Email may have tainted Stanford coronavirus antibody study | Hacker News Reader

Email may have tainted Stanford coronavirus antibody study

sfgate.com

46 pointsMilnerRoute33 comments

Side question:

The very first comparison of covid-19 and seasonal flu I read (before covid-19 was called covid-19) said that covid-19 would have a mortality rate of 1.5% and seasonal flu of 3%, therefore covid-19 would be less deadly. In recent months I often read the same statement turned around, which is already confusing. Now here we talk about a tenth of the same numbers turned around.

I mean, either it's not standardized what "mortality rate" means or the numbers are all over the place and one can't really say which is more deadly.

Could someone with more knowledge in this area summarize the current status and maybe where the confusions might come from?

Maybe loose terminology so ending up getting case fatality rate confused with mortality rate?

It's the difference between IFR and CFR. The IFR is the infection fatality rate meaning everyone in the population who has been infected. The CFR is the case fatality rate, those who are sick enough to show up at a hospital, or well, get tested and confirmed positive. Which nowadays is essentially the same and a small fraction of total cases.

We see here that flu seems to have an overall IFR rate of 0.002%. Many orders of magnitude smaller than COVID19.

https://www.cdc.gov/nchs/fastats/flu.htm

Where do you get IFR for covid19?

The IFR rate for covid-19 seems to be trending towards 0.1-0.5% based on the serological tests over entire population cohorts that are starting to be published.

I thought it was between 0.5% to 1.0%, at least according to the NYC study: https://twitter.com/trvrb/status/1253436656742305793

By doing serologic studies you can estimate the global number of contaminated people: looking at the number of viral mutations, as well as different strands of the virus that can be identified as being generated in a specific region.

It’s going to be very imprecise. For H1N1 the global infection rate could have varied from 11% to 21% of the global population, which is a +/- 350 million people estimation.

https://www.cidrap.umn.edu/news-perspective/2011/08/study-pu...

0.002% is again a different number than 3%, 1.5%, 0.3% and 0.15%.

Also to compare two numbers don't you need a second one as well? 0.00002 is a magnitute smaller than what number?

The IFR rate for covid-19 seems to be trending towards 0.1-0.5% based on the serological tests over entire population cohorts that are starting to be published.

According to the CDC, the seasonal flu kills 12,000-79,000 Americans a year by the same standard as used for COVID-19. Much closer to the widely cited 0.1% IFR.

https://www.cdc.gov/flu/about/burden/faq.htm

Seems high. Here is a systemic review of studies done on the 2009 influenza coming to a conclusion of an IFR of 1-10 deaths per 100 000 infections.

"The consistent estimates of the infection fatality risk at around 1 to 10 deaths per 100,000 infections identified in our review"

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3809029/

CEBM concurs, the H1N1 CFR was between 0.1% and 5.1%, based on country and the fatality risk settled out at 0.02%. [1]

Some more napkin math, the CDC estimates between 12,000 and 61,000 flu fatalities per year out of approximately 9,300,000 to 45,000,000 infections so pretty consistently a 0.1-0.15% IFR. [2]

Not sure how to reconcile these. The worst year, 2017-2018, had 45M cases and 61K deaths. The dominant strain was H3N2 not H1N1. The vaccine wasn't particularly effective that year (38%). Not sure how else to account for the delta. [3, 4]

[1] https://www.cebm.net/covid-19/global-covid-19-case-fatality-...

[2] https://www.cdc.gov/flu/about/burden/index.html

[3] https://www.cdc.gov/flu/about/burden/2017-2018.htm

[4] https://www.cdc.gov/flu/vaccines-work/effectiveness-studies....

> Deaths per 100,000 population: 2.0

Your 0.002% number is the POPULATION Fatality Rate not the Infection Fatality Rate (IFR).

Remember that IFR of a disease alone does not tell you how many people ultimately die from it. The R0 number, which determines how quickly the diseases spread, determines total cases and then cases X IFR = deaths.

My understanding is that Covid-19 has a much higher R0 than the seasonal flu.

One thing that didn’t seem to be noted in the rest of the comments: the actual mortality rate can’t be formally calculated until the pandemic has subsided. What you are seeing now are estimates and projections for the number that they’ll end up being able to calculate when the disease has stopped spreading.

If someone said the flu has a 3% ifr (infection fatality rate) they were way off.

The best guess for the covid-19 IFR I've seen is ~1%. New York City has has .02% of it population die above the base death(excess deaths). And the last serological study said 20% of people have been infected. So that gives us an approximation of 1%.

This is 10x-100x worse than the flu.

> New York City has has .02% of it population die above the base death(excess deaths)

Where did you get that number? The numbers I’ve seen are several orders of magnitude higher.

2% (2 orders of magnitude above 0.02%) of 8 million (approximately the population of NYC) is 160,000 people. So that's clearly wrong being more than 3 times the entire US deaths.

The numbers you've seen are probably deaths of infected people not the entire population.

> The numbers you've seen are probably deaths of infected people not the entire population.

No, I just slightly misread OP’s confusingly worded stat. Base deaths in NYC is around 6k/month. Covid-19 deaths alone have been way above that number. Numbers are obviously not in for the month of April yet, but it’s looking more like it’ll be closer to 30k deaths in NYC per CDC[0]. So that’s 24k more deaths per month, aka 5x (where my several orders of magnitude came from). If you use the scale OP used, that’s about .3% of NYC population in excess deaths, so one order of magnitude higher.

[0] it’s at 26k right now with a bit under a week left - https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm

Excess deaths in New York City: 19,200.

Population of New York City: 8.4 million.

19,200 / 8.4 million 0.22% dead.

20% infection rate = IFR of ~1.1%

https://www.nytimes.com/interactive/2020/04/21/world/coronav...

Yes, I know those numbers. That’s not what OP said that I was responding too (see reply to sibling comment for more context).

> COVID-19 antibody testing - FREE

How did that email manage to get through any spam filter?

Sent by <name>@stanford.edu with host email server verifying the email came from there?

It sounds like it was sent from the researchers wife who did not work for stanford.

The problem with the Stanford study, as I understand it, is that the 95% confidence interval on the false positive rate includes 1.5%, meaning that it is not more than 95% certain that anybody among the people tested was not a false positive.

Yesterday:

https://news.ycombinator.com/item?id=22972658

> A 0.35% mortality rate is still nothing to sniff at, however, as it is three to four times deadlier than the seasonal flu (with a mortality rate of 0.1%)

Which means a response three to four times as strong as the one to the seasonal flu would be appropriate.

> and can be significantly higher for elderly residents and individuals with underlying conditions.

Which doesn't make the pandemia any more dangerous because the risk is correspondingly lower for everyone else. In fact, it's a good thing because we could focus our efforts on protecting those who are in actual danger. We could use our functioning economy so they can avoid unnecessary contacts and to help them cope with the increased isolation, making sure they aren't part of the ~70% that will inevitably be infected.

Could.

>We could use our functioning economy so they can avoid unnecessary contacts

What economy?

https://fortune.com/2020/04/23/us-unemployment-rate-numbers-...

The US now looks worse than Greece did in 2008.

Because of the misguided response to Corona. So at least for the US, "could have" may be more correct.

> Which means a response three to four times as strong as the one to the seasonal flu would be appropriate.

I'm not sure that's a meaningful statement. Our response to a seasonal flu - for which we spread a vaccine in advance and have a population with scattered levels of resistance due to similarities to previous iterations - cant be "done but 3 to 4 times more".

No, but we can spend 3-4 times more/accept 3-4 times the inconvenience/etc. responding to the flu is worth to us.

Except that math doesn't work either. 3-4 times the morality rate is not the same as 3-4 times as many dead, for the same reasons I listed above - we've already made lots of societal adjustments to change the input number before the mortality rate of the seasonal flu gets involved.

The flu has vaccines, and anti-viral treatments. If it did not the reaction to it might be very similar to how we're being told to respond to COVID-19. These are not comparable things because the options for combating them are drastically different.

Reply on news.ycombinator.com