Hacker News Reader

Comment by danuker | Hacker News Reader

I am stunned so few governments publish such "in the wild" data.

The only way to judge the effectiveness of measures is to see how many people have actually been infected, not to count "confirmed" cases.

For instance, it seems only 8.8% is confirmed (or 88k out of 1 million) [1] . Add that to the 69% vaccinated with at least one dose [2], and you get 77.8% accounted for, which is less than the 95% quoted.

[1] - https://ourworldindata.org/explorers/coronavirus-data-explor...

[2] - https://ourworldindata.org/explorers/coronavirus-data-explor...

View on HN

The CDC publishes that data.[1][2] It runs 6 weeks behind, though.

[1] https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/abou...

[2] https://covid.cdc.gov/covid-data-tracker/#national-lab

Are these numbers comparable to OP? It's hard to reconcile that FL is at 24%, or NY at 13% w/ antibodies vs the headline.

Your implication is correct that the CDC chart is mostly incomparable to the OP. The chart tries to combine incomparable data, which annoyingly requires interpretation of the legend to understand correctly.

The circles indicate nucleocapsid (i.e., infection only) while the squares indicate spike (i.e., infection or vaccination).

So, Puerto Rico is 70% infected and/or vaccinated (a square). Whereas NY is 13% infected (a circle).

Note that the recent numbers in the CDC chart all seem to be measuring antibodies that only appear from infection, not vaccination (indicated by a circle instead of a square in the chart), while the numbers from the article appear to be measuring antibodies that appear from either infection or vaccination, so they are not directly comparable.

Yes. Which is annoying. Both can be measured, but apparently the commercial lab data only runs one of the two antibody tests.

There's also a blood bank antibody test. That's useful as a crosscheck. People who have blood drawn for medical purposes are in general sicker than average, and those who donate blood are healthier than average. So those set bounds on the measured values.

> I am stunned so few governments publish such "in the wild" data

[warning: tinfoil hat...] Can you think of an incentive for governments to be honest with such "in the wild" data?

If you were a senior politician trying to assist your government in driving up the vaccination rate, then publishing hard data that 94% of your population already have the antibiodies is hardly going to help, is it?

> The only way to judge the effectiveness of measures is to see how many people have actually been infected

Can anyone provide links to recent peer-reviewed primary sources that estimate the number of individuals who have been infected in the U.S.?

If ~70% of people are asymptomatic, and we assume just about everyone has been exposed to and infected by the virus - haven't we already achieved herd immunity in the U.S.?

EDIT: Answered in solid source provided by animats - the seroprevalance estimate for the U.S. is 21.6% (95% CI - 21.2% to 22.22%). I also found a couple recent primary sources regarding the asymptomatic rate, see thread below.

This assumes infection leads to immunity. It doesn't seem to be clear that this is true.

Here are some primary sources which support the fact that infection leads to robust and durable immunity.

[1] SARS-CoV-2 infection induces long-lived bone marrow plasma cells in humans https://www.nature.com/articles/s41586-021-03647-4.pdf

[2] Necessity of COVID-19 vaccination in previously infected individuals https://www.medrxiv.org/content/10.1101/2021.06.01.21258176v...

I know three people who got it twice and another who is a friend of a friend. Only one of those is known to be immunocompromised. Two of those got it the second time after also being vaccinated. I do wonder though if some of the second cases aren't people going in with the flu and having some covid fragments in their nose.

> I do wonder though if some of the second cases aren't people going in with the flu and having some covid fragments in their nose.

This is why using PCR tests cranked up to absurd cycle counts isn't the best way to determine _infection_.

though nations like Australia that did flu checks last year estimated it was almost absent, and of those that tested positive very small percentages were vaccinated, so if its the flu, the covid vaccine is the best flu vaccine ever.

N=1, but I now know many people who have been vaccinated that have since been infected, where as up until recently I knew 2 people who got it this entire time (I'm in SF). One of these new people is a co-worker who got covid last year, then got Pfizer vaccinated, then got it again and almost was hospitalized. He's 29 and fit, and as far as I know, is not immunocompromised.

All the data shows that vaccination provides protection, but with all these new variants coming out (expect more over time), I don't believe herd immunity is any more possible than herd immunity with the flu or common cold.

> I don't believe herd immunity is any more possible than herd immunity with the flu or common cold.

This seems to me to be a very strange thing to say.

We know precisely what the pattern is for herd immunity in the case of a coronavirus, because four of them cause what you are calling a "common cold". It's endemic equilibrium. To my knowledge, there is no dispute among experts on this point.

So to say that we can't achieve herd immunity because of its similarity to a virus against which we are indeed herd immune is, at best, begging the question.

If you have six minutes, this is a very acute description of the endemic equilibrium effect of herd immunity: https://www.youtube.com/watch?v=RsmDJYhteQs

It's the default assumption given practically every other known virus, and is the core assumption of all epidemiological modeling. Somehow all what is known from epidemiology and virology has been thrown out of the window during COVID. If SARS-CoV-2 is somehow involved the sky is falling on us on any moment unless otherwise proven. Bizarre.

Don't know if they have a part in this, but at least it's extremely profitable for pharma to assume there's no immunity but what they provide.

The fact that more people are in the hospital in Florida right now than at any time in the last two years would seem to disprove any speculation of here immunity.

How many have been infected twice?

Can we stop with this please?

> The fact that more people are in the hospital in Florida right now than at any time in the last two years

THIS IS NOT TRUE.

The metric you are describing is hospital census, which is _not_ at an all-time high right now.

What is at an all time high? The percentage of people in the hospital who have tested positive for SARS-CoV-2, regardless of their presentation.

The roundtable just yesterday confirmed that there are hospitals which have near 100% positivity, despite 80% presenting for things unrelated to a respiratory infection.

Some people are there for blunt trauma or routine surgery, but have tested positive for SARS-CoV-2.

I just don't understand the point of constantly misrepresenting this metric. It muddies the waters of actual concerns about the spread of the virus by making it seem like those are are concerned are dishonest.

Stop it.

I think the media is at least partly to blame for this. I have the greatest respect for the BBC, but look at this headline: Covid third wave: Florida surpasses all-time record for hospital admissions. Totally unsupported by the article. https://www.bbc.co.uk/news/world-us-canada-58077209

With regards to metrics that are peaking, FT reports that "Occupancy in the south-eastern US state’s intensive care units is on course to hit an all-time high by this weekend." (But then again, I've no idea whether this is actually accurate since they don't cite their source. Besides, it's clearly an extrapolation.) https://www.ft.com/content/21ed4e9c-2b20-4a9d-bf38-6b76212be...

The first line of the article is "The number of Covid patients in Florida hospitals has risen to a new high, breaking records set during previous waves before vaccines were available."

Then, there's a quote to that effect from a doctor, specific to their hospital.

Right. The headline is about hospital census; the article does not support that point, and instead is about a different (and much less important, from a public health perspective) metric - admissions which are coincident with a positive test for a specific pathogen.

"The Florida Hospital Association on Monday reported 10,389 Covid-19 hospitalizations, the most statewide during any point in the pandemic. "

https://www.politico.com/states/florida/story/2021/08/02/flo...

> I just don't understand the point of constantly misrepresenting this metric. It muddies the waters of actual concerns about the spread of the virus by making it seem like those are are concerned are dishonest.

> Stop it.

That must be this "projection" I keep hearing about.

I don't understand. It sounds like we agree. The quote you've pulled confirms what I'm saying.

Hospital census is not what's at an all-time (or 2-year) high. SARS-CoV-2+ tests among admitted patients ("Covid-19 hospitalizations") are.

"People in the hospital" and "Covid-19 hospitalizations".

These are not the same thing.

>I just don't understand the point of constantly misrepresenting this metric. It muddies the waters of actual concerns about the spread of the virus by making it seem like those are are concerned are dishonest.

>Stop it.

It's hopeless. The exact same thing happened last year, driven off of media that loved talking about how the local hospital is at the brink of collapse because "the ICU is 90% full". The few voices that pointed out that 90% ICU occupancy is a) normal and b) what hospitals plan for were overwhelmed.

The only time US hospitals anywhere were actually overrun with patients was a brief time in NYC in March 2020. Even that cleared up quickly; the US military hospital ship sent to NYC was barely used and soon left, and I don't believe the one sent to LA was used at all. All those tents built in hospital parking lots everywhere? Not used either.

If we assume everyone has been infected? Why would we assume that? The estimates I've seen (including asymptomatic) are around 1/3 infected. The source you cited is from 2020, so it doesn't tell us much about today.

Yup good catch about the old source, I removed that from my comment.

The source provided by animats puts the seroprevalence at 21.6% for the U.S. I think that should be a lower bound of the number of individuals who have been infected, given antibody levels wane over time (even though immunity does not necessarily wane).

The CDC estimates that 36% of Americans have been infected. Of those, only 15% were completely asymptomatic but most had mild symptoms.

https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/burd...

Thanks for pointing that out and doing the math.

I did some digging, here is the paper outlining the methods and data sources for the numbers on the page you linked to [1]. That paper cites two other papers used to estimate the asymptomatic proportion [2][3].

Here is a more recent systematic review [4], which estimates that at least 33% of infections are asymptomatic.

And another recent paper which estimates the lower bound for asymptomatic proportion (AP) to be 37%, with an upper bound of 67% [5]. "The estimated AP varied with age, from 74% (95% CI 65–90%) for < 20 years, to 61% (95% CI 57–65%) for the 50–59 years age-group." [5]

[1] Estimated Incidence of Coronavirus Disease 2019 (COVID-19) Illness and Hospitalization—United States, February–September 2020 https://academic.oup.com/cid/article/72/12/e1010/6000389

[2] Estimating the extent of asymptomatic COVID-19 and its potential for community transmission: Systematic review and meta-analysis https://jammi.utpjournals.press/doi/pdf/10.3138/jammi-2020-0...

[3] Age-dependent effects in the transmission and control of COVID-19 epidemics https://www.nature.com/articles/s41591-020-0962-9

[4] The Proportion of SARS-CoV-2 Infections That Are Asymptomatic https://www.acpjournals.org/doi/full/10.7326/M20-6976

[5] Estimating the asymptomatic proportion of SARS-CoV-2 infection in the general population: Analysis of nationwide serosurvey data in the Netherlands https://link.springer.com/article/10.1007/s10654-021-00768-y

Asymptomatic is kind of a fuzzy concept so the numbers will vary based on methodology. If you test positive and then have a bit of diarrhea one day is that a COVID-19 symptom or the consequences of a bad burrito? In mild cases there's no way to be sure.

https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/s...

I agree there can be many confounding factors.

To clarify your point: this leads you to the conclusion that the asymptomatic proportion is likely under-estimated? And conversely, the symptomatic proportion is likely over-estimated?

Remember this includes vaccinated persons

Yep, just realized and edited.

Assuming the vaccination was successful in triggering the production of antibodies. It is unclear to me in what percentage of cases that is the case.

Hong Kong requires 2 weeks quarantine even for vaccinated people entering, unless they have proof not only of vaccination, but of antibody count.

The immune response from vaccines is generally far less variable than the immune response from fighting off an infection naturally.

Reply on news.ycombinator.com