Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

reuters.com

Post reply on HN
Print view View original

421–430 of 544 posts

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#421

Post by cm2187 »

jpxw wrote: post #416
cm2187 wrote: post #408

Earlier quoted context omitted.

Though my understanding is that the survival rate of a covid19 patient once he requires a ventilator is around 50%. If that's true, at worst, if we ran out of ventilators, the death rate would double. That's bad and we should avoid it but the doubling of a tiny number will still be a small number. Which is why I am still confused by the ratio of 1:10 that the White House has shown between the death toll under lockdow…

Some 15% of people become “seriously” unwell, and require medical intervention such as oxygen. In a situation where the peak is hit unmitigated, these people will not get the medical care they need, and a lot of them will die.

The 15% looks implausible given the fatality rates mentioned here.

But I would expect a lot of the medical care to be scalable so unmitigated doesn't mean no medical care. We have been building field hospitals in the UK and US (though have barely used them so far), using hotels, etc. Providing oxygen should be fairly straightforward. I understood ventilators was the part that caused a serious scalability bottleneck.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#422

Post by xorfish »

nostromo wrote: post #376
vrtx0 wrote: post #351

IMHO, this article is intentionally misleading. The incubation period is currently estimated to be 2-14 days (mean is 5.6 days per CDC, similar per WHO). The article doesn’t mention any dates or time frames, but does mention: ”Roughly 60 percent of the over 600 sailors who tested positive so far have not shown symptoms of COVID-19” — note how “so far” is ambiguous in that sentence. It also states, “The Navy’s testing…

Just as one data point, the captain of the ship raised alarms about the virus spreading late last month. So it’s not exactly a recent outbreak. Of course, without testing from that period it’s impossible to know.

If the virus spreads exponentially, you would expect most cases to be presymptomatic at any given time.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#423

Post by xenonite »

xenonite wrote: post #332
system2 wrote: post #322

Earlier quoted context omitted.

If 0.37% is true, why are the hospitals in this shape?

These were among the very first cases in Germany where capacity was abundant, hence the low fatality rate. Also, this is a “up to now” value, more cases might turn out fatal. “Only” 7 deaths is furthermore not a good data basis for calculating this number. Moreover, the report is not even public nor peer reviewed. So I can’t even cite it here. However you find criticism of the study by Prof. Streeck online.

At least, here is a link to a preliminary summary of the study, in German language. Please note how shallow the text is in its current state.

https://www.land.nrw/sites/default/files/asset/document/zwis...

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#424

Post by sfj »

fzeroracer wrote: post #354
sfj wrote: post #256

Earlier quoted context omitted.

> Someone has cancer for six months, gets COVID and dies 17 days later and now Tucker Carlson is here telling me not to count that death. If someone has cancer and then gets a bad cold and dies, what should the cause of death be? Or what if they fell, etc.?

The cause of death would be the cold, yes. That's how tallying these statistics work and yes, people can die from the cold. If someone ends up catching pneumonia because their immune system was compromised as a result of immunotherapy, pneumonia was the direct cause of death and the cancer lead up to pneumonia. The fact that we have to have this argument about basic death statistics is depressing. If you really wante…

> If you really wanted to play around with the numbers like what you're suggesting, you could easily make it so covid-19 causes zero deaths just by looking far back enough to find a contributing factor. Like age, immune system issues, other injuries, obesity etc.

Or that cancer causes zero deaths...

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#425

Post by sfj »

projektfu wrote: post #286
sfj wrote: post #256

Earlier quoted context omitted.

> Someone has cancer for six months, gets COVID and dies 17 days later and now Tucker Carlson is here telling me not to count that death. If someone has cancer and then gets a bad cold and dies, what should the cause of death be? Or what if they fell, etc.?

If someone had AIDS and died of pneumocystis pneumonia, the latter is the cause of death, and the former is the cause of the latter.

So, if there was an epidemic of pneumonia going around, would you count that as a death due to pneumonia, and then use that statistic to help justify lock downs and enforced social distancing?

Because that is what is happening when you take an extremely sick person who finally died and slap a "death by covid-19" label on them.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#426

Post by jobigoud »

ImaCake wrote: post #105
andrewla wrote: post #13

There's always the chance that the test itself is not particularly good and just has a high false positive rate. Although it is difficult for a PCR test to produce a match if there is no genetic material present, it may amplify trace signals, depending on the cutoffs chosen.

I looked into this a few weeks ago. The PCR test has an incredibly low false positive rate, below 1%. It is low enough that it is feasible to test a 100,000 people and expect basically zero false positives. The false negative rate is somewhere around 30%. So the real danger with PCR is that you will miss cases. The opposite is true for the antibody tests. The false negative rate is around 10% and the false positive r…

> The false negative rate is around 10% and the false positive rate is also around 10%. So in 100,000 tests, you would expect to miss about 1,000 true positives. > you would also mistakenly get 1,000 false positives

I think you mean missing 10,000 true positives and getting 10,000 false positives.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#427

Post by xorfish »

zephyrnh wrote: post #340
gdubs wrote: post #193

Stanford’s study should give us more info soon. They recently screened a decent sized group in Santa Clara County to get a sense of how widespread it is, using their newly developed finger-prick test which looks for both antigens and antibodies (whether you have it / had it and recovered). Currently the test is being prioritized for frontline healthcare workers while they work on scaling up: https://www.stanforddaily…

It's been 12 days since they tested everyone. My understanding is that the results for serological tests are available right away, so I was expecting them to publish results within a few days. Any idea why they haven't yet?

They would need to collect data again now to get a better result. Otherwise the data is the same as on the day the test was made.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#428

Post by jobigoud »

GANG718 wrote: post #2

This is because they're testing the entire population as opposed to simply those who end up hospitalized and those around them. We need population testing to get an actual idea of how many people in an area have it. There's likely way more cases nationwide than being reported. In NYC, most asymptomatic/low symptom cases of coronavirus aren't going to be tested until antibody testing later on.

> This is because they're testing the entire population

Not really, they are testing a population of only young, healthy and fit individuals with an obesity rate and other conditions absolutely not in line with the general population.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#429

Post by yomly »

cm2187 wrote: post #421
jpxw wrote: post #416

Earlier quoted context omitted.

Some 15% of people become “seriously” unwell, and require medical intervention such as oxygen. In a situation where the peak is hit unmitigated, these people will not get the medical care they need, and a lot of them will die.

The 15% looks implausible given the fatality rates mentioned here. But I would expect a lot of the medical care to be scalable so unmitigated doesn't mean no medical care. We have been building field hospitals in the UK and US (though have barely used them so far), using hotels, etc. Providing oxygen should be fairly straightforward. I understood ventilators was the part that caused a serious scalability bottleneck.

Yes - part of the strategy of flattening the curve is to buy time to scale capacity. Dedicated covid19 hospitals also have the advantage of isolating covid19 patients from high risk non-covid19 patients (e.g immunocompromised cancer patients)

As the disease is so easily transmitted hospitals are partly buckling due to contamination within hospitals. A regular 50 year old probably has a good chance of surviving coronavirus. A 50 year old who is midway through chemo, probably less so...

All of this nuance is hard to express through topline death tolls and mortality rates.

Re: Coronavirus clue? Most cases aboard U.S. aircraft carrier are symptom-free

#430

Post by neuronic »

sveme wrote: post #378
manfredo wrote: post #103

Random testing in Germany has also revealed higher than expected rates of people with antibodies. Their IFR (infected fatality rate) was 2% overall based on non-random testing. Random testing has yielded an IFR or 0.37%. This does not increase the risk. These tests only test people who had coronavirus. The larger this portion, the more herd immunity there is in the population. It implies that there were a larger than…

Though that study has been criticized for the use of antibodies with cross-reaction with other Coronaviruses, so the numbers aren‘t necessarily to be trusted.

This is a general problem of antibody tests and precisely the reason why people should not buy those from any retailer (you can in fact do so online or in pharmacies). Unreliable and non-specific.

Let's wait for the development of specific viral antigen tests.

Post reply on HN