1. In the op-ed pages of newspapers and on cable news shows, I'm seeing frequent confusion over what it means to reach herd immunity—and whether we can relax COVID precautions once we do so.

It will take a few posts, but let me try to explain.
2. The key thing to note is that the herd immunity threshold is the point at enough people are immune (by vaccination or previous infection) to prevent a new epidemic from starting from scratch.

It is *not* the point at which an ongoing epidemic disappears.
3. When you reach herd immunity, a pandemic is far from over. In fact, in a basic SEIR model, an ongoing epidemic is at its *peak* when the herd immunity threshold is reached.
4. The problem is that while a new epidemic can no longer start from scratch once you reach herd immunity, when we reach this point for COVID we'll still have the old epidemic underway—and epidemics have something akin to momentum.
5. All of the people who are currently infected will continue to transmit disease after you reach the herd immunity threshold. They just will infect fewer than one additional person, on average. The people they infect? Those will infect others downstream. And so on.
6. This gives rise to what we call "overshoot"—additional cases above and beyond the herd immunity threshold.

@nataliexdean and I wrote an OpEd about this almost a year ago.
nytimes.com/2020/05/01/opi…
7. Another thing to keep in mind is that the herd immunity threshold depends on the transmissibility of the disease—which depends on behavior.

For example, in many places we could already be reaching the herd immunity threshold for a masked, socially distant world.
8. But that does not mean we've reached the herd immunity threshold for a world in which everyone returns to how things were in 2019.

In other words, as you open up, the herd immunity threshold shifts higher.
9. The other thing to recognize is that implementing non-pharmaceutical interventions (masking, social distancing, gathering restrictions) around the herd immunity threshold is a *very* efficient way to reduce the total size of the epidemic, by reducing the overshoot.
10. We can see the basic principle in a simple SEIR model. Implementing aggressive controls for even a short period around the herd immunity threshold reduces the overshoot and prevents many cases that would have occurred without controls.
11. For this reason, it's a big mistake to open up right when you're reaching the herd immunity threshold but before the number of current cases—the momentum of the pandemic, so to speak–is low. This will generate a larger overshoot and lead to many preventable infections.
12. In the above I have used simple SEIR models to illustrate the basic concept of herd immunity during an ongoing pandemic, and the notion of overshoot.

In practice, we're dealing with a complex situation where these lessons hold but the dynamics are more complicated.
13. In particular, we have

a) ongoing vaccination as well as natural immunity
b) ongoing behavioral changes
c) highest-risk groups already vaccinated
d) increasing cases due to variants of concern that spread at higher rates than the previous strains.

These all interact.
14. So I'm not aiming to predict exactly what the next couple of months will look like.

My main point: Reaching herd immunity is necessary but not sufficient to relax control measures. To do that, you also want to have a low number of cases to avoid overshoot.
15. Finally, I'm already being accused of moving the goalposts. For the record, the goalposts were always here. Below, a thread I wrote nearly a year ago explaining these same points and stressing the need for controls around the herd immunity threshold.

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Carl T. Bergstrom

Carl T. Bergstrom Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @CT_Bergstrom

6 Apr
1. People have a lot of questions about the concept of herd immunity and what it means for COVID-19. What is it, can we get there, what does it get us if do, etc. In this thread, I'll try to answer some of these.
2. Q: What *is* herd immunity?

A: When a large fraction of a population is immune to a disease, a *new* outbreak of that disease can no longer take off and spread. The herd immunity threshold is the point at which enough people are immune for this to be the case.
3. Q: How does it work?

A: An outbreak grows from one case to many when each person who is infected transmits disease to more than one other person on average.

Conversely, if each person transmits to fewer than one person on average, an outbreak will quickly fizzle out.
Read 31 tweets
2 Apr
New arXiv preprint: Kathy Brauer's, “I’ll Finish It This Week” And Other Lies

tl;dr: We suck at estimating how long it will take us to do things that don't have deadlines, and we don't improve much over the course of a career.

arxiv.org/pdf/2103.16574… Image
Serious point: academia would grind to an absolute halt if it were not for closely related cognitive distortions.

Example: It's 1:30 AM and I'm working on a promotion letter that is due tomorrow.

When I agreed to do it, I figured "Why not, I won't be busy then like I am now."
I was wrong, of course.

And that's generally true of ever manuscript or grant review I agree to write, every seminar I agree to give, every collaboration I agree to add to my plate....
Read 6 tweets
30 Mar
I find it odd that the @IHME_UW would choose to advertise its broad impact on the US Covid response by tweeting a picture of Deborah Birx and the IHME model predicting that the pandemic would go to zero with 100% probability by July 2020.

And indeed one year later there are indeed many unanswered questions.

Most importantly: what happened, why were the serious (and ultimately correct) concerns expressed by much of the research community ignored, what has been learned, and what is going to be different in the future?

Read 5 tweets
28 Mar
In his latest paper about COVID infection fatality rates, John Ioannidis does not address the critiques from @GidMK, but instead engages in the most egregious gatekeeping that I have ever seen in a scientific paper.
John's defenders have done this in the past, but I'm stunned that he'd stoop to the same.

Science doesn't work like that, to say the least. Gideon's degree status is irrelevant and in the entirety of my career I've never seen this issue raised in a scientific paper before.
The condescension and hypocrisy here is mind-boggling.
Read 7 tweets
25 Mar
There's an odd website going around right now that purports to point out twitter accounts that have a left-wing bias.

Now, I don't deny my own left-of-center leanings.

But I think the algorithms need work.

I mean, consider their report on known leftist @megynkelly.
Let's look at what they say about me.

I lean further left than Megyn.

But why? Here are my top influencers.
One of these accounts, @callin_bull, is an account I run.

Another, @stephaniemlee, is one I do retweet.

The third is really weird. @janweider is a very talented photographer who I follow, but have interacted with online only *once*, as shown below.
Read 6 tweets
24 Mar
1. When developing COVID testing protocols, it
is critical to know the sensitivities of alternative COVID testing methods and how they change over the course of an infection.

But this is hard to measure, particularly for individuals who are pre-symptomatic or asymptomatic.
2. A new preprint from UIUC provides some of the best data I've seen, comparing antigen, saliva-based PCR, and nasal swab-based PCR.

medrxiv.org/content/10.110…
3. Instead of using the onset of symptoms as "day zero", they use the onset of culturable virus from nasal swab. Doing so, you get the following sensitivity curves.
Read 8 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!