Coronavirus 2019-nCoV

[Napoleon Dynamite]

I read that too, when it came out. Unfortunately, it predates the WHO's more recent CFR estimate by about a week, and the studies the article cites are older still. For example, the paper by Guan et al. cited there does indeed report a CFR of 1.4%, but if you read Guan itself, you'll see that of 1099 patients studied, 15 (1.4%) had died, 55 (5.0%) had been discharged, and 1029 (93.4%) were still in the hospital at the time the study was done. The data is from January 28th. It was a helpful early look at an early patient population, but the outcomes for the overwhelming majority of those patients weren't known then. It's not a relevant sample or time frame now.

It seems now, over a month later, that lots of those who die take weeks to expire. That's not good news for how the numbers will trend. On the other hand, as has been discussed above, there's also bias in the other direction from all the unreported cases (which we're assuming are disproportionately mild). We'll have to see where the final balance lies. For now, don't rely on old data, and be sure to check your underlying sources to be sure they're relevant for whatever you're trying to learn. Pay special attention to the population of interest, time frame, outcomes definitions, etc

As it turns out the NEJM article (dated Feb. 28) is in line with the data reported by South Korea. Maybe I'm missing it, but I can't find any sort or retraction or disclaimer stating that the article relied on older or incorrect data. You would think a journal with the reputation of NEJM would be quick to correct or retract an incorrect study.

[moshock]

I deal with a couple of business' in China on a daily basis. This is a copy of an email that I received last night:


"For the virus situation, we are sure that it's getting way better than it was in Spring holiday. Originally we are afriad that people would get infected after they return back to work due to the mass population travel flow, but there is 0 case patient found in many of the provinces in China for many days, which means that all the methods we are taking to fight the virus are working and it's getting better gradually. And for our city, we have not had new patients since we are back to work. There are 49 patients committed into our local hospital since the virus outbreak and for now all of them have already checked out and stayed at home for observation. That's really good news for us. I even took my son out to the river side to hang out the whole afternoon yesterday. And I am very glad to know that you are safe from the virus zone in the US. Please just be alert but not panic and also take care. Please let me know you need other info.




Have a nice day!




Michael "

The "spring holiday" that he is talking about is Chinese New Year. It's two weeks (around the first of February) that their country basically shuts down. It's an annual deal and it's one of the many difficult situations that we deal with.

The companies that I deal with re-opened two weeks after CNY (about 2 -3 weeks ago) and are working 2 shifts to catch up. One is in Shanghai and the other outside of Qingdao. They both employ several hundred people. Neither have any employee's who have the virus.

For those who have never visited China - it is nothing like the US. There are masses of people all shoved in very small places. They all share these small spaces. There is no privacy or "personal space". The air has a stench to it and filth is everywhere. There are literally people cooking in the streets and selling it. There is no health department (that I could tell) to over see the sidewalk restaurants. Most of the living conditions for normal Chinese people were slums - worse than any place I've ever seen in Wichita.

I can see how a virus could spread throughout that whole nation - but it hasn't.

I can also see how it wouldn't be near as bad in the USA. We have many codes and regulations that require us to be sanitary in just about everything we do. We may well have an outbreak - but I have to believe that it will be minimal compared to other countries.

I flew from Phoenix to DFW last night. There were people cleaning everywhere I looked in the airports. I doubt that those buildings have ever been cleaner or more sanitary than they are now.

IMHO, this is being overblown by the media - is it serious ? Yes - so is the flu. Is it the plague ?? Nope. Just be smart about things and watch your cleanliness and contact with others....we are gonna be ok ShockerNation ! Go Shox

Spock.jpg

Live Long and Prosper !!

MoShock Out

[wufan]

As it turns out the NEJM article (dated Feb. 28) is in line with the data reported by South Korea. Maybe I'm missing it, but I can't find any sort or retraction or disclaimer stating that the article relied on older or incorrect data. You would think a journal with the reputation of NEJM would be quick to correct or retract an incorrect study.

NEJM comes out weekly. The January data would probably be the most up to date available for a peer reviewed article. Why would they retract it?

[DCShockerFan05]

Yeah I wasn’t saying NEJM is wrong, I’m saying that article is old and in a situation evolving as rapidly as this the shelf life is pretty short. It’d be like checking stock quotes from a month ago to see if the virus is affecting the economy today...

[Napoleon Dynamite]

NEJM comes out weekly. The January data would probably be the most up to date available for a peer reviewed article. Why would they retract it?

Agreed. Exactly my point.

The article was published on Feb. 28 using January data, and somehow that is "old" data. Really?

[Napoleon Dynamite]

I deal with a couple of business' in China on a daily basis. This is a copy of an email that I received last night:


"For the virus situation, we are sure that it's getting way better than it was in Spring holiday. Originally we are afriad that people would get infected after they return back to work due to the mass population travel flow, but there is 0 case patient found in many of the provinces in China for many days, which means that all the methods we are taking to fight the virus are working and it's getting better gradually. And for our city, we have not had new patients since we are back to work. There are 49 patients committed into our local hospital since the virus outbreak and for now all of them have already checked out and stayed at home for observation. That's really good news for us. I even took my son out to the river side to hang out the whole afternoon yesterday. And I am very glad to know that you are safe from the virus zone in the US. Please just be alert but not panic and also take care. Please let me know you need other info.




Have a nice day!




Michael "

The "spring holiday" that he is talking about is Chinese New Year. It's two weeks (around the first of February) that their country basically shuts down. It's an annual deal and it's one of the many difficult situations that we deal with.

The companies that I deal with re-opened two weeks after CNY (about 2 -3 weeks ago) and are working 2 shifts to catch up. One is in Shanghai and the other outside of Qingdao. They both employ several hundred people. Neither have any employee's who have the virus.

For those who have never visited China - it is nothing like the US. There are masses of people all shoved in very small places. They all share these small spaces. There is no privacy or "personal space". The air has a stench to it and filth is everywhere. There are literally people cooking in the streets and selling it. There is no health department (that I could tell) to over see the sidewalk restaurants. Most of the living conditions for normal Chinese people were slums - worse than any place I've ever seen in Wichita.

I can see how a virus could spread throughout that whole nation - but it hasn't.

I can also see how it wouldn't be near as bad in the USA. We have many codes and regulations that require us to be sanitary in just about everything we do. We may well have an outbreak - but I have to believe that it will be minimal compared to other countries.

I flew from Phoenix to DFW last night. There were people cleaning everywhere I looked in the airports. I doubt that those buildings have ever been cleaner or more sanitary than they are now.

IMHO, this is being overblown by the media - is it serious ? Yes - so is the flu. Is it the plague ?? Nope. Just be smart about things and watch your cleanliness and contact with others....we are gonna be ok ShockerNation ! Go Shox

Spock.jpg

Live Long and Prosper !!

MoShock Out

Unfortunately one population (other than the elderly) that I suspect will be very hard hit by COVID-19 is the homeless population in this country.

[SB Shock]

Keep an eye on South Korea. That's closest to our future.

SK case curve has started going from exponential to "S" curve. So it is looking like their aggressive measures are turning tide.

[Kung Wu]

SK case curve has started going from exponential to "S" curve. So it is looking like their aggressive measures are turning tide.

I'm cautiously optimistic. Still want to see a tad bit more data though.

[DCShockerFan05]

As usual, Randall Munroe at xkcd makes the point in four panels instead of four pages.
scientific_briefing.png

https://xkcd.com/2278/

[Kung Wu]

Per this site, 51 dead and 166 cases recovered in S Korea. That’s over 20% of resolved cases (assuming all positives were hospitalized) leading to death.

The 20% isn't correct.

You can't compare dead today (with a timeline of about 10 days) with the recovered today (with a timeline of about 22 days). If you want the real % dead of resolved cases, you have to use the number of dead from 12 days prior to the 166 day.

If we open the data that is driving the chart, I can go to day 2020-03-08 and see your 166 number. Then I can walk up the chart about 12 days and count the total deaths at that point in time (I'll use 2020-02-26 - you decide if it's the 25th, 26th or 27th): There were 12.

So it's really 12 / 12 + 166 = 0.067

P.S. I made the 10 and 22 days numbers up. Use whatever number you have read that you think creates an accurate delta between being recovered and average days until death.

[DCShockerFan05]

I'm not sure I follow what you're saying, Kung Wu. But I also don't think wufan's analysis is watertight either--my issue with it is the assumption that all positives were hospitalized--that's almost certainly not true. If we only look at hospitalized cases we're selecting for a much sicker subgroup and we would expect to see much higher mortality in that subgroup. So if the question is "what's the risk of dying if you're hospitalized with COVID" then those numbers may be relevant. If the question is "what's the risk of dying if you're infected," they're not.

I do think that fatalities among resolved cases is a meaningful number. Not sure why the time to death vs time to recovery has anything to do with it.

[Kung Wu]

I do think that fatalities among resolved cases is a meaningful number. Not sure why the time to death vs time to recovery has anything to do with it.

Long term it doesn't matter and statistically fades into the CFR you are familiar with. In the ultra-short term you cannot compare an event that occurs in 10 days with an event that takes 22 days to occur. You will get a really, really false narrative if you do that.

Here's a severely oversimplified example just to show the math ...

100 people get sick simultaneously, and nobody else in the world ever gets infected:

Day 1: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 2: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 3: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 4: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 5: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 6: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 7: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 8: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 9: Deaths - 0, Recoveries - 0, CFR = 0/0
Day 10: Deaths - 1, Recoveries - 0, CFR = 1/1 + 0 = 100%

If you "run the numbers" right here, you would think that 100% of people will die that get the virus, because there are no recoveries yet.

But if you wait until both events have reached a conclusion ...

Day 21: Deaths - 1, Recoveries - 0, CFR = 1/1 + 0 = 100%
Day 22: Deaths - 1, Recoveries - 99, CFR = 1/1 + 99 = 1%

Now that both events from the first batch have reached their conclusion, you know that there is a 1% CFR.

As enough time and data rolls in that "problem" of calculating a CFR in the ultra-early stage fades away (the error gets closer to zero as more data pours in).

[wufan]

Granted, it’s almost certainly a false assumption that all cases are hospitalized. I was just looking for deaths vs recoveries. It could be that the info I was searching for is either unavailable or is irrelevant as 7000 or so were mostly asymptomatic.

[WstateU]

[WstateU]