My guess is IFR will settle around 0.7 - 0.8%

It will depend on age distribution more than anything. If I recall, median age of Italy is almost 10 years higher than US. Could be something like 1.3% there and 0.8% here.

When the seroprevalence studies find something around 1-2%, I find it hard to trust the data that much just because of issues with sensitivity and specificity make a huge impact on the result. For example, all positives could be a false positive if the test found 2% positive in a test with 2% false positives. They do inform a higher end, though.

The studies that have more widespread infection rates are really interesting, like Spain, Italy, NYC, etc because the error rates are small by comparison. The Spain one seems well done.

Diyala Nick said:

---

One thing I have not heard discussed is the idea of there being a higher fatality rate earlier in the course of the pandemic. By that I mean, if nursing homes produce a disproportionate portion of total deaths, and they are prone to infection, then as the infection progresses, there will be fewer "candidate" nursing homes succeptible to infection. It would seem that this will substantially decrease the overall fatality rate over the course of the pandemic.

---

Good question. I'm also curious if there will be any noticeable effect on the fatality rate as doctors are becoming better at treating. Some of the early studies on convalescent plasma use and some of the anecdotal info from the doctors that post here seem to indicate greater success in preventing death. I wonder if it will be enough to start showing up in the overall fatality rate soon or if it will take a really big breakthrough in treatment to see a noticeable change.

Rachel 98 said:

---

Diyala Nick said:

---

One thing I have not heard discussed is the idea of there being a higher fatality rate earlier in the course of the pandemic. By that I mean, if nursing homes produce a disproportionate portion of total deaths, and they are prone to infection, then as the infection progresses, there will be fewer "candidate" nursing homes succeptible to infection. It would seem that this will substantially decrease the overall fatality rate over the course of the pandemic.

---

Good question. I'm also curious if there will be any noticeable effect on the fatality rate as doctors are becoming better at treating. Some of the early studies on convalescent plasma use and some of the anecdotal info from the doctors that post here seem to indicate greater success in preventing death. I wonder if it will be enough to start showing up in the overall fatality rate soon or if it will take a really big breakthrough in treatment to see a noticeable change.

---

Great point. A few modest gains in treatment can have a big impact on lives saved. If convalescent plasma turns out to be highly effective, it will be a game changer.

Keegan99 said:

---

How did Miami get to 10% without a huge fatality count?

That's not in any way trying to be argumentative. It's bewildering. Miami-Dade has had 500 deaths in a county of 2.7 million, but have a whopping 10% of the population with antibodies.

That's something akin to a 0.2% IFR.

---

The drive-thru testing methodology will skew the data toward people who believe they've had symptoms. To get a more accurate number you need a truly random sample of the population.

Moxley said:

---

Keegan99 said:

---

How did Miami get to 10% without a huge fatality count?

That's not in any way trying to be argumentative. It's bewildering. Miami-Dade has had 500 deaths in a county of 2.7 million, but have a whopping 10% of the population with antibodies.

That's something akin to a 0.2% IFR.

---

The drive-thru testing methodology will skew the data toward people who believe they've had symptoms. To get a more accurate number you need a truly random sample of the population.

---

I'm curious how they use this antibody test data to estimate. Surely they don't say that 10% of tests come back positive and then apply that to the whole population, right?

It was a random sample in Miami. It was not people reporting for a test because they had symptoms.

Keegan99 said:

---

How did Miami get to 10% without a huge fatality count?

That's not in any way trying to be argumentative. It's bewildering. Miami-Dade has had 500 deaths in a county of 2.7 million, but have a whopping 10% of the population with antibodies.

That's something akin to a 0.2% IFR.

---

Vitamin D? Interesting that Miami and California both show lower symptomatic rates than anyone expected. Maybe the fact that you can be outside year round comfortably - including outdoor cafes and shopping - means that the general population has a much higher level of D than other places.

And I'm not talking about hot climates, I'm talking about comfortable climates. I live in Houston and literally don't go outside for 7 months of the year because it's miserable. So no D for me.

Anyway, just a random thought.