Coronavirus Pandemic (COVID-19) (SARS-CoV-2) [2020]

[dobwal]

Iroboto was talking about more cases presently out there than reported. If so, it means 1) there are (far) more people infected than currently mentioned. 2) Their symptoms are so mild as to not make anything of it. 3) The current death rate calculated from known cases would actually be lower if we included the cases currently unrecorded. Theoretically, let's argue that half the population infected by Covid19 have a very mild reaction such that they aren't even aware it's Covid19; then the death rate reduces to 1% instead of 2%.

Given 'healthy' people aren't screened for infections, we can't tell real danger of infection. For the really deadly diseases that are tracked like SARS, all the known cases are tracked because it's pretty obvious whether you have it or not.

How can ML help with creating vaccines?

Hard to know because in China they aren’t just counting those that self report to the hospital. They are screening the public and testing anyone who is registering above normal body temperatures.

Some countries are screening people entering and hospitalizing anyone that meet certain criteria as a quarantine measure.

SARS and MERS are both caused by Coronaviruses. Covid-19 is caused by the SARS-CoV-2 virus.

 

[Shifty Geezer]

The particular examples were Korea and Italy.

Less than a 1000 people have it in South Korea. A country with over 50 million people. Not even 300 have it in Italy, a country with over 60 million people

those are only confirmed numbers. Meaning they were sick enough in a way that their health care provider decided to issue the COVID test.

there could be many carriers with mild symptoms walking about thinking nothing is wrong.

Taking the idea of there being many carriers with mild symptoms, the 300 confirmed cases, and all those to follow, and the death rate therefrom, would exclude all the mild-symptom people.

Your flu data points to this - estimates are between 10 and 20%. That's a huge range. Reported cases for Covid19 may hit 10% of the world population, say, and 2% of those cases may die, but then untracked, another 10% of the world population also had Covid19 without ever realising, such that the 2% death rate among reported cases is 1% of total cases.

 

[iroboto]

How can ML help with creating vaccines?

It cant help with vaccine creation. It can help with identifying people who could have COVID before having to do the test. This way we can get ML separating out a population of high probable COVID infections that should receive a test from those that shouldn’t.

Taking the idea of there being many carriers with mild symptoms, the 300 confirmed cases, and all those to follow, and the death rate therefrom, would exclude all the mild-symptom people.

Your flu data points to this - estimates are between 10 and 20%. That's a huge range. Reported cases for Covid19 may hit 10% of the world population, say, and 2% of those cases may die, but then untracked, another 10% of the world population also had Covid19 without ever realising, such that the 2% death rate among reported cases is 1% of total cases.

Possibly yea. The R0 for COVID is 2-3 right now as per CDC. Meaning each person infected will infect 2-3 others. So yes this could grow very rapidly if we didn’t take all these precautions that we are now.
It’s the critical balance between not causing panic to the world population, but taking the right steps to lock the virus from spreading.

the fatality rate could be significantly less. That is correct. It could also be significantly higher if people are dying from it before being tested.

 

[dobwal]

The particular examples were Korea and Italy.

Taking the idea of there being many carriers with mild symptoms, the 300 confirmed cases, and all those to follow, and the death rate therefrom, would exclude all the mild-symptom people.

Your flu data points to this - estimates are between 10 and 20%. That's a huge range. Reported cases for Covid19 may hit 10% of the world population, say, and 2% of those cases may die, but then untracked, another 10% of the world population also had Covid19 without ever realising, such that the 2% death rate among reported cases is 1% of total cases.

CFR or Case Fatality Ratio is the ratio of confirmed deaths over laboratory confirmed cases, which is around 2.3% for Covid-19. Modeling is used to produce an IFR or Infection Fatality Ratio which includes diagnosed infections as well as undiagnosed and sits at 0.3 to 1% for Covid-19.

So there are some estimates that place undiagnosed cases close to 120K and others that place them at ~575K.

 

[zed]

In Iran, 43 infected -> 12 deaths there, also ppl from iran have now caused cases in 5 other countries
suggests theres prolly many hundreds if not 1000s of cases in Iran

I wonder how much longer they can pretend in that country its business as normal when stuff like the follwing happens
like some thing out of a comedy film

Iran's deputy health minister tests positive for coronavirus - reports

Iraj Harirchi had been working as normal on Monday, and gave a news conference with journalists in Tehran about the virus during which he reportedly had been sweating and looking uncomfortable.

I can imagine what happened
Iraj Harirchi - "There is no reason to panic theres only a few cases in the country, trust our numbers"
projectile vomits over the journalists
(gets carted away on a stretcher) - "theres no cause for alarmmmmm ...... Its just sun stroke "

 

[Silent_Buddha]

The particular examples were Korea and Italy.

Taking the idea of there being many carriers with mild symptoms, the 300 confirmed cases, and all those to follow, and the death rate therefrom, would exclude all the mild-symptom people.

Your flu data points to this - estimates are between 10 and 20%. That's a huge range. Reported cases for Covid19 may hit 10% of the world population, say, and 2% of those cases may die, but then untracked, another 10% of the world population also had Covid19 without ever realising, such that the 2% death rate among reported cases is 1% of total cases.

Generally true but not quite so simple. If an undocumented Covid19 case results in a death, the death may not end up being attributed to Covid19. One would hope that if someone got sick enough that they were in danger of dying they would seek medical care and should at that point get diagnosed, but that isn't always going to happen, unfortunately.

I almost lost my Mother in Japan to Pneumonia last year. She was really REALLY sick, had a hard time breathing, and couldn't sleep. She's also generally afraid of going to the doctor, although she's gotten a little better about that in recent years. Anyway, she wasn't going to go see a doctor, but luckily her partner made her go in and it's a good thing he did. The physician said that if she had waited another day to come in that it was quite likely she would have died. As it was she ended up being hospitalized for quite a while as they treated her.

And that's in Japan, not some 3rd world country. Thankfully, because of that experience, she's being extra careful of Covid19. She's deathly afraid of catching it from someone.

Regards,
SB

 

[pascal]

CFR or Case Fatality Ratio is the ratio of confirmed deaths over laboratory confirmed cases, which is around 2.3% for Covid-19. Modeling is used to produce an IFR or Infection Fatality Ratio which includes diagnosed infections as well as undiagnosed and sits at 0.3 to 1% for Covid-19.

So there are some estimates that place undiagnosed cases close to 120K and others that place them at ~575K.

I was reading the CNN Live and today there are 80,000 cases registered with 2700 fatalities, and most cases are not finished yet. So my guess the CFR is higher.


https://edition.cnn.com/asia/live-news/coronavirus-outbreak-02-25-20-hnk-intl/index.html

 

[iroboto]

https://www.thestar.com/news/canada...than-reported-canadian-research-suggests.html

The true number of coronavirus cases in Iran may be upwards of 18,000, according to a preliminary analysis by Canadian researchers, suggesting an epidemic there nearly 200 times larger than what the country has reported — and a situation with potentially grave global consequences.

A spokesperson for Iran’s health ministry said Tuesday that the country had 95 confirmed cases of COVID-19, as the disease caused by the novel coronavirus is called, and 15 deaths.

Well this is very worrying if this estimate is accurate. @zed seems you beat me to it.
Unlike China, Iran is unlikely to be able to tightly contain their population from fleeing (thus spreading). I hope this does not keep progressing like this.

 

[orangpelupa]

Oh my God. Now I'm really selfish here. I really really hope Indonesia didn't get infected.... I don't have a strong confidence in the ability to contain the infection.

Hopefully an antiviral or vaccine can be developed quickly. Or the weather changes again to something that the virus can't stand

 

[Shifty Geezer]

Oh my God. Now I'm really selfish here. I really really hope Indonesia didn't get infected....

If the infection isn't there yet, it will get there once it breaks confinement. World travel ensures as much. It won't be a case of 'if' but 'when'. Short of closing all borders and ceasing interactions with the rest of the world until the virus has passed, it cannot be avoided. If it's rife and uncontrolled in Iran (15 deaths suggests some 1000+ infections going by other death rates), it'll be spreading from there all over the place.

 

[Gubbi]

The first 8 minutes 45 seconds are pleasantries.

China has done a top-notch job of containing Covid-19.

Cheers

 

[pharma]

So much for globalization. Companies that manufacture medical products (face masks) have moved production to China because of cheap labor. Any face masks in the US are currently out-of-stock (all retailers) and masks in China are either being diverted locally or manufacturing lines are shut down.

The govt should mandate US companies need some factories in the US as a matter of national security.

 

[AlNom]

The govt should mandate US companies need some factories in the US as a matter of national security.

Sounds a little authoritarian. Anyways, someone has to pay for it, and clearly it'll be expensive.

I suppose there might be some way a govt. could raise money via "crowdfunding", but that might need people to opt-in, so I don't know! ¯\_(ツ)_/¯

 

[Kaotik]

First confirmed case for Finnish citizen, got it from Milano, Italy, from where she returned to Finland on Sunday.

 

[BRiT]

Ooooof...

• California, Nebraska and Illinois are the only U.S. states that can currently test for coronavirus, the Association of Public Health Laboratories (APHL) told Reuters on Friday.
• The U.S. Centers for Disease Control and Prevention (CDC) last week said some of the testing kits sent to U.S. states and at least 30 countries produced “inconclusive” results due to a flawed component.

https://twitter.com/i/web/status/1232318905156276224

 

[BRiT]

https://twitter.com/i/web/status/1232777025145974791

Microsoft says "the supply chain is returning to normal operations at a slower pace than anticipated". No longer expect to meet its 'More Personal Computing' segment guidance as Windows OEM and Surface sub segments are more negatively impacted than previously anticipated.​

 

[iroboto]

Ooooof...

• California, Nebraska and Illinois are the only U.S. states that can currently test for coronavirus, the Association of Public Health Laboratories (APHL) told Reuters on Friday.
• The U.S. Centers for Disease Control and Prevention (CDC) last week said some of the testing kits sent to U.S. states and at least 30 countries produced “inconclusive” results due to a flawed component.

https://twitter.com/i/web/status/1232318905156276224

oooh

 

[dobwal]

oooh

doh. They could of sent out the protocol to hospitals willing to develop their own test while sending out test panels to ensure quality of accuracy and lower limit of detection.

The test panel part is required for their test kits anyway. You need one to make sure your staff is correctly following the protocol and producing reliable results.

 

[Picao84]

Sounds a little authoritarian. Anyways, someone has to pay for it, and clearly it'll be expensive.

I suppose there might be some way a govt. could raise money via "crowdfunding", but that might need people to opt-in, so I don't know! ¯\_(ツ)_/¯

I see what you did there.

 

[zed]

Come on, trump has declared there will be zero cases in a few days in the states.
And just like he showed us with his hurricane path prediction the guy is a stable genius and is the Delphi reincarnate