Coronavirus

[Norbert]

Decent reporting this morning on wide criticism of the Stanford study in the San Jose local paper.
Stanford coronavirus study triggers feud over methodology and motives

"Statistician John Cherian of D. E. Shaw Research, a computational biochemistry company, made his own calculations given the test’s sensitivity and specificity — and estimated the proportion of truly positive people in the Stanford study to range from 0.5% to 2.8%.

Adjusting for demographics, Cherian estimated a county infection rate of close to 1%. This would lead to a substantially higher mortality rate, rising above 1%."

Even with the same self-selected sample !

EDIT: As far as I can tell this number would bring the factor down to 20x, which is within the range we have been looking at here. (Keep in mind this factor reflects the local testing level at a specific point in time, more than anything else.)

 

[Daniel in SD]

Only in America would we expend so much energy examining something in the theoretical that had and is already happening all around us.

10k random tests in NYC would tell us everything we need to know about mortality. 3M tests around the country would have let us begin opening back up in a tiered fashion today.

I actually agree with you about something! haha.
It is finally happening: Gov. Cuomo announces plan to start 'most aggressive' antibody testing of New Yorkers to see just how widely the coronavirus has spread

Should be interesting to see the updated article and new Appendix addressing comments/criticisms they've received:

What I don't understand is why they didn't run it by a statistician before releasing it to the public? I saw a major flaw (that many others have pointed out now) skimming the paper for 5 minutes. And I think I've only taken one or two statistics courses in my life.

 

[MXLRplus]

The SBA paycheck loan status:

The $342 billion went to 1,661,000 businesses averaging $205,900.
4,412 checks were cut for over $5 million. Some recipients had up to 10,000 employees and also had existing credit lines so they were just looking for the <1% APR.

 

[TEG]

Different takes on virus mutation news:

First 'significant' coronavirus mutation discovered in preliminary study

'Significant' coronavirus mutation discovered, could make vaccine search 'futile,' study says

Scientists: Virus mutating, but won't affect drug, vaccine work - Chinadaily.com.cn

The new coronavirus may be mutating to a less deadly form

 

[TEG]

UV-C light to kill airborne COVID-19:

How UV light may protect us from the coronavirus

Ultraviolet LEDs prove effective in eliminating coronavirus from surfaces and, potentially, air and water

Full Page Reload

UC Santa Barbara Researchers Tweak UV LEDs To Neutralize COVID-19 Virus | CleanTechnica

An ultraviolet light may help fight spread of novel coronavirus

 

[dondy]

That’s a change that’s only a couple of days old. The USA may have changed their counts, but I doubt our neighbouring countries have changed their stats. Note that the CDC means nothing in Europe, we follow WHO guidelines.
This article explains the difference between our stats and the neighbouring countries: België vestigt een triest record covid-19-doden :
”
Maar niet alle landen tellen hun doden op dezelfde manier. België telt de doden in de ziekenhuizen, de rusthuizen en daarbuiten. Ook vermoedelijke gevallen worden meegeteld, zonder dat er een formele covid-19-test was.

Italië telt alleen de doden die eerst al een positieve covid-19-test afgelegd hadden, maar in rusthuizen werd lang niet systematisch getest. In Duitsland telt het Robert Koch-instituut enkel diegenen bij wie een test bevestigde dat ze covid-19 hadden. Ook in Nederland zijn er volgens de Vereniging van Specialisten voor Ouderengeneeskunde (Verenso) meer mensen aan covid-19 overleden dan in de officiële statistieken zichtbaar is. In Nederland is 35 procent van de overledenen 85-plusser, in België is dat 45 procent.”
It says that Italy and Germany only count positively tested deaths, and the same is assumed for The Netherlands based on the difference in mortality of 85+yo deaths. (A couple of days ago our national TV news stated as a fact that the Netherlands only count confirmed cases and that the death rate is basically identical in Belgium versus The Netherlands).
The article also states that the ‘oversterfte’ (‘over-deaths’, i.e. the additional number of deaths compared to the same period a year ago) in the last couple of weeks of march is the same in Belgium as in The Netherlands, but the number of reported covid deaths differs by 40%.

EDIT: This Zo komen de verborgen coronadoden aan het licht Dutch TV news link confirms that only confirmed cases are counted in the Dutch stats, and that the real mortality rate is suspected to be twice as high as the official stats. The article links to the youtube clip of their news, and that contains an astonishing graph of the weekly deaths.

bad journalism is bad. These jerks were reporting "lack of ventilators" in the Netherlands as well which led to the cases of stubborn old people refusing to go to the hospital "because it is useless".

The RIVD (dutch CDC) reports aggregated data provided by regional GGDs. GGD aggregate data coming from death certificates with covid-19 as cause of death. Cause of death is reported by respective doctors who signes death certificates. It is choice of these doctors to report or not suspected case as "covid". In most of the nursing associations and huisarts associations it was decided by the end of march to report.
In the specific case presented in this news all deaths were reported as covid. i.e. ALL. Everybody during that short period.

The Dutch have around 1000-2000 unreported deaths in the second half of march. It's not clear how many because GGDs were expecting golf of the influenza deaths a-la 2018 spring.

 

[MXLRplus]

This might really get people wake up to the C19 threat. The US is about to have a bacon shortage as some major pork packing facilities are shutting down due to sick employees.

Noooooo! Not the BACON!!!

Sidebar - oil futures dipped into the negative numbers. Wow.

 

[Oil4AsphaultOnly]

The SBA paycheck loan status:

The $342 billion went to 1,661,000 businesses averaging $205,900.
4,412 checks were cut for over $5 million. Some recipients had up to 10,000 employees and also had existing credit lines so they were just looking for the <1% APR.

10,000 employees?!?! How's that a "small business"?! At least some of them have returned their loan money:

Shake Shack Returns $10 Million PPP Loan Amid Criticism Of Restaurant Chains Receiving Stimulus Funds

 

[MXLRplus]

10,000 employees?!?! How's that a "small business"?! At least some of them have returned their loan money:

Shake Shack Returns $10 Million PPP Loan Amid Criticism Of Restaurant Chains Receiving Stimulus Funds

There was chain restaurant exemption. Apparently some Senator was a major shareholder?

 

[Oil4AsphaultOnly]

There was chain restaurant exemption. Apparently some Senator was a major shareholder?

Maybe we should carve some bacon off of that "pork"?! Kills two birds with one stone!

 

[dfwatt]

I actually agree with you about something! haha.
It is finally happening: Gov. Cuomo announces plan to start 'most aggressive' antibody testing of New Yorkers to see just how widely the coronavirus has spread

What I don't understand is why they didn't run it by a statistician before releasing it to the public? I saw a major flaw (that many others have pointed out now) skimming the paper for 5 minutes. And I think I've only taken one or two statistics courses in my life.

Actually there's an even more damning assessment that one can make about this Stanford antibody study. When a study's senior author is associated with multiple videos dismissing covid-19 as an overblown and relatively minor Health threat, you would think that the study authors would have gone over the study with a fine-tooth comb and dotted every i and crossed every T to make sure that there was no appearance of bias. This did not happen. This says something about the quality of the science . . and the professional caliber and integrity of the scientists involved. When you have taken a loudly visible and politically divisive public position around an undecided scientific issue, and then released data sets that show poor methods and / or cooked statistics, as far as I'm concerned your scientific credibility is shot to s***.

You've done the equivalent of pouring gasoline on your scientific reputation and lit a match.

 

[SVMike]

Even if that isn't a common transmission path, studies have also shown the virus can live on metal and plastic surfaces up to 72 hours and on cardboard for 24.

True, but we should understand what it means to survive for 72 hours.
This is in a lab, not outdoors (for example). It doesn't mean that the virus is just as infectious at hour 71 as a hour 0. It means that numerous half lives killed most of the virus until there was just a detectable amount to reproduce.
Curiously, different metals have different properties. Stainless steel ...up to 72 hours, while copper is just a few hours.
Also, I think all these studies were on other corona virus types, not CV-19.

 

[SMAlset]

More symptoms like blue lips in addition to rashes and frostbitten toes being linked to CV19. Reports are being taken seriously by medical staff.

“The CDC lists “bluish lips or face” as an emergency warning sign for COVID-19 that requires immediate medical attention.“

https://bgr.com/2020/04/20/coronavirus-symptoms-frostbite-lesions-bluish-lips-and-face-hives/#

 

[TEG]

...Also, I think all these studies were on other corona virus types, not CV-19.

No, they have specifically been testing SARS-COV-2 (the virus that can cause COVID-19)
New coronavirus stable for hours on surfaces
Can COVID-19 Survive on Surfaces?
How long can the virus that causes COVID-19 live on surfaces?

The 72 hours was for "on plastic".
I also wonder how long it can be viable in the fridge?

Of all the things I am dealing with during shelter-at-home, the thing that concerns me most are the plastic containers I get from the market and put right into the fridge/freezer at home.

How long can coronaviruses survive in a freezer? Up to two years, warns expert

Also, fresh vegtables in plastic bags tend to get consumed fairly rapidly (while they are fresh) so don't get to spend 3 days in the normal quarantine routine I do for dry and canned goods.
Sort of a tradeoff... Maybe safer to eat non-fresh things, but I hear that it is extra important to eat well now to keep your immune system in top shape.

 

[jerry33]

True, but we should understand what it means to survive for 72 hours.
This is in a lab, not outdoors (for example). It doesn't mean that the virus is just as infectious at hour 71 as a hour 0. It means that numerous half lives killed most of the virus until there was just a detectable amount to reproduce.
Curiously, different metals have different properties. Stainless steel ...up to 72 hours, while copper is just a few hours.
Also, I think all these studies were on other corona virus types, not CV-19.

Copper is famous for killing virus and bacteria. One reason I have a copper sink in the kitchen. Four hours pretty well kills them all.

 

[EinSV]

What I don't understand is why they didn't run it by a statistician before releasing it to the public? I saw a major flaw (that many others have pointed out now) skimming the paper for 5 minutes. And I think I've only taken one or two statistics courses in my life.

Here's a good question to ask: why do we apply a higher standard to new studies that challenge our views than to previous studies that formed our views in the first place?

Quality assessment and confirmation bias
The quality of any experimental findings must be appraised. Was the experiment well performed and are the outcomes reliable enough for acceptance? This scrutiny, however, may cause a confirmation bias: researchers may evaluate evidence that supports their prior belief differently from that apparently challenging these convictions. Despite the best intentions, everyday experience and social science research indicates that higher standards may be expected of evidence contradicting initial expectations. Effect of interpretive bias on research evidence
​

Many on this board have been dismissive of antibody testing done to date because the tests aren't randomized, were not adequately powered, etc.

Have any existing studies using RT-PCR provided statistically valid estimates of the percent infected of any large population, with adequate and statistically valid analyses of false negatives and the short time period RT-PCR can detect the presence of virus? I don't think so, but if I've missed something, I'm sure someone will let me know. If not, what's the justification for the double standard?

 

[AlanSubie4Life]

Have any existing studies using RT-PCR provided statistically valid estimates of the percent infected of any large population, with adequate and statistically valid analyses of false negatives and the short time period RT-PCR can detect the presence of virus? I don't think so, but if I've missed something, I'm sure someone will let me know. If not, what's the justification for the double standard?

I believe there are tons and tons of false negatives on RT-PCR (they are highly specific, but sensitivity is an issue), but since even if every one of the 4 million tests conducted in the US were positive, I believe it would still be undercounting the number of US cases, based on epidemiologists' best estimates based on infection dynamics, I don't really see the sensitivity as an issue right now.

Another point is that the maximum positivity we've seen is 45% positive or so in NJ, so in the worst case of everyone being positive in that tested population, we know that RT-PCR testing is off by no more than about a factor of 2. It only has a 55% false negative rate at worst (assuming all the PCR tests are pretty similar and done at about the same point in an infection in general), so that seems pretty good.

Do I know what the prevalence is, exactly? No. Just seems like people who do this for a living think it's about 10-20x what we're measuring, so that seems like a good starting point.

Empirically, mostly I rely on circumstantial evidence from South Korea to inform me of what the likely IFR is (they don't have a massive outbreak, so must have found most of the cases), look at US deaths (recognizing they are somewhat undercounted), back-propagate to the number of infections a couple weeks ago, forward propagate to likely cases today, based on published virus properties and suppression measures in place, and use that as my estimate of infections. And then compare to known cases. That seems to generally align pretty well with epidemiological estimates of 10x-20x, so that is good enough for me. For now.

I look forward to antibody test results in large samples from populations with relatively high prevalence. We definitely need them.

 

[EinSV]

LA County of Public Health and USC have issued a press release on the first round of antibody studies conducted by USC in Los Angeles.

Here is a summary of results:

"Based on results of the first round of testing, the research team estimates that approximately 4.1% of the county's adult population has antibody to the virus. Adjusting this estimate for statistical margin of error implies about 2.8% to 5.6% of the county's adult population has antibody to the virus- which translates to approximately 221,000 to 442,000 adults in the county who have had the infection. That estimate is 28 to 55 times higher than the 7,994 confirmed cases of COVID-19 reported to the county by the time of the study in early April. The number of COVID-related deaths in the county has now surpassed 600.​

"We haven't known the true extent of COVID-19 infections in our community because we have only tested people with symptoms, and the availability of tests has been limited," said lead investigator Neeraj Sood, a USC professor of public policy at USC Price School for Public Policy and senior fellow at USC Schaeffer Center for Health Policy and Economics. "The estimates also suggest that we might have to recalibrate disease prediction models and rethink public health strategies." Press Release:USC-LA County Study: Early Results of Antibody Testing Suggest Number of COVID-19 Infections Far Exceeds Number of Confirmed Cases in Los Angeles County
​

More info in the release for anyone interested (I haven't reviewed in detail yet).

 

[GFernandes]

LISTING OF DEPARTMENT OF PUBLIC HEALTH PRESS RELEASES
LA Antibody

https://t.co/ku0Z8S64DO?amp=1
Ohio prison antibody

 

[Daniel in SD]

Here's a good question to ask: why do we apply a higher standard to new studies that challenge our views than to previous studies that formed our views in the first place?

Quality assessment and confirmation bias
The quality of any experimental findings must be appraised. Was the experiment well performed and are the outcomes reliable enough for acceptance? This scrutiny, however, may cause a confirmation bias: researchers may evaluate evidence that supports their prior belief differently from that apparently challenging these convictions. Despite the best intentions, everyday experience and social science research indicates that higher standards may be expected of evidence contradicting initial expectations. Effect of interpretive bias on research evidence
​

Many on this board have been dismissive of antibody testing done to date because the tests aren't randomized, were not adequately powered, etc.

Have any existing studies using RT-PCR provided statistically valid estimates of the percent infected of any large population, with adequate and statistically valid analyses of false negatives and the short time period RT-PCR can detect the presence of virus? I don't think so, but if I've missed something, I'm sure someone will let me know. If not, what's the justification for the double standard?

I think it's funny that you think I'm capable of quickly finding mathematical errors in other papers and the only thing holding me back is confirmation bias. I'm not that smart, the errors have to egregious. Is your argument that we should apply lower standards to all research?
I have already admitted that I will continue to view anything you post with a high degree of skepticism. I guess I should put a disclaimer whenever I respond to your posts so that everyone knows my bias.