

Are you saying there is no mask wearing in Sweden or just that the government doesn't dictate you must wear one?

They did also make changes to bars and restaurants, but did also keep them open. Only table service, no bar or counter service, and table distancing. Some restaurants chose to close, on their own, at least temporarily due lack of customers. Far fewer governmental restrictions, no doubt, but nothing is really binary - we weren't totally locked down everywhere, Sweden did not have zero changes in behaviors.

Sweden has no mask mandate and while some citizens do wear them, they are not common.

None of the Nordics are big on masks.

Duncan Idaho said:

---

Also

Quote:

---

But the effective reproduction number, Rt, has decreased dramatically since then due to a variety of reasons such as greater population awareness, mask-wearing, reduced larger gatherings, and implementation of social distancing guidelines.

---

So to have a HIT of 10-30%, you have to maintain all of these things like mask wearing, social distancing, reduced social gathers, etc.



Ever step you take to open it up, pushes the HIT higher and higher until you get back to 60%

---

Excellent. Self-quarantine the vunerable and open it up. We are so freaking stupid not doing this yet.

cone said:

---

Quote:

---

While still spreading, it is slow and the system can easily handle it.

---

that will be true even after vaccine development and roll out

you will get covid, in all likelihood, during your lifetime

it's endemic and it's not going away

---

which begs the question, when do we go back to school, sports, etc.? I know it has some nuance, but at some point people will have to understand that having a couple hundred cases at a time in TX is okay.

I have tried to do a lot of reading on the epidemiologists who are predicting the 10-20% herd immunity for a naturally selective virus versus a vaccine scenario which may have the more typical HIT formula resulting in 60%, so I can be better educated.

The paper is below:
https://www.medrxiv.org/content/10.1101/2020.07.23.20160762v1.full.pdf

I have tried to read a lot of the author's defense and discussion around the topic:
https://twitter.com/mgmgomes1

This is the best thread I found dissecting it from an epidemiologist (from a critical standpoint):

And here is the best response I found to those arguments:

I am convinced that the HIT is probably lower than the traditional model, since it makes sense that the traditional model would apply well to randomly selected vaccines, whereas a naturally selected disease propogation would have a different behavior, infecting more susceptible people first and breaking down social chains.

A lot of this comes down to the homogeneity or lack there of of the social mixing in the population. What Wes argues is that some edge cases of high seroprevalence could be explained by a combination of naturally higher R_0 in densely populated areas and also overshoot (HIT does not imply that infections magically stop and the level of infection when HIT is reached implies how many future people will be infected.

What I think ties back to my original understnading is that we have certain social networks/mixing right now in our current semi-quarantined status. For example, in Houston I am not going to work, but I still occasionally see my family or neighbors. We try to be more mindful of distance around each other and hugging/shaking hands. We are not making as many shopping trips or exposing ourselves indoors, and when we do we wear masks. Other people are doing it differently, some only leave home to go grocery shopping once a week and some people are out and about seeing people every day and regularly mixing. So today, we have a certain level of social network/mixing and heterogeneity that has developed as a result of the virus. I do believe that it's possible under these circumstances, there could be a much lower HIT.

In some imaginary world, where we told everyone to go back to normal and forget the virus, those social networks and heterogeneity would change. I would be exposed to my coworkers every day and since I'm in sales going to customer meetings and such I would be a far more effective vector for transmission. You would probably consider me one of the more susceptible people. So, heterogeneity would increase and the HIT would increase and the level of infection would go up.

You're probably seeing this right now in Spain where they probably hit the HIT at their current level of infections (~10% of the total population), but then once they re-opened and encouraged tourism to return, they increased the number of social interactions and homogeneity and the infections returned:


I am not advocating for a shutdown or anything and I want schools opened, I think it just remains important to be vigilant on avoiding the virus in the easy to control ways -- wear a mask, wash hands, keep distance where possible, etc.

This week I found out my cousin, who for her own reasons (which I understand) is a bit of a COVID denialist and has been one of the more cavalier people I know with regards to the virus. Last week, she lost her sense of taste and smell along with some conjestion. Her son (4) was experiencing the same. Despite that, she assumed that she just had a cold and 2 days after her first symptoms she went to go visit my mom and dad while still symptomatic and got my parents to babysit her son while she went to go workout in a gym. My dad has T2 diabeties with 4 stents in his heart. Not just that, but 2 days after that, my mom was supposed to go visit her mom who is in her 80s and has a littany of comorbities. Once my mom feared that she might have caught the virus from my cousin and her son, thankfully she cancelled their visit. My mom started getting symptoms yesterday. To me, this personal situation perfectly encapsulates the issue that I want to express -- people just need to be careful. My cousin is a young healthy person who will not have any severe consequences from the virus and neither will her son. But, the least she could do is not visit my parents when she's sick, especially with tell-tale symptoms of loss of smell and taste, not to mention her visit to the gym. I'm hopeful that everything will be fine, but it just highlights that at a baseline we should all take the virus seriously and do the simple/easy things to protect others. I'm not asking people to lock themselves away forever, just be careful and thoughtful.

So if the IFR is really 0.25%, then based on the number of deaths in Texas, that implies that Texas has had over 3 million real infections, which is right around 10% of the population. You have to imagine that the percentage of people that have had it is actually a lot higher in the major metro areas and lower in rural areas. So we may be close the the HIT in many places. If this is true, then what will happen in the next couple of months is that you'll see a decline in Dallas, Houston, San Antonio, Austin, but maybe some mini outbreaks in smaller cities.

HotardAg07 said:

---

I have tried to do a lot of reading on the epidemiologists who are predicting the 10-20% herd immunity for a naturally selective virus versus a vaccine scenario which may have the more typical HIT formula resulting in 60%, so I can be better educated.

The paper is below:
https://www.medrxiv.org/content/10.1101/2020.07.23.20160762v1.full.pdf

I have tried to read a lot of the author's defense and discussion around the topic:
https://twitter.com/mgmgomes1

This is the best thread I found dissecting it from an epidemiologist (from a critical standpoint):

And here is the best response I found to those arguments:

I am convinced that the HIT is probably lower than the traditional model, since it makes sense that the traditional model would apply well to randomly selected vaccines, whereas a naturally selected disease propogation would have a different behavior, infecting more susceptible people first and breaking down social chains.

A lot of this comes down to the homogeneity or lack there of of the social mixing in the population. What Wes argues is that some edge cases of high seroprevalence could be explained by a combination of naturally higher R_0 in densely populated areas and also overshoot (HIT does not imply that infections magically stop and the level of infection when HIT is reached implies how many future people will be infected.

What I think ties back to my original understnading is that we have certain social networks/mixing right now in our current semi-quarantined status. For example, in Houston I am not going to work, but I still occasionally see my family or neighbors. We try to be more mindful of distance around each other and hugging/shaking hands. We are not making as many shopping trips or exposing ourselves indoors, and when we do we wear masks. Other people are doing it differently, some only leave home to go grocery shopping once a week and some people are out and about seeing people every day and regularly mixing. So today, we have a certain level of social network/mixing and heterogeneity that has developed as a result of the virus. I do believe that it's possible under these circumstances, there could be a much lower HIT.

In some imaginary world, where we told everyone to go back to normal and forget the virus, those social networks and heterogeneity would change. I would be exposed to my coworkers every day and since I'm in sales going to customer meetings and such I would be a far more effective vector for transmission. You would probably consider me one of the more susceptible people. So, heterogeneity would increase and the HIT would increase and the level of infection would go up.

You're probably seeing this right now in Spain where they probably hit the HIT at their current level of infections (~10% of the total population), but then once they re-opened and encouraged tourism to return, they increased the number of social interactions and homogeneity and the infections returned:


I am not advocating for a shutdown or anything and I want schools opened, I think it just remains important to be vigilant on avoiding the virus in the easy to control ways -- wear a mask, wash hands, keep distance where possible, etc.

This week I found out my cousin, who for her own reasons (which I understand) is a bit of a COVID denialist and has been one of the more cavalier people I know with regards to the virus. Last week, she lost her sense of taste and smell along with some conjestion. Her son (4) was experiencing the same. Despite that, she assumed that she just had a cold and 2 days after her first symptoms she went to go visit my mom and dad while still symptomatic and got my parents to babysit her son while she went to go workout in a gym. My dad has T2 diabeties with 4 stents in his heart. Not just that, but 2 days after that, my mom was supposed to go visit her mom who is in her 80s and has a littany of comorbities. Once my mom feared that she might have caught the virus from my cousin and her son, thankfully she cancelled their visit. My mom started getting symptoms yesterday. To me, this personal situation perfectly encapsulates the issue that I want to express -- people just need to be careful. My cousin is a young healthy person who will not have any severe consequences from the virus and neither will her son. But, the least she could do is not visit my parents when she's sick, especially with tell-tale symptoms of loss of smell and taste, not to mention her visit to the gym. I'm hopeful that everything will be fine, but it just highlights that at a baseline we should all take the virus seriously and do the simple/easy things to protect others. I'm not asking people to lock themselves away forever, just be careful and thoughtful.

---

I would have come the **** unglued on anyone who irresponsibly and selfishly exposed my family to this. How horrible. It takes very little educating of oneself to realize how stupid of a move that would be. Even if you're "regular sick" you don't go to people's houses until you're better. I don't understand why you understand her reasoning, personally.

Covid will be mostly gone in Texas by end of August. Maybe not the cities that border Mexico but the rest , yes.

total lunacy

Duncan Idaho said:

---

Also

Quote:

---

But the effective reproduction number, Rt, has decreased dramatically since then due to a variety of reasons such as greater population awareness, mask-wearing, reduced larger gatherings, and implementation of social distancing guidelines.

---

So to have a HIT of 10-30%, you have to maintain all of these things like mask wearing, social distancing, reduced social gathers, etc.



Ever step you take to open it up, pushes the HIT higher and higher until you get back to 60%

---

Posts like this remind me of how my neighbor has reacted to Covid. I've lost a lot of respect for my neighbor.

Go hide in your hole. Let the rest of us live our lives.

.25% and dropping, comparable to .1% for the common flu

I. Can't. Even.