Coronavirus

[AlanSubie4Life]

I don't see that happening. As load on the testing system improves - more people with symptoms will go for testing and keep the confirmed case numbers flat.

I guess I should have said "infections" - I don't expect the actual case numbers to change much, either, over the next week or two, though I hope for a gradual downward trend while testing trends up.

BTW, is there a good site that shows daily confirmed case charts for all states ?

You could try this:

Tableau Public

The "My State" tab allows you to look at different states. Every now and again there's bad data in there but I think that's due to the source.

Honestly, all of this state data is basically crap since it doesn't account for date of symptom onset, date of test, date of death, etc. It totally changes all the curves. Would be nice to have epidemiological curves for all the states in one place. (Look at the WA State data with all the missing data and compare to the actual curve on the WA State website.) This can have a big impact on actual Rt. So I'd ignore the Rts on this site.

Santa Clara have re-pre-printed their study, and now they're at a minimum of 15x the infections vs. cases.

They still, unbelievably, seem to be adjusting for the test sensitivity/specificity after doing the population weighting which will mean that the uncertainty in the test performance will be underestimated.

COVID-19 Antibody Seroprevalence in Santa Clara County, California

But anyway, seems to be borderline converging on the generally accepted 10x number (that most people recognized was the scaling a month ago!). Not sure what value the study really has at this point.

 

[FoverM]

I've been using Worldometer. Seems to agree with MA website.

Do you manually add the data to your database or does worldometers.info they have a data file available?

 

[ReddykwRun]

Pardon me, should have said "healthy 49 and under crowd", though I did also say anyone at risk with underlying conditions should still isolate.

Plenty of 30-50 year olds are very much at risk of hospitalization or death. Diabetics, heart conditions....

The point is, the hospitals are in fact half empty.

In a statement, the Navy said Aviation Ordnanceman Chief Petty Officer Charles Robert Thacker Jr., 41, of Fort Smith, Arkansas, died on April 13 at the U.S. Naval Hospital in Guam of COVID-19, the illness caused by the novel coronavirus. He died 11 days after his captain was fired for pressing the Navy for greater action to safeguard his crew from the virus.

 

[TEG]

I dread 420 tweet and reaction to it visiting us all over again this month … I can see a complete meltdown if CA pushes out stay-at-home order.
BTW, the Republican governor of Arizona did just that. A small relaxation to the existing rules but most rules stay for another 2 weeks. Inslee in WA made some small changes (like allowing some construction work to resume). I can see CA do something similar ….

Yeah, construction & landscaping starting back up...
6 Bay Area counties relax some shelter-in-place restrictions; here are changes starting May 4
Gardening and nursey supply firms. Some outdoor summer camps with small groups too.

On the other hand, the state beaches and parks are being closed down in response to recent swarming during a heat wave.
Memo Says California Governor Will Order All Beaches Closed

 

[TEG]

Our governor sounded like a disappointed parent when talking about the "occupy beaches" movement.

" You kids were bad - back to your rooms - you are grounded! "

#Covid19FlashMob

 

[AlanSubie4Life]

On the other hand, the state beaches and parks are being closed down in response to recent swarming during a heat wave.
Memo Says California Governor Will Order All Beaches Closed

Incorrect, OC only, all beaches state & local. He seemed to clearly indicate that he wanted to work out some measures with OC so that they could be safely reopened. Sort of like what San Diego has been doing (no sitting down or lying down, or stopping, allowed)

 

[AlanSubie4Life]

Aviation Ordnanceman Chief Petty Officer Charles Robert Thacker Jr., 41, of Fort Smith, Arkansas, died on April 13 at the U.S. Naval Hospital in Guam of COVID-19, the illness caused by the novel coronavirus

He was probably very frail and weak. Just a fluke!

I'm a fit 43 year old, so I have nothing to worry about, or at least so says @TheTalkingMule. Of course, I'd have to be pretty dumb to listen to a Talking Mule.

 

[renim]

Not sure it means much.

Taiwan took measures very early - so they didn't have to lock down at all. Italy was so late to respond - a thousand dead before lockdown was imposed, so they had to be quite strict (compared to Wuhan lock down after 30 deaths).
....

View attachment 537598

It is difficult to make useful comparisons useing differing countries. But NZ and Australia are quite similar, actually considering the bulk of Australia used to reside is the region bounded by Syd, Melb, ACT. NZ is closer in style to that Australia than much of the crocodile dundee rest of Australia.

Australia/ NZ took comparable timing and quarantine, and it seems that NZ lvl 4 stay at home was about twice as damaging to employment and GDP as Australia's lvl 3.x supression, for no additional health benefit. (Link currently lost but it was from an ABC report)

Point is, re covid 19, for a given employment loss, there is a set of government actions with very high health benefit.

Another set of government actions have negligible ( or even negative) health benefit.

An average government, can overshoot/undershoot on maximising health benefits while minimising employment/mental damage.

A better than average can maximise health benfits without causing excessive enployment/mental damage.

The corollary is that there is also a choice of restictions that are ineffective at health benefits, yet still cause employment/mental damage....

 

[AlanSubie4Life]

Mobility index site to compare a few interesting cities:

Sydney still looks pretty immobile.


Citymapper Mobility Index

 

[EinSV]

The Stanford scientists issued an updated version of their Santa Clara paper. They added data that allowed them to tighten up the specificity error bars, and also added a section responding to the extensive input they received on the earlier draft.

"Results
The raw prevalence of antibodies to SARS-CoV-2 in our sample was 1.5% (exact binomial 95CI 1.1-2.0%). Test performance specificity in our data was 99.5% (95CI 99.2-99.7%) and sensitivity was 82.8% (95CI 76.0-88.4%). The unweighted prevalence adjusted for test performance characteristics was 1.2% (95CI 0.7-1.8%). After weighting for population demographics of Santa Clara County, the prevalence was 2.8% (95CI 1.3-4.7%),using bootstrap to estimate confidence bounds. These prevalence point estimates imply that 54,000 (95CI 25,000 to 91,000 using weighted prevalence; 23,000 with 95CI 14,000-35,000 using unweighted prevalence) people were infected in Santa Clara County by early April, many more than the approximately 1,000 confirmed cases at the time of the survey."

They calculate a 0.17% IFR in Santa Clara County:

"We can use our prevalence estimates to approximate the infection fatality rate from COVID-19 in Santa Clara County. Through April 22, 2020, 94 people died fromCOVID-19 in the County. If our estimates of 54,000infections represent the cumulative total on April 1, and we assume a 3 week lag from time of infection to death, up to April 2224, then 94 deaths out of 54,000 infections correspond to an infection fatality rate of 0.17% in Santa Clara County. If antibodies take longer than 3 days to appear, or if the average duration from case identification to death is less than 3 weeks, then the prevalence rate at the time of the survey was higher and the infection fatality rate would be lower. On the other hand, if deaths from COVID-19 are under reported or the health system is overwhelmed than the fatality rate estimates would increase. Our prevalence and fatality rate estimates can be used to update existing models, given the large upwards revision of under-ascertainment." https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v2.full.pdf

Of course the most obvious issue with their estimate is that antibodies are unlikely to detectable in only 3 days per their very conservative assumption, so their IFR estimate may be too high. For some odd reason nobody mentions this issue.

Let the next round of flame wars begin.

 

[Remus]

The Stanford scientists issued an updated version of their Santa Clara paper. They added some additional data that allowed it to tighten up the specificity error bars, and also added a section responding to the extensive input it received on the earlier draft.

"Results
The raw prevalence of antibodies to SARS-CoV-2 in our sample was 1.5% (exact binomial 95CI 1.1-2.0%). Test performance specificity in our data was 99.5% (95CI 99.2-99.7%) and sensitivity was 82.8% (95CI 76.0-88.4%). The unweighted prevalence adjusted for test performance characteristics was 1.2% (95CI 0.7-1.8%). After weighting for population demographics of Santa Clara County, the prevalence was 2.8% (95CI 1.3-4.7%),using bootstrap to estimate confidence bounds. These prevalence point estimates imply that 54,000 (95CI 25,000 to 91,000 using weighted prevalence; 23,000 with 95CI 14,000-35,000 using unweighted prevalence) people were infected in Santa Clara County by early April, many more than the approximately 1,000 confirmed cases at the time of the survey."

They calculate a 0.17% IFR in Santa Clara County:

"We can use our prevalence estimates to approximate the infection fatality rate from COVID-19 in Santa Clara County. Through April 22, 2020, 94 people died fromCOVID-19 in the County. If our estimates of 54,000infections represent the cumulative total on April 1, and we assume a 3 week lag from time of infection to death, up to April 2224, then 94 deaths out of 54,000 infections correspond to an infection fatality rate of 0.17% in Santa Clara County. If antibodies take longer than 3 days to appear, or if the average duration from case identification to death is less than 3 weeks, then the prevalence rate at the time of the survey was higher and the infection fatality rate would be lower. On the other hand, if deaths from COVID-19 are under reported or the health system is overwhelmed than the fatality rate estimates would increase. Our prevalence and fatality rate estimates can be used to update existing models, given the large upwards revision of under-ascertainment." https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v2.full.pdf

Of course the most obvious issue with their estimate is that antibodies are unlikely to detectable in only 3 days per their very conservative assumption, so their IFR estimate may be too high. For some odd reason nobody mentions this issue.

Let the next round of flame wars begin.

they did not address the self selection bias.

 

[eevee-fan]

Part of the reason our crisis is lasting longer than China's is our lockdown is relatively lax compared to China's because we are trying not to be fascist. If the Govt in mid March sent the National Guard to enforce a hard 2 week quarantine where no one leaves their house even for supplies, rather the army does drop offs, this whole thing would've been over but we didn't do that because personal rights and we aren't fascist.

I was expecting this while others were out panic shopping. Did not happen. Now we know what happens when we do not act logically. Just more pain.

 

[EinSV]

they did not address the self selection bias.

To my knowledge, there are no studies of covid-19 infection rates or IFRs of any kind that avoid selection bias.

The authors discuss selection bias at length in the response to comments.

 

[EVNow]

Our governor sounded like a disappointed parent when talking about the "occupy beaches" movement.

" You kids were bad - back to your rooms - you are grounded! "

#Covid19FlashMob

If people keep their distance - frankly beach is probably a low risk place. Winds, humidity should work to reduce risk.

I think outdoors in general are less risky than indoors. Gyms, bars, restaurants and offices are probably the worst.

 

[GoEV]

Moody’s is awaiting the last criterion they set that Tesla demonstrate they can reliably produce 250k Model 3’s per year



Elon just personally attacked some scuba guy? Really?

Elon was spending his own personal wealth, his company’s resources in a race against time to help in any way possible the boys trapped in the cave, when Unsworth with zero provocation told him he could stick the sub up his a**. Unswortth behaved beyond despicably and it is quite hilarious that you characterize him as just some scuba guy who Elon for apparently no reason decided to attack.

Really?

* * *
I GREATLY admired Musk's efforts to help the kids, but why would a great leader get into an "8th grade name calling" rant with some largely unknown guy who had offended him? Did Musk really need to do that? What did Musk gain?

 

[AlanSubie4Life]

Of course the most obvious issue with their estimate is that antibodies are unlikely to detectable in only 3 days per their very conservative assumption, so their IFR estimate may be too high. For some odd reason nobody mentions this issue.

At this point, they have kind of lost their credibility and their thunder, as in the interim so many other seroprevalence studies have pointed at closer to the expected 10x-20x ratio with IFR slightly below 1%, with much tighter confidence intervals (as a %).

There are still the sampling/selection bias issues and how they handle the adjustments for sampling. Obviously there is probably some significant undercount in February of deaths, and in addition 3 weeks is not long enough to capture the nasty long death tail, unfortunately.

In any case they seem pretty close to the expected 10x number, as I would have expected, and looks like they are going to end up with an IFR of about 0.7-1% as expected, once they get all their issues sorted (which may be impossible).

Summary: in the end, it turns out it is hard to measure incidence of 1% with a test with between 0.5% and 1.5% false positives. It was a tough study to have any success with - I don’t think anyone could have been successful.

On the plus side, it does confirm (with reasonable confidence) the one thing it could confirm:

Nearly no one has had the virus in Santa Clara County

 

[TheTalkingMule]

In a statement, the Navy said Aviation Ordnanceman Chief Petty Officer Charles Robert Thacker Jr., 41, of Fort Smith, Arkansas, died on April 13 at the U.S. Naval Hospital in Guam of COVID-19, the illness caused by the novel coronavirus. He died 11 days after his captain was fired for pressing the Navy for greater action to safeguard his crew from the virus.

Yes, it's bad policy to leave hundreds of known coronavirus patients to mix with the general population on a boat. Everyone of all ages should practice safe social distancing to the best of their ability, test aggressively, trace contacts with those found positive, and isolate.

 

[AlanSubie4Life]

To my knowledge, there are no studies of covid-19 infection rates or IFRs of any kind that avoid selection bias.

Yeah it is tough. Denmark had a blood donor test done recently which worked out to an IFR of 0.3 to 0.8, but that doesn’t account for inadvertent selection bias due to socioeconomic factors which affect exposure (blood donors are probably more likely to be exposed, but hard to know for sure).

Carl T. Bergstrom on Twitter

 

[eevee-fan]

... when your Intel comes from talking heads at FAUXNEWS...

 

[eevee-fan]

Oh great, New York-New York hotel to be in the Control Group. I assume bookings open only to residents of New York?