Days after Stanford University researchers issued an early draft of a study suggesting that up to 81,000 residents of Santa Clara County had been infected by COVID-19 as of early April, a team at the University of Southern California (USC) released its own serological study that similarly concludes that the disease is far more widespread — and less deadly — than official estimates indicate.
The study by USC and the Los Angeles Department of Public Health concluded that between 2.8% and 4.6% of the adult population in Los Angeles County has an antibody to the virus. This translates to between 221,000 and 442,000 adults — an estimate that is 28 to 55 times higher than the roughly 8,000 confirmed cases that the county had in early April, when the study was conducted.
Led by Neeraj Sood, a USC professor of public policy, the study took blood samples from 863 people who were randomly selected from a list obtained through a marketing firm. According to Sood, 4.1% of those people tested positive for COVID-19. The rate was adjusted to incorporate the statistical margin of error, which was assessed at a lab at Stanford University using blood samples that were positive and negative for COVID-19, according to the university.
The methodology differed slightly from the Stanford study of 3,330 people, which relied on targeted Facebook ads to find participants for its finger-prick exams, which took place on April 3 and 4. The Los Angeles study, which relied on testing at six sites on April 10 and April 11, had fewer participants, though USC had indicated that it is just the first round in a series of antibody-testing studies.
But researchers from Stanford and USC, who collaborated on the studies, found plenty of similarities in their test results. The Stanford study concluded that the number of COVID-19 cases in Santa Clara County is 50 to 80 times higher than the number of confirmed cases. The USC one also found that the number of cases is likely far higher than experts had projected.
Because so many residents with COVID-19 have not been officially tested, both studies conclude that the two counties' mortality rates based on confirmed cases are far higher than mortality rates based on the two studies' estimated numbers of infections.
The Stanford study, led by Assistant Professor Eran Bendavid, concluded that the mortality rate in Santa Clara County is between 0.12% and 0.2%. (In contrast, the county's mortality rate based solely on official cases and deaths as of last Friday, April 17, was 3.9%.)
Sood likewise said at a Monday news conference that because the number of infections in Los Angeles County cases appears to be so much higher than the number of confirmed cases, the actual mortality rate is lower.
"Maybe the good news is that the fatality rate is lower than what we thought it would be," Sood said.
He added, however, that this shouldn't be the only number that the county focuses on. The study's finding that 4% of the county's population has been infected suggests that "we are very early in the epidemic and many more people in Los Angeles County could potentially be infected."
"And as the number of infections arise, so will the number of deaths, the number of hospitalizations and the number of ICU admissions," Sood said.
Find comprehensive coverage on the Midpeninsula's response to the new coronavirus by Palo Alto Online, the Mountain View Voice and the Almanac here.
Comments
Palo Alto Hills
on Apr 21, 2020 at 5:39 pm
on Apr 21, 2020 at 5:39 pm
The technology used in these tests is what's known as a lateral flow process similar to the home pregnancy test only this one relies on blood. The problem is that the technology is still fairly unreliable and is based upon a platform that came out of China with exaggerated claims of 100% sensitivity and 99.5% specificity, when real world cohort studies done in California have shown the sensitivity to be only 59% and specificity to be 98.5%.
Another Palo Alto neighborhood
on Apr 21, 2020 at 6:07 pm
on Apr 21, 2020 at 6:07 pm
Posted by DrWaycool, a resident of Palo Alto Hills
>> when real world cohort studies done in California have shown the sensitivity to be only 59% and specificity to be 98.5%.
So, based on that, what do you think an appropriate range for the multiplier would be? Before all this, I recall seeing estimates of 9X-20X. 28X-55X, or, 50X-80X. Regardless, it looks like we are still far from a "herd immunity" level of infection history -- I don't see that the need for social distancing is diminished. What do you see?
Greenmeadow
on Apr 21, 2020 at 6:27 pm
on Apr 21, 2020 at 6:27 pm
Um, big problem here. Let's do the math.
They're saying the mortality rate of Covid19 is .1 to .2% (roughly). There are currently just over 10,000 deaths due to Covid19 in NYC (a figure that's almost certainly low, as it only counts hospital deaths where Covid tests were done). Applying even the higher .2% rate to that number, over half of the population of NYC would have had to have been exposed to Corona (10k/0.02 = 5,000,000, current population ~9M). Anyone think that's the case?
Barron Park
on Apr 21, 2020 at 6:42 pm
on Apr 21, 2020 at 6:42 pm
NYC night actually be close to 50% infection rate actually. The first serological study out of the northeast found that on about April 1, Chelsea, a Boston suburb, had around a 32% positive covid antibody rate (meaning roughly 32% of their population had already been infected and recovered). Clearly that number has increased since then and the Boston area wasn’t hit quite as bad as nyc.
College Terrace
on Apr 21, 2020 at 7:13 pm
on Apr 21, 2020 at 7:13 pm
Bruce- 10000 divided by. 02 =500000
Another Palo Alto neighborhood
on Apr 21, 2020 at 7:37 pm
on Apr 21, 2020 at 7:37 pm
Posted by BruceS, a resident of Greenmeadow
>> Applying even the higher .2% rate to that number, over half of the population of NYC would have had to have been exposed to Corona (10k/0.02 = 5,000,000, current population ~9M). Anyone think that's the case?
Could be that high, or, nearly. The rate of new infections seems to be growing much more slowly this week (in NYC). Time will tell. If enough people have gotten it/significantly exposed, the rate of new infections will actually decline soon. A lot of people in NYC died performing this "experiment". I would much prefer it if we, in Santa Clara County, can defer the experience.
Midtown
on Apr 21, 2020 at 8:29 pm
on Apr 21, 2020 at 8:29 pm
Basket of deplorables, USC and Stanford releases a report to support their good news claim as stated in the articles:
"Maybe the good news is that the fatality rate is lower
than what we thought it would be," Sood said.
The fatality rate is meaningless unless the high-school caliber research is implying that we should not be concerned about mitigation strategies since the mortality rate is "low" - are they following Dr-Phil-Dr-Oz-Dr-Pinsky way?
There is no medical research in this report, just statistics, and
drawing potentially misleading conclusions - seeking publicity.
This is as good as high-school research that looks at the CDC data on COVID-19 and concludes that Asians are not that impacted by COVID19 due to their low mortality rate. And that explains the low death count in China due to COVID19. And then add that Asians have antibodies for COVID19 to make it sound like a medical research.
Flu death rate is between 12,000 to 61,000 in 6 or 7 months, according to CDC, and COVID19 death rate so far 43,000 in 2 months. Rate by time not by population. And we don't know whether any of the deaths attributed to flu might have been actually due to COVID19.
The bottom line is that there is no "official" mortality rate for
COVID19 since widespread testing is not available. Hence, this exercise of downplaying the severity of COVID19 by comparing it with flu, ignoring the major fact that COVID19 does NOT have a vaccine, is self-serving and irresponsible.
(Wish high school students had the same peer review process for their finals, update their answers after grading, and resubmit for grading.)
Palo Verde
on Apr 21, 2020 at 9:03 pm
on Apr 21, 2020 at 9:03 pm
"Bruce- 10000 divided by. 02 =500000"
Pesky decimal points. Note that .2% is not .02 but rather .002
Another Palo Alto neighborhood
on Apr 21, 2020 at 9:28 pm
on Apr 21, 2020 at 9:28 pm
Today's NYC data (Web Link
Cases: 134,874
Hospitalized*: 35,746
Confirmed deaths: 9,562
Probable deaths: 4,865
Updated: April 21, 1 p.m.
If we assume exposure is 9X Cases, then 1.2M, divided by 8.3M, = 14%.
If we assume exposure is 20X Cases, then 2.7M / 8.3M = 32%.
If we accept the largest estimate is possible, then, 10.8M > 100%. Of course it won't be that, but NYC could be getting "saturated" at this point, where new cases should start to decline.
Note that SCC is nowhere close to any of these numbers, so, what we can do is observe strict social distancing, and reduce the exponent to less than zero -- each new case infecting less than one additional case. In order for that to happen, companies like Spectrum Communications have to get a clue. If this is true: Web Link then Spectrum just doesn't get it.
Palo Verde
on Apr 22, 2020 at 2:44 am
on Apr 22, 2020 at 2:44 am
Sciene.news attacks the Stanford study, although its slant is obvious. Nonetheless, one of its tenents is that the tests are giving false positives. Their claim is that, by suggesting that the infection rate is higher, that Americans will believe the virus isn't as fatal as it may be.
That's more conspiracy than I'd like, but it seems to be working.
In any case, I personally think the stories we're reading about exhausted healthcare workers, how the virus attacks blood cells and the immune system, and how quickly someone can go from a cough to a ventilator suggests that this virus should be taken seriously.
> false positives could account for many if not all of the 50 reported positives in their study.
So at least one question is whether or not this LA study used faulty test kits. With all the articles I've read about faulty test kits, it's almost as if you're flipping a coin.
Web Link">Web Link
Web Link">Web Link
Palo Verde
on Apr 22, 2020 at 2:58 am
on Apr 22, 2020 at 2:58 am
Sorry, forgot to mention the article criticizes that the test kits were faulty ones made in China, thus suggesting the conspiracy.
The article points to a less vociferate peer review article that discusses the false positives in the Stanford article. "this is actually a high enough rate to potentially mess up the whole study."
The peer review then discusses how the self-selected participants may have, as others have said, "enriched" the study. Certainly, we've read the articles of Bay Area residents exhibiting symptoms but were denied tests.
"After all, in the Bay Area in early April, it was really hard to get a test for people with mild symptoms or exposure. So people who thought they were exposed or symptomatic may have signed up for the study to get access to a free COVID-19 test they could get no other way. We only need ~50 out of 3330 to exhibit this behavior. And there are at least some who appear to have done just that."
The peer review then discusses the deaths in NYC from the virus and says, for the study's numbers to be correct, this would have to be the fastest transmitted virus transmitted "significantly faster than past pandemics like H1N1, many of which had multiple waves and took more than a year to run their course."
Web Link
another community
on Apr 22, 2020 at 3:03 am
on Apr 22, 2020 at 3:03 am
Neeraj Sood was a researcher on both studies - Stanford & USC. They were led by the same team.
They used the same test, a non FDA approved test that has not been validated.
The tests were done in communities that were not widely infected by covid, where testing has been scarce.
There was sample bias in the Stanford study for sure, and potentially in the USC study, but maybe not to the same extent (still volunteers that signed up to be tested for antibodies, not a totally random pool of data).
Too many issues with both tests, and I can't trust the USC test given that they had some of the same researchers conduct both tests. Too much bias there. They are mainly looking for publicity.
Palo Alto High School
on Apr 22, 2020 at 3:32 am
on Apr 22, 2020 at 3:32 am
[Post removed.]
Another Palo Alto neighborhood
on Apr 22, 2020 at 11:41 am
on Apr 22, 2020 at 11:41 am
Posted by C, a resident of Palo Verde
>> The article points to a less vociferate peer review article that discusses the false positives in the Stanford article. "this is actually a high enough rate to potentially mess up the whole study."
The Science.News article is kind of odd IMHO, but, the peer review link in your second post is pretty interesting and worth reading.
I'm puzzled by one thing though-- the general acceptance of the idea that the results of the two studies are "good news". Good news maybe, I guess, if you are under age 30, and hoping to inherent your parents' house? People in high-risk groups are looking at NYC and see bodies piling up so fast in Brooklyn they're storing them 18-wheeler trailers. 14,000+ people in NYC is a little over a month. Good news?
Crescent Park
on Apr 22, 2020 at 12:40 pm
on Apr 22, 2020 at 12:40 pm
ITS ALL VERY INTERESTING TO LOOK AT THE NUMBERS BUT AT THE END OF THIS THE LIVES LOST ARE THOSE OF LOVED ONES THE STRUGGLES ARE SHARED BY MANY.. YES WE WILL GET THROUGH THIS BUT HIND SIGHT WILL BE OUR ONLY CHANCE AT A 20 20 VISION AS TO WHAT THE NUMBERS ARE....
Old Palo Alto
on Apr 22, 2020 at 1:12 pm
on Apr 22, 2020 at 1:12 pm
[Post removed.]
Palo Verde
on Apr 22, 2020 at 1:17 pm
on Apr 22, 2020 at 1:17 pm
> Good news?
My guess is that some groups want support the "herd immunity" theory, and demonstrate that a low lethality of the virus means that you can acceptably create immunity through the natural formation of antibodies through exposure, rather than artificial immunity through vaccines. However, you don't have to look further than NYC for how unacceptable letting the virus spread is.
Showing my tinfoil hat here (: but I think many of the advice given by authorities, are more economically driven than health-driven. A further conspiratorial theory, which I don't believe because of Hanlon's Razor (who needs evil intentions when incompetence is more freely available), is that China is intentionally shipping faulty test kits so that these studies downplaying the lethality of the virus. I don't believe this because China is already known for producing products with low QC. (Even the Quon Yick Noodle Company, whose product is aimed towards Asians, now has a big red label (used to be green) saying MADE IN USA. : Web Link
Menlo Park
on Apr 22, 2020 at 1:56 pm
on Apr 22, 2020 at 1:56 pm
Seems to me that 10000/.002 = 5,000,000 :-) .02x100 = 2%, .002x100 = .2%. I'm sure it was a typo.
I think the Stanford and USC studies are close to correct. It borders on absurd to base the death rate on the number of known cases alone. It's extremely important to know the unknown cases. That's what these two studies are trying to do. Note that the vast majority of cases a minor, and many are asymptomatic. The later reason is why we must wear masks. To protect others from ourselves.
This is going to take a long time to unwind. The first cases are not known and are much earlier than we suspected. Check this out:
Web Link
Bill1940
Old Palo Alto
on Apr 22, 2020 at 4:25 pm
on Apr 22, 2020 at 4:25 pm
Why are people in the media ignoring John Ioannidis MD, the Stanford epidemiologist and medical faculty member? He makes a lot of sense in his evaluation of COVID-19 and the numbers.
Web Link
Barron Park
on Apr 22, 2020 at 4:46 pm
on Apr 22, 2020 at 4:46 pm
@ what will they do next
yep, we are under the hypnosis from Svengali Governor Newson. We must submit and obey all of his orders even those that don't make sense. Meanwhile 22,000,000 unemployed and counting. For those who have jobs and do their zoom cocktail parties all is great. I thought the Democrats were for the working class, guess not
Los Altos
on Apr 22, 2020 at 5:16 pm
on Apr 22, 2020 at 5:16 pm
@DrWaycool and others (comments, not criticisms):
"real world cohort studies done in California have shown the sensitivity to be only 59% and specificity to be 98.5%." In other words, the antibody test under-counts positives I assume? I hope that the researchers didn't "adjust the data" by inflating the number of reported positives by assuming only 60% sensitivity. That would be scientifically unacceptable. Just report the results as measured with caveats and a detailed account of the data adjustments and assumptions.
The Stanford study had unacceptable sample bias due to its reliance upon highly unreliable Facebook(!!!???) users, which would exclude most minorities, older, and medically ignorant & disbelieving people. The USC study used a "random" "study took blood samples from 863 people who were randomly selected from a list obtained through a marketing firm." But, just how that "random" sample of 863 people was collected and ramdomized is very critical to determine sample bias, which probably still is quite considerable. Also, a sample size of only 863 people is grossly insufficient to provide statistical significance from such a huge population. USC needs to conduct at least 10x more tests in the future and to ensure that its sample equally includes ALL people in the Greater LA population and not those just data minded from the Internet.
BTW, same is even more true for Stanford. Only Facebook and word of mouth??? Gimme a break!!! Get some applied math majors on board. Next time, choose your sample using mathematically accepted models and procedures. And make it 10x bigger.
Mountain View
on Apr 22, 2020 at 5:17 pm
on Apr 22, 2020 at 5:17 pm
Gee sport, don't like it in California? Why don't you move to Georgia -- the governor there is about to "open up the economy."
Of course, if you come down with COVID-19, that's your own fault.
Barron Park
on Apr 22, 2020 at 7:32 pm
on Apr 22, 2020 at 7:32 pm
@barron park
I'm willing to take my chances since i'm 98% sure to survive. Keep living in your bubble. Please come out when the world is virus free, which will be ..... never
Mountain View
on Apr 22, 2020 at 8:18 pm
on Apr 22, 2020 at 8:18 pm
Sure, keep telling that to yourself.
Come to think of it, "I'm willing to take my chances since i'm 98% sure to survive" would be just the thing to put on your tombstone after COVID-19 gets you...
Green Acres
on Apr 23, 2020 at 12:16 am
on Apr 23, 2020 at 12:16 am
[Post removed.]
another community
on Apr 23, 2020 at 10:50 am
on Apr 23, 2020 at 10:50 am
There have now been three significant COVID prevalence tests as far as I know, including the ones by Stanford in SC, USC in LA, and one in a small town in Germany, all suggesting very high infection rates Do we know of any such tests indicating the generally presumed 5x or so rates? The German test included serum antibody tests and PCR, and it seems to be rigorous at first glance, although the town had an early outbreak and high confirmed case and death rates.
There has been a pattern perceived by some as a 'lag' between initial exposure of a community to COVID-19 and the onset of hospitalizations and then deaths. Such a lag, if present, would allow a few extra initial doublings, increasing the prevalence literally exponentially.
Another Palo Alto neighborhood
on Apr 23, 2020 at 11:40 am
on Apr 23, 2020 at 11:40 am
Numbers will be analyzed and argued about for years.The only way to move forward is for Mayors to require masks, social distancing and hand sanitizer, and let businesses start to open May 1 in a phased way...doctors, dentists, elective surgery, outside businesses, beaches, parks, the first week, retail week 2, then other businesses with a limit to customers..10 for small retail etc. No one has to go out if they don’t want to.
Los Altos
on Apr 23, 2020 at 12:01 pm
on Apr 23, 2020 at 12:01 pm
I guess every study is wrong according to you guys. According to Bloomberg news, New York State conducting a similar study.
Web Link
Another Palo Alto neighborhood
on Apr 23, 2020 at 2:44 pm
on Apr 23, 2020 at 2:44 pm
How convenient for Bloomberg news, considering mike Bloomberg volunteered to lead the Tracing Army....totally unnecessary and possibly illegal according to HIPPA guidelines, and worthless with community spread already here since early Feb.
another community
on Apr 23, 2020 at 3:17 pm
on Apr 23, 2020 at 3:17 pm
Imagine we are at a very high infection rate right now. Then herd immunity is not far away. Imagine that the current infection rate is in fact so high that the infection fatality rate is the same as for a severe flu: in that case, the significant difference between COVID-19 and the flu is the doubling rate. Then mitigation is not helpful, not considering the beneficial effects of flattening the curve to avoid hospital overload. We will soon find out about herd immunity, because many herds are unable or unwilling to fully mitigate, such as the populations of Sweden, developing countries, and some US states. I wonder whether herd immunity always progresses to a virus-free state, or whether it sometimes reaches an endemic state, where there is a continous background of infection. If it is complete, then subpopulations could be allowed to reach herd immunity, say young children. This is being discussed in Germany.
Another Palo Alto neighborhood
on Apr 23, 2020 at 4:28 pm
on Apr 23, 2020 at 4:28 pm
I know everyone is getting antsy, but, the SCC program did what it was supposed to do: keep ICUs and Acute Care from overflowing, and avoid a crisis like NYC. Now, a few of you out there are saying, "See-- there's no crisis!" You *want* SCC to look like NYC?!? Look at the numbers: Web Link
So, yes, the NYC rate of new cases is dropping. Part of that may be that "enough" people have had the infection now that the infection rate is dropping. "Herd immunity". Or, it could just be social distancing and they still aren't at the herd immunity level. That is something that they should understand before they "re-open" NYC. Either way, look what they have gone through to get there. Most of us want to avoid that. (Oh, and, they aren't done yet, either, but, yes, the rate of new infections is going down.). I just can't understand why some folks are criticizing SCC for (so far) successfully avoiding an NYC-type crisis.
another community
on Apr 23, 2020 at 5:44 pm
on Apr 23, 2020 at 5:44 pm
Abon is of course, absolutely right that NYC was worse off by far than SCC, which was very successful. Many deaths and much suffering was avoided here.
I expect that the infection fatality rate will in fact not be as low as that of the seasonal flu, in part because the NYC experience was so horrific. That 20% infection rate in NYC is probably not enough to account for the tailing off, which would happen at about 60% we are told, but it probably contributed, so they are still going to have to mitigate.
In fact, we are all going to continue to have to mitigate for a long time.
Being closer to a herd immunity makes it more attractive to choose a strategy of allowing herd immunity to subpopulations, however. Some subpopulations, such as younger people without compounding factors could be allowed to reach herd immunity, and then possibly the virus would disappear from them. Schools could operate normally. Those subpopulations could freely mix with the rest, after the period of rapid infection and recovery is over.
another community
on Apr 24, 2020 at 11:47 am
on Apr 24, 2020 at 11:47 am
Here is a very controversial calculation, and I hate to be the one to do this, but here goes:
In the US, the cost so far given a shutdown very roughly is about $1T in federal aid bills. The cost to the economy so far is very roughly about a month of output. If the GDP is $20T, that is about $2T. For 30K deaths, that is $100M per death.
I am not using this number to justify or detract from any particular point of view.
another community
on Apr 24, 2020 at 6:20 pm
on Apr 24, 2020 at 6:20 pm
My father was a rather brilliant designer of automatic process control for environments like electric power plants and steel mills, back when such systems were essentially analog computers as the digital computer was barely invented. One of his favorite admonitions to researchers was "the figure of zero is not an adequate estimate for a critical assumption you find difficult to estimate". So much of the COVID debate violates this rule. Bravo to the Stanford and USC folks for focussing on the assumption that is in fact essential to rational allocation of medical and economic resources. (We do still have children dying of the flu for inadequate medical care. NOT a $3 trillion dollar problem to solve.) The way to get to a good estimate is to start somewhere, be explicit about how you got there, and keep improving the process and the estimate. Disparaging the researcher with insinuation of dispicable motivations... Under that model, the sun still goes around the earth, and those who think otherwise are condemned to hell.
another community
on Apr 25, 2020 at 7:19 am
on Apr 25, 2020 at 7:19 am
@DrDataCruncher:
Really? Bias? The study "exclude(s) most minorities, older, and medically ignorant & disbelieving people"?
What difference does it make? Isn't the point of the study to try and detect the % of a particular county's exposure? So if I'm a minority and didn't volunteer to participle, I somehow negate the fact my white, young, and educated neighbor was infected?
another community
on Apr 26, 2020 at 5:27 pm
on Apr 26, 2020 at 5:27 pm
" there have been 12 deaths of people ages 19 to 44 who had no reported underlying medical conditions as of April 7, of 3,202 deaths total."
"And 11 percent of more than 19,000 hospitalizations were for those ages 18 to 44 as of April 5. "
The above is from Web Link
Vox's source here is CDC, although it is a few weeks old.
There is also a chart from China relating age range to case fatality rate, showing a very small risk for the young The 50 to 59 range showed a 1.2% fatality rate, right about in the middle with a heavy tilt. For those at and below 59, then, the fatality rate is about 1%, but the denominator is not clear: it must be the confirmed case rate, diagnosed at hospital admission, since no general surveillance antibody testing was apparently available then. Given the recent discoveries about a possibly very large infection rate compared to the confirmed case rate, the actual infection fatality rate for Chinese at and below 59 may be very small, possibly in line with the seasonal flu. This seems not to take compounding factors into account. This is a lot of loose deduction, and I would appreciate anyone that can find better more recent data.
another community
on Apr 27, 2020 at 11:52 am
on Apr 27, 2020 at 11:52 am
If you consider the Santa Clara county Covid-19 dashboard data. It reports a positive test rate of 8.48%. Which when scaled to the population of Santa County of 1.928-Million would suggest a ~164,000 cases within the county. Data soure: Web Link
Old Palo Alto
on Apr 27, 2020 at 12:22 pm
on Apr 27, 2020 at 12:22 pm
The Santa Clara numbers are based on testing of patients and others who sought and were approved for testing. It has the same problem as the flawed Stanford study.
The percent who've been infected could be either much higher or much lower. Until there's an actual randomized sample taken and Santa Clara County, all of this speculation is meaningless. Hopefully there are some real scientists on the cases and now conducting such a study.
Barron Park
on Apr 28, 2020 at 8:56 am
on Apr 28, 2020 at 8:56 am
Brilliant real time peer (ish) review of this study by the highly educated Palo Alto people.
Learned more here than most places.
Good work PAOnline
another community
on Apr 28, 2020 at 11:24 am
on Apr 28, 2020 at 11:24 am
This seems to be the main source of data for COVID-19, a page in the CDC site:
Web Link
I do not see any info there indicating prevalence, i.e. infection rate, based on age, and the entire data set is based on 'hospital associated data', which makes sense because prevalence is only now being estimated by the (poorly randomized) serum antibody tests, such as the SCC, LA, NYC, and Heinesburg Germany tests. The latter two seem stronger.
So all the data on the site is uncorrected for age selection, meaning we do not know at all the probability of people in any particular age range to visit a hospital given infection,. Hence all of the hospitalization data seems irrelevant to me. The death rate data is probably less subject to selection, but is weeks out of date, and selected base on being in a hospital and being declared a COVID-19 death.
We need age-stratified data on actual infection fatality rate determined by truly randomized serum antibody tests. The randomization is vital, and also the false positives must be substantially less than the measured IFR so statistical noise is not a significant factor. This age-stratified measurement is difficult because it will require a larger sample set.
Without age-stratified IFR we cannot judge whether younger non-compromised individuals or other subpopulations could be reasonably allowed to go for herd immunity.
Another Palo Alto neighborhood
on Apr 28, 2020 at 11:52 am
on Apr 28, 2020 at 11:52 am
Posted by rld, a resident of another community
>> Without age-stratified IFR we cannot judge whether younger non-compromised individuals or other subpopulations could be reasonably allowed to go for herd immunity.
Note also that some compact households are 3-generational and have children and grandparents together. In such cases, the entire household still needs to shelter together and let other kids elsewhere "go for herd immunity".
another community
on Apr 28, 2020 at 12:39 pm
on Apr 28, 2020 at 12:39 pm
Three-generational households are a problem, I agree.
However, once the younger non-compromised herd reaches herd immunity, presumably the virus would disappear there. I don't know whether the process of reaching herd immunity always ends in a virus-free state or ends with an endemic state, so that is a potential problem. In order to avoid an endemic state, maybe the herd infection rate could be accelerated in some way temporarily, so it goes beyond the endemic rate. Then, once the virus is completely gone, the subpopulation could be allowed to freely mix into the rest of society. The three-generational households might actually be better protected at that point. Three-generational households would need to protect themselves temporarily, and that would need creativity.
Old Palo Alto
on Apr 28, 2020 at 6:10 pm
on Apr 28, 2020 at 6:10 pm
There's no hope of eliminating the virus through herd immunity unless we completely shut down travel. The most we can hope for is that it would become endemic. However, based on the current rates of infection I would expect a vaccine will arrive before herd immunity anyway.
another community
on Apr 28, 2020 at 8:19 pm
on Apr 28, 2020 at 8:19 pm
Good point on the travel.
But, isn't it possible for a herd to absorb a certain amount of externally injected infection if there is a larger proportion of immune individuals? If the remaining susceptible exposed proportion ('S') is small enough, and R is small enough, then the virus will die away in a few hops exponentially. The endemic state happens when SxR=1, so if SxR is less than 1, then there is a certain amount of resilience. Maybe your point is that we will in fact never get to SxR < 1. That makes sense, assuming that there would be an asymptotic approach to 1.
Also, I think we may have no choice but to cope with a spike caused by people giving up on mitigation. We have no headroom, because R is just a bit below 1 right now: supposedly it is 0.9 in NYC. I think there are many reasons for people to give up, such as restlessness, frustration, anger, economic desperation, selfishness, misunderstanding, or particular religious or political points of view. A big spike may bring us quickly nearer to herd immunity for some subpopulation.
another community
on May 3, 2020 at 12:05 pm
on May 3, 2020 at 12:05 pm
The criticism of the Santa Clara findings are speculative where one critic asserts since the study advertised to participants that it was more likely that people who experienced flu like symptoms were inclined to participate. There is no substance to that criticism as the level of concern created around Covid-19 and the prospect of immunity after having been exposed to it causes the general public to be as inclined to participate as those who experienced flu like symptoms. The second criticism relates to the tests themselves claiming they were less specific in other studies and the positive results could be a statistical error. The problem with that assertion is these tests have been used in various places across the country and in the world. If the tests were only returning an unusually high number of false positives larger sample sizes would typically yield larger proportion of positive results. If we compare the Santa Clara study to the Los Angeles County study, the Los Angeles county study has roughly ¼ the number of participants and returned a slightly higher rate of infection. The less accurate the test is, the greater chance there is for false positives the larger the sample size is. The fact that larger sample sizes have returned lower rates than smaller sample sizes suggests the test is as accurate as the researchers claim.
another community
on May 3, 2020 at 2:02 pm
on May 3, 2020 at 2:02 pm
Another interesting possibly controversial calculation:
In the US, the 300M people divided into the current number of COVID-19 deaths, 67975 is about .0002, or .02%.or one in 4413.
The average age of death is roughly 80, hence about 1/80 of the population dies each year, or 1.25%, 3.75M people.
Then the proportion of COVID-19 deaths is 1.8%
There are about 38K driving deaths, 67K opiate-related deaths, 22K suicides, and 15K homicides in the US each year.
Numbers are crude and are from wiki etc.
another community
on May 3, 2020 at 3:17 pm
on May 3, 2020 at 3:17 pm
Who will have access to the data generated by the forthcoming personal tracking App?
That data will be very valuable to many people and organizations, including some bad actors and foreign governments. Will it be deleted once the pandemic is over? How will that be ensured? Is the App opt-in? Will we be required to keep the App installed? How will we be able to ensure to ourselves that the App is secure as claimed? How can we be sure that our civil liberties will not be violated? Valuable data is hard to delete once created, especially securely, and it is very difficult to track once backed up or otherwise copied.
Mountain View
on May 3, 2020 at 3:24 pm
on May 3, 2020 at 3:24 pm
[Post removed.]
Mountain View
on May 3, 2020 at 3:56 pm
on May 3, 2020 at 3:56 pm
Your basing your viewpoint in not because of what the CDC puts out?
Given their performance during this pandemic, the CDC has no credibility left. None. Zero. Nada.
[Portion removed.]
Another Palo Alto neighborhood
on May 3, 2020 at 4:13 pm
on May 3, 2020 at 4:13 pm
@ Posted by Numbers count
Two reasons that numbers have been revised upwards from official death certificates count:
1) In many cases, when someone died without a conclusive test-based diagnosis, it wasn't attributed to COVID-19. If people died at home, weren't tested or diagnosed, they weren't counted. Web Link
2) People who died of "heart attacks" and "strokes" turned out to have Covid-19 and were suffering from hypoxia. Web Link
Given that many people have died without being diagnosed/confirmed Covid-19 is by looking at "excess" deaths. Web Link
Mountain View
on May 3, 2020 at 4:18 pm
on May 3, 2020 at 4:18 pm
Again, if you’re not comfortable going out then by all means, stay in. There are many who need, and are ready, to get back to Lori, back to life.
This is insane. The numbers do NOT measure up. I have MANY friends at both STANFORD and ECR. They are EMPTY.
Mountain View
on May 3, 2020 at 4:19 pm
on May 3, 2020 at 4:19 pm
*ready to get back to Work. Sorry for typo
another community
on May 3, 2020 at 4:26 pm
on May 3, 2020 at 4:26 pm
Not propaganda. -45 is irrelevant and I personally couldn't care less about him. I'm neutral. Numbers and facts and science count. I have no personal dog in this fight. Even the devil can speak the truth at least partially.
But,
Also note that 65K deaths is about at the top of the range for a normal seasonal flu, which goes from about 10K to 60K deaths. I forget where I got that from - CDC probably.
Here is an analogy.
The wind blows through the trees, and dead limbs come falling down. Then in between winds, there are very few dead limbs falling. Can you say the wind caused the limbs to fall? The limbs die at a certain rate regardless. Without the wind, they fall at a constant rate, but not more of them.
Perhaps after the COVID-19 is over, there will be a period during which the deaths will decrease, based on the fact that many people on the edge died, pushed over by the COVID-19 wind. That decrease may be hidden in the noise of the other deaths from whatever reason including seasonal flu that come and go.
.
Another Palo Alto neighborhood
on May 3, 2020 at 4:46 pm
on May 3, 2020 at 4:46 pm
Posted by rld, a resident of another community
>> Can you say the wind caused the limbs to fall? The limbs die at a certain rate regardless. Without the wind, they fall at a constant rate, but not more of them.
This is why in medicine people frequently cite the effect of something-or-other on life expectancy. Some risks may not have much effect on the life expectancy of a 90 year old, since many will die of something fairly soon. Some risks have large effects on life expectancy. e.g. smoking. Some don't.
>> Perhaps after the COVID-19 is over, there will be a period during which the deaths will decrease, based on the fact that many people on the edge died, pushed over by the COVID-19 wind. That decrease may be hidden in the noise of the other deaths from whatever reason including seasonal flu that come and go.
That is a reasonable hypothesis. But, it appears to be flat-out wrong. COVID-19 appears to have a major impact on life expectancy.
Web Link
Because a lot of people have been cut down by it while still in their highest-productivity working years, it appears that Covid-19 has the potential of having a major effect on both longevity and the economy. That is why it makes a lot of sense to limit activities/businesses which contribute little to long-term productivity, and, have significant virus transmission rates.
another community
on May 3, 2020 at 5:39 pm
on May 3, 2020 at 5:39 pm
I believe this is the source article:
Web Link
This is very misleading.
The paper refers to years of life lost YLL, the number of years of life lost for a patient before and after diagnosis with COVID-19.
The roughly one decade YLL number does not apply to people without COVID-19. It is the change in life expectance based on a diagnosis. Every possible additional serious diagnosis that a person could have added to their current state of health, especially if co-morbid will produce a YLL. We do not have such numbers for other serious diagnoses added to previous morbidity for comparison.
The number of YLL for the entire population is a different number. It will be much less than the YLL for a person with co-morbidity who receives a diagnosis of COVID-19, or even someone without co-morbidity. We know that the risk to a co-morbid patient given a new COVID-19 diagnosis, i.e. a "confirmed case" is in the range of one to ten percent or even more as outside numbers.
So this is not relevant to anything I have mentioned before. The population in general is not losing a decade. You have to work with the infection fatality rate, for example, not the case fatality rate.
another community
on May 3, 2020 at 6:47 pm
on May 3, 2020 at 6:47 pm
Anon:
"Because a lot of people have been cut down by it while still in their highest-productivity working years, it appears that Covid-19 has the potential of having a major effect on both longevity and the economy."
The case fatality rate is extremely dependent on age. I remember a report from Italy that said the average age of death was 80 years, and that did not take comorbidity into account. There was also data in a paper I referenced above about that heavy age dependence in China. We have no data on the age-stratified infection fatality rate, even though the SCC (flawed?) and LA and NYC and German serum antibody tests showed a very high degree of infection overall, reaching orders of magnitude larger than the case rate. The lack of data on the infection prevalence or the infection fatality rate by age leaves us no way to determine the actual YLL for various ages, hence the YLL for the general population. Getting that age-based data will require bigger sample sets.
The overall economy will be affected by the loss of about 65K more people than usual, not accounting for age bias of the work force or deaths. That is a very tiny fraction of the work force. It's in the noise.
Do you have more info on that major effect you mention?
Another Palo Alto neighborhood
on May 3, 2020 at 7:06 pm
on May 3, 2020 at 7:06 pm
Posted by rld, a resident of another community
>>I believe this is the source article: Web Link
Yes.
>> This is very misleading.
I don't think so at all. We have frequently heard comments along the lines that these people all had underlying conditions or were very old and therefore would have died soon anyway. This paper directly refutes that.
>> The paper refers to years of life lost YLL, the number of years of life lost for a patient before and after diagnosis with COVID-19.
Correct; it does. That is the whole point.
>> The roughly one decade YLL number does not apply to people without COVID-19.
You are asking a different question, one that the paper doesn't answer. We don't yet know how the pandemic will affect life expectancy overall, because, among other things, it depends on characteristics of the population, and, it depends on our response.
Thus it is with any pandemic response. If you succeed, as New Zealand is getting close to, then, almost nothing happens, and you ask what the big deal is. OTOH, you look at NYC, and you can ask why they didn't act more forcefully sooner.
another community
on May 3, 2020 at 10:14 pm
on May 3, 2020 at 10:14 pm
Anon, we are not far from agreement. I agree that the paper does not answer the question of the general YLL for the population, which depends on the infection prevalence and infection fatality rate based on age. I agree that the paper is about cases, not infections.
However, then you do seem to be saying that the general population's YLL will be significant, and the life expectancy for the general population and the economy and so on will be significantly affected, but again these depend not on case-related stats, but on infection-related stats, which are only starting to come out, and only in the aggregate.
I think in the end there will be very little affect on the stats for the general population.
Of course with respect to the economy, the lockdown is having a very dramatic effect, which must be compared with the effects of COVID-19 on the workforce later, if significant. How do we amortize this enormous expense?
Yes, one could say 'look at what would have happened without the mitigation?'. "We would have another NYC". But we don't know. The models were so wrong that I think it's a stretch to give them much credibility in determining the hypothetical worst-case. They were off by two orders of magnitude. Even the best cases under hypothetical mitigation according to the models were far off. It is easy to assume we dodged a 2M death situation with mitigation.
I don't think anyone can reasonably say that the effects of mitigation in any particular case were significant or not. Each case is so different: in the NYC case you have people living in tall buildings sharing elevators and recirculating building air and so on, so staying locked down may be more effective, and there was much international travel, but rural areas have different exposures. We are not a herd, but more like a herd of herds. So many random hotspots.
Still, it does seem overall that mitigation helped a lot, just intuitively. But now where are we? Do we continue this delicate balance between mitigation and openness for a year until the vaccine? Will there be outbreaks anyway? If there were a subpopulation - such as younger non-compromised people, - who actually did safely reach and exceed herd immunity for some reason, we might be done with this. Keep looking at Sweden, developing countries, and some US states for herd immunity experience. Sweden says they are in it for the long run, and herd immunity may be inescapable generally. Anders Tegnel thinks so. Let's see.
Web Link