Hacker News Reader

Comment by bla3 | Hacker News Reader

NYC has 20x the deaths / 100k people, so under the assumption that number of deaths are on comparable trajectories (due to both locations being under lockdown, maybe not as unrealistic as a few weeks ago), NYC's prevalence could be 56%-112%. This could be one explanation for why numbers are going down in NYC, or the 112% being larger than 100% could suggest that either the situations aren't comparable or that the LA study's numbers are a bit high.

View on HN

LA's numbers seemingly agree with Santa Clara's numbers as well. I would guess NYC probably has a much higher rate of infection (possibly due to http://web.mit.edu/jeffrey/harris/HarrisJE_WP2_COVID19_NYC_1... ?) But 50%+ seems pretty high.

Is it also possible different strains are dominant in California and NYC with different mortality characteristics? There is also the issue of weather. A recently leaked DHS report indicates sunlight radically reduces virus half-life and it's also affected by temperature and humidity. Maybe subway + weather + (possibly) strain of differing virulence can explain the discrepancy?

'More subway' doesn't really capture it. NYC has ~70x the daily ridership on metro rapid transit.

https://en.wikipedia.org/wiki/List_of_United_States_rapid_tr...

Overall population density is much higher as well:

https://en.wikipedia.org/wiki/List_of_United_States_cities_b...

NYC also has the lowest car ownership in the nation, so you're basically left with walking or uber/taxi/public transportation just to get to the store/doctor/etc.

https://www.governing.com/gov-data/car-ownership-numbers-of-...

Also ~25% more airline traffic hauling bugs in from everywhere.

https://en.wikipedia.org/wiki/List_of_busiest_city_airport_s...

Combined with a long transmissable dwell time before symptoms, wouldn't surprise me if 90%+ of the people in NYC were materially exposed.

Spain & Italy have very similar weather to California, but have been hit really hard.

Because of demographics. The disease really, really doesn't affect young people. The younger, the better you do, to the extent that the percent of people under 20 that have died in every major area of the world rounds to 0%. Older demographics are hit hundreds of times harder than young to middle aged folks.

Italy has the oldest average population in Europe, and Lombardy has the oldest average population in Italy. 99.2% of those who died in Italy averaged 80.5 years old and had an average of 3 co-morbid conditions. [1]

[1] https://www.bloomberg.com/news/articles/2020-03-18/99-of-tho...

Italy also appears to have suffered from tragic public health errors. Apparently they moved covid-19 infected patients around. First to other hospitals that had not yet had outbreaks and then into empty beds at long-term care facilities.

https://www.cbc.ca/news/covid-19/italy-covid-19-outbreak-les...

It affects young people, it just doesn't kill them. They spread it as well as anyone else.

"spread" isn't "affected" -- someone with no symptoms who spreads it really isn't affected. The people who get it from them may well be.

So far the data is clear: it doesn't really do much to young people. [1] Speaking broadly, almost nobody under 20 has died, practically nobody under 30 has died, and a handful of 30-40 year olds have died. The trendline is super clear: if you're young, you're gonna be just fine. The CDC says young folks aren't even likely to wind up with serious disease or in the ICU ("no ICU admissions or deaths were reported among persons aged ≤19 years. Similar to reports from other countries, this finding suggests that the risk for serious disease and death from COVID-19 is higher in older age groups.") [2]

When I say "affected" I mean that they are not likely to develop serious symptoms, certainly severe symptoms and they're definitely not likely to develop life-threatening symptoms.

[1] https://www.cebm.net/covid-19/global-covid-19-case-fatality-...

[2] https://www.cdc.gov/mmwr/volumes/69/wr/mm6912e2.htm

We know for a fact(well based on this report from a Chinese univesity[1]) that there are two strains of covid-19, I wonder if the less aggressive strain has been around awhile more than the more aggressive strain, it may have been around a lot longer than the outbreak was announced (think fall 2019) and may have been misidentified as the flu/bad cold. And it would explain why some people are getting extremely mild cases, they either have the less agressive strain or already had it but now have the antibodies to fight it. This is total conjecture mind you and the weather aspect is also interesting, look at Hawaii's cases and death rate(580 cases to 10 deaths, they are both very low numbers).

[1] https://www.forbes.com/sites/lisettevoytko/2020/03/04/discov...

That study is really not very good and doesn’t tell you anything.

Sure, mutations happen all the time, however since the virus has error correction those are rare and very unlikely to have any effect on the phenotype.

I don’t think 50% seems that high given the death count.

If the true infection rate is 50% they are very unlucky with their tests because less than 40% of the people tested give a positive.

Not really. Antibody tests like this one measure the total number of people who've been infected at some point in the relatively recent past, whereas the diagnostic tests New York is using only detect current active infections, so it's easy for the percentage of infections measured by the former to be higher than the latter.

When articles like this compare the number of known cases detected with the PCR diagnostic tests with the higher estimate based on antibody testing, they're comparing the sum of all cases detected since the start of PCR coronavirus testing with a point-in-time estimate of the proportion of the population that currently has antibodies for exactly this reason.

That’s true. We could have 25% active infections and 25% not currently infected but who were infected in the past. I still think the actual numbers may be be lower (but higher than the reported figures, of course).

> less than 40% of the people tested give a positive.

FWIW, Queens and Brooklyn (Kings county) are at 50.7% and 48.3% positive tests respectively.

https://covid19tracker.health.ny.gov/views/NYS-COVID19-Track...

I meant people being tested nowadays, not cumulative numbers.

Today: Kings 548/1835 = 29.9% and Queens 786/2428 = 32.4%

If we're going to talk about the total number of people who have been infected, wouldn't it make more sense to use the overall test rate?

My point was that the number of people currenlty infected is surely lower than 30%, given that only 2% of the population is already confirmed and testing a suspected case today will return a negative 70% of the time.

But as another commenter said, it's not impossible that there are many other people who had infection which is not currently active.

Using the overall test rate doesn't really help, I think.

50% means every other person would have to have been infected. That seems a bit unlikely to me but I suppose it's possible.

It seems very improbable since the positive rate for tests in general are around 10% (at least in San Mateo county).

It doesn't seem that unlikely to me (I live here). I know so many people, me included, who had it but aren't included in any official count for lack of testing, and so many others who must have had it (e.g. my cohabiting SO) who were asymptomatic.

You can't just take numbers and multiply them like that.

What about average age? Number of icu beds? Effect of temperature/humidity on the virus? Population density? &c. There are so many variables

The point is that either these numbers are not accurate, or that there are crucial differences between LA and NYC. Either conclusion is interesting.

I don’t really understand your point. The error bars would be calculated differently for a larger sample. If you had 100/100 positives with a five percent error you wouldn’t say the true amount is 95 to 105. You’ve done some sloppy statistics and claimed this invalidates the work...

NY has performed 32k tests per 1M population, CA 7k per 1M population [1]. If CA PCR-tested as much as NY, there would probably be way more official cases and your calculation would make more sense.

[1] https://www.worldometers.info/coronavirus/country/us/

edit: typo

Death rate higher could be due to an overwhelmed health care system, so it is not linear. Once that threshold is hit, it will probably edge close to rate of severe cases among the patients.

NYC is massively overreporting deaths. Just the other day they added 3700 "new" deaths that didn't even test positive, they just suspected (ie. pulled it out of their behind) that it was cause of death.

Death numbers are overreported in general - Dr. Birx herself even claimed that if you test positive but die of something other than the virus they track it as a COVID death. It's preposterous.

Belgium is doing something similar. They're being more aggressive than any other nation in Europe at labeling deaths Covid. It has contributed to their extreme case mortality rate.

https://www.politico.eu/article/why-is-belgiums-death-toll-s...

I listened to Birx[0] and perceived the same. I continue to wonder what I'm missing. If she's correct, there would be serious problems with such methods. Strange stuff.

0. https://m.youtube.com/watch?v=blZpgra3XbU

Reply on news.ycombinator.com