CV: Stanford Antibody study results are out ...
Figured this was worth a new thread - the first truly rigorous and large CV antibody test results from anywhere in America.
This is Santa Clara County, the SF South Bay. 3330 residents tested in a county of about 1.9 MM.
Key takeaway - estimated that 2.49%-4.16% of the population there has been infected by CV by early April (the study was done on April 3-4) already. That prevalence estimate represents a range of 48,000 to 81,000 people infected in Santa Clara County by early April, which is 50x-85x the number of actually confirmed cases.
Link to the paper:
https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v1.full.pdf
Of course, that is just Santa Clara County, but it gives us another estimate of the ratio between actual and confirmed cases. I've always been pegging that number at 10x-30x. This ratio is obviously larger than that.
More and more antibody studies to come in the time ahead. I think a similar Stanford study conducted down in SoCal (as opposed to NorCal) is next, and should be out within days.
April 17th, 2020 at 12:01 PM ^
Not being an epidemiologist I don't really know what to make of this. That is really high for number of actual cases implied. But also a low number for the percentage of the population that's been infected. Seems far from being herd immunity. Of course they're probably a lot closer to that in NYC.
Also, not gonna make a separate post for it, but China revised its Wuhan death toll 50% higher. For all of our disagreements, I think we can agree the Chinese government sucks balls.
April 17th, 2020 at 12:04 PM ^
I had the same reaction. If I'm interpreting this correctly, the low percentage implies that we may be closer to a vaccine than herd immunity.
I don't know why people speak of this dichotomy between a vaccine and herd immunity. Herd immunity is normally achieved via a vaccine. To try to get it otherwise is to accept a massive death toll, which the British government realized (after wasting valuable time).
It's also not clear how long the post-infection immunity would even last.
Remember though that California has a relatively low per-capita death rate among states.
New York's per-capita death rate is ~20x higher than Santa Clara's. If ~2.5-4% of the Santa Clara population has the virus, that would suggest 50-80% of New York's population has already contracted the virus. If true, New York is already completely over the hump and could lift their lockdown immediately without seeing any increase in death rate.
Other states are going to need to be far more cautious than New York though.
April 17th, 2020 at 12:05 PM ^
There was a theory going around that the Bay Area was doing so well with coronavirus because they may have gotten it super early (think Nov or Dec) and not diagnosed it. This is yet another example of wait for the science before proclaiming one doctor's thoughts as fact. There is no herd immunity.
April 17th, 2020 at 12:21 PM ^
If I understand correctly, the COVID death rate is a known number. And the number of confirmed cases are known by testing, which we know is very limited. So this study says the actual number of people infected is 50-85 TIMES the number of confirmed cases? If that's correct then the death rate of those infected is MUCH lower than previously thought. So COVID would then be much quicker spreading but far less fatal (in percentage terms) that has been thought up to now.
I'm not making any specific point other than trying to understand the implications, if accurate, of the Stanford data. One of the hallmarks of this pandemic is that nobody really knows anything for sure. We are flying blind and relying at best on educated guesses and at worst on rumor and speculation.
April 17th, 2020 at 12:31 PM ^
That was my takeaway, but the only thing I hate more than legal documents (just wrapped up my refinance) are scientific research documents, so I could be misinterpreting. Would be really interesting to see similar studies in more hard hit areas like northern Italy and NYC and also areas with less severe distancing guidance like Sweden.
April 17th, 2020 at 12:42 PM ^
Yes it is good news on this front, but not groundbreaking. The death rates thrown around have always assumed that we are missing cases, but not not 50x. For example, right now around 5% of the US cases have ended in death and this doesn't account for those currently sick who may sadly pass away.
Caveats: - we are likely missing deaths
- it is one (affluent) area.
- the study has to control for age/income demographics which always makes the confidence interval a bit wider than reported.
In most places, they’re adding deaths by counting anyone who dies and tested positive as a Covid death, even if it wasn’t actually the cause of death. This week, NYC added nearly 4,000 deaths to the total for which no tests were ever done and the virus was only suspected to have been present.
I think that for the 3800 they added to the count, it's not just that it was suspected to be present but that it was suspected to be the cause of death. But regardless, the bottom line is that *something* is killing thousands more New Yorkers than usual. Did you see the graphs in this article?
You're calling this an affluent area (I don't know the region). Wouldn't this make it likely that the infection rate is even higher in big cities because of population density and all the reports we're hearing of less discipline on social distancing? That would certainly be good news.
April 17th, 2020 at 12:49 PM ^
I have been saying that for months.
also said that the posted stats on mortality (total deaths divided by total number of infected) is the WORST case scenario at that specific date, as there are likely multiples of people who got it and were asymptomatic.
I was told I was wrong ny a number of people who apparently don’t understand math or the idea of best case worst case scenarios, called names, and negged to shit.
thats why I just say we’re all gonna die and those that survivor will be back in the Stone Age.
Except for OSU. If you're going to go worst case go big or go home.
We'll all be in the stone age. Except for OSU. Which will be fine and get even more recruits.
also said that the posted stats on mortality (total deaths divided by total number of infected) is the WORST case scenario at that specific date,
Only if you're looking at closed cases. A lot of people are trying to calculate mortality by dividing total cases (most of which are still active) by deaths. That assumes that 100% of active cases will result in recovery, which of course isn't going to happen.
I think you’re making a flawed assumption when you say we “know” the death rate. Why do you think the death count isn’t underreported? Why would they test corpses when we don’t even have enough testing for the symptomatic? I think if anything, the death number is likely underreported due to the widespread lack of testing capacity. The lack of testing applies to deaths, too— there are many cases where the person dies and they were likely CoVid positive but were never tested.
April 20th, 2020 at 12:45 PM ^
What makes you think the true case load isn't grossly under-reported, much more so than any loack of accurate reporting on deaths? These same Stanford researchers are also on record saying that the real infection rate is 30 to 50X the reported cases. Its tough to inflate death numbers as you are counting bodies. The people who are infected and never bother to get reported are immensely more likely to be missed
Deaths/Recovered is currently at about 26%.
So that means this would be projecting a death rate of about 0.3-0.5% which is pretty close to initial estimates I heard when this all started.
This is also quite consistent with the number of 0.37% from the Gangelt study in Germany
I found it to be interesting that countries in Europe are estimating that roughly half of their total deaths have occurred in elder care homes.
This would suggest that a Swedish approach but with even more aggressive steps to protect the elderly would allow for quicker herd immunity, less economic despair and allow us to achieve a lower IFR as the higher IFRs are mainly driven by the extremely high death rate among the elderly.
The COVID death rate is far from a known number. Just as there are (living) people who had it and didn't know, there are also people that died because of it and didn't know since we still don't have enough testing and we haven't bothered testing dead people (or at least didn't when I last read about it). The unknown deaths are probably a smaller % today as we are able to test more but who knows the backlog
If I understand correctly, the COVID death rate is a known number.
How could it be known when the disease's existence has only been known for a few months?
I am pretty sure from the context of the post the poster means number of deaths.
Either way, it isn't a known number, it's just the data we have at hand right now. How complete/incomplete this data is remains to be seen.
The CDC will come up with an official estimation of COVID-19 deaths, but it won't be until after the epidemic is finally over, when it makes an extrapolation of the data to come up with a final range of deaths (as it does for influenza and other viruses).
Right now, trying to figure out the "real" death rate is chasing a moving target as the data keeps changing. China just quietly acknowledged an extra 1200 or so deaths in Wuhan, for instance. And France started counting COVID deaths outside of a hospital setting a little while ago, which caused its death toll to soar. Other countries, including the U.S. for the most part, are only counting deaths in hospitals.
The death rate is an estimated number, very far from being “known.“
Its exaggerated high by hospitals being encouraged to count presumptive positive cases due to testing issues, while also under counted by those who pass outside the hospital without being tested.
April 17th, 2020 at 12:06 PM ^
The thing about herd immunity is that we don't know if, or for how long, you're immune to this thing once you've had it and survived. I've seen several reports of people testing positive after having been cleared of the virus. Were they reinfected? Does it just go dormant and then reactivate? Is it something that stays with you forever and comes back when your immune system is weak?
Many questions need answering.
April 17th, 2020 at 12:23 PM ^
I've seen this but also repeatedly seen doctors say they think this is just due to testing errors. Having said that, no one is sure if the immunity lasts. They think there is an immunity to it, but it's also possible that only lasts for 6 months or a year or whatever.
April 17th, 2020 at 12:36 PM ^
One other issue to factor in that analysis is false negatives. We know little about the sensitivity of the nasal swab PCR testing, but there are some reports that up to 30% of the negative tests are false negatives due most likely to inadequate test administration and sample collection. The swab needs to go deep in nose to back of throat to get sample, and it can cause coughing and gagging that makes the health care worker who is administering the test to pull back early.
It's possible some of these people who were cleared originally were actually positive in the first place. Besides lack of widespread testing, this could be an issue when the country starts to reopen. We need to either shift to an easier method for test sample collection(saliva, etc.) or provide better training to test administrators if we are missing that many positive cases.
April 17th, 2020 at 12:47 PM ^
"We don't know how long you can be immune for" is something that's technically true, but it seems to me the medical community is sandbagging like hell on this.
All our lives we've been told that exposure and sickness from a particular virus makes you immune to that same virus again for, effectively, the rest of our lives. You get chicken pox once and never again. (Before the vaccine, anyway.) Measles too. Lifelong immunity is what makes vaccines work. If viruses commonly reinfected people and caused symptoms again, vaccines would never have existed. If it's possible to be symptomatic a second time with this virus, and I mean a real re-infection, not intensive testing that discovered remnants of it in someone who previously tested negative, it would truly be unique in the annals of virology.
Doctors are saying what you've said because they can't say with 100%, rock-solid, unassailable certainty that it's not true. But in all likelihood, it's not especially different from every other virus known to humans.
I'm not a science talking guy, or a sawbones, but I think it all depends on how many different strains or mutations of the virus there are.
We have to get a flu shot each year to account for slightly different strains of the virus. And even if the strain is the same year to year, the immunity gained from being vaccinated or fighting the disease off doesn't last forever.
Hopefully this one turns out to me more like the measles than the flu.
However if you get sick from that strain of flu you won't get sick from ever again.
False - experts have not been “saying that all our lives.” I agree that some people often describe pathogen immunity as lifelong, but it’s never been true. It’s just a common and honest mistake of the lay community. Nothing lasts forever and some immunity never gets off the ground in the first place.
This can be true for many reasons:
(a) pathogen mutagenesis outpaces a patient’s immune system the first time (e.g., most battling HIV never develop immunity, even across the multi-decade battle),
(b) or because the pathogen mutates too quickly between a patient’s first and next infections for that same species of virus or bacteria to be recognized by that B cells, T cells and/or antibodies that responded last time (e.g., many viruses that cause common colds and seasonal influenza change this quickly between exposures)
(c) or because, even in certain cases where a viral genome is somewhat stable, sometimes our immunity fades -- be it because T cells & B cells can of course die of old age or because circulating antibody levels must ultimately decay -- and that means that some pathogens will make a return (e.g., dormant Varicella commonly makes comebacks in the elderly, categorized as shingles, after having been defeated when the disease was categorized as chickenpox.
And of course this isn’t an exhaustive list. The point being that our immune system is neither perfect at pattern recognition, nor are its successes forever. As with everything in the universe, entropy goes to work on it. Things fall apart.
And experts have been confirming that for decades.
/medical doctor and biochemist
Thanks for that explanation Dr. I was immediately thinking of shingles when the poster you replied to said we assume immunity lasts a lifetime.
Isn't it far too early to make any estimates or assumptions whether COVID acts in similar fashion? I guess I'm not clear where this whole sidebar about people getting "reinfected" came from? We don't have even a clue how many have been infected, much less whether any are being reinfected.
April 17th, 2020 at 12:07 PM ^
Yeah, it's a long way from herd immunity - but at the same time it helps us better determine the Infection Fatality Rate, as opposed to the Confirmed Case Fatality Rate. Which is important in its own right.
Santa Clara's case load has gone up ~ 1.6x times since April 3-4, when this study was conducted. So theoretically that range could be 3.5%-6% as of today. Community spread is on-going even in the current conditions.
It's just one study. Interested in seeing more.
April 17th, 2020 at 12:15 PM ^
I think it's interesting data but I can already see some misinterpretations on the thread.
MAJOR CAVEAT: they are testing exposure to SARS-CoV-2 virus that causes COVID-19, they are NOT testing immunity in the population. The presence of a circulating antibody tells you one, but not the other.
This is an often confused point and I think the reporting on these kinds of science issues around the pandemic has been poor, or at least could be communicated much better.
I should sit down this weekend and do the equivalent of a "neck sharpies" on ABC's of antibody tests, natural immunity, vaccines etc and post it on the board, I think it could really help everyone interpret some of the lab news that is being firehosed at you.
Still, interesting data from the local team (I'm down the road from Palo Alto).
April 17th, 2020 at 12:17 PM ^
How does one test immunity then?
April 17th, 2020 at 12:26 PM ^
Not sure you can. We don't even know for sure if people who were 100% sick and recovered are truly immune or if they are, for how long.
April 17th, 2020 at 12:27 PM ^
Great question. At an individual level, it's tricky, but the real answer is the question has to be addressed to the population level.
For the individual, you would have to do a much more involved experiment that wouldn't be practical for high-throughput assays and definitely not for home testing. Basically, you would have to use viral particles that infect target cells (Type II Pneumocytes from the lungs) in sample 1, and viral particles plus target cells plus plasma from the subject in sample 2, and measure the decrease in viral infectivity. A dramatic drop-off in the ability to infect a cell in vitro would be a reasonable (not 100% conclusive) way to determine the subject is likely immune. At that time. Tough to say what happens months or years later for reasons that would require a whole neck sharpie on immunologic memory.
For the population, you measure immunity the old fashioned way. How many people are getting sick relative to how many people are getting exposed? That's old hat to epidemiologists who study the flu every year and it's reliable because of the numbers involved.
April 17th, 2020 at 12:18 PM ^
In case you needed any small show of support as motivation, I would really appreciate such a post.
April 17th, 2020 at 12:19 PM ^
I think that would be a very worthwhile post, as long as you're able to objectively view and relate the data.
How predictable. Timmay posts an Ad hominem attack. Literally. Every post.
I'm the hominem and I didn't feel attacked.
You have serious cognitive issues, guy. I was voicing support for what he said he wants to do. I seriously feel bad for you, and absolutely loathe you at the same time.
April 17th, 2020 at 12:26 PM ^
you should do the neck sharpie
April 17th, 2020 at 12:35 PM ^
No one should do the "neck sharpie." Stupid name.
April 17th, 2020 at 12:35 PM ^
I really have appreciated your contributions to these threads. One of the few posters whose posts actually help inform vs just pushing personal viewpoints on limited understanding of the underlying science.
April 17th, 2020 at 12:45 PM ^
Oh man, I'm as mouthy and opinionated as anybody around here, I just have a carve-out for science because I revere it and don't want to see it devolve into politics.
Other than that, I've got a lot of problems with you people, and you're going to continue to hear about them. Every day is Festivus on MGoBlog.
I'm also tipping my hat to you. When this is over, if you live anywhere in Michigan, the beer's on me. You have absolutely earned it.
That would be interesting and very Mgoblog.
Maybe Brian could include that as part of HTTV.
I'm curious what to make of the study participants. The study did not indicate if the participants were selected randomly in any way, just that they were recruited on facebook and asked a series of questions. Could it be that the participants who actually responded to the advertisement were not an accurate representation of the actual population, since perhaps they may have had reasons to believe they were already exposed or had had the disease? It seems to me that information gathered from randomly selected study participants would be more accurate than information collected from participants that sought out testing by clicking on an ad, as the latter could skew the the information.
April 17th, 2020 at 12:57 PM ^
China’s death total they are reporting is still a blatant lie. Probably over 100k.