Hacker News Reader

Comment by javagram | Hacker News Reader

Your quoted paper is 5 weeks old, as shown by the excerpt “ compared with less than 4000 deaths for SARS-CoV-2 at the time of writing.”

Now, today, we are at 50,000 confirmed COVID-19 deaths just in the USA, and an analysis of excess mortality data already shows that a significant % of COVID-19-related deaths are not included in that count.

This is with a massive mitigation campaign and nationwide social distancing. The only place where the virus seems to have established itself is in NYC (20% of people infected according to a recent study) where we have seen incredible death tolls and overwhelmed morgues. (Edit, needless to say, none of the 4 common cold coronaviruses do this, ever.)

View on HN

The data presented by the two doctors is fresh, and shows exactly what the paper says. COVID-19 is not worse than the flu.

The “data presented by the two doctors” has obvious reasoning errors to anyone who has taken even high school math.

At this point, how can you look at the number of people who died from COVID-19 in just the last month in NYC and claim this is no worse than the flu? https://www.nytimes.com/interactive/2020/04/27/upshot/corona...

The last flu to cause such a severe spike in mortality statistics was the devastating 1918 Spanish flu. Even the Asian Flu and Hong Kong Flu pandemics didn’t do this.

Edit: corrected the Spanish flu year. My bad :-)

>The last flu to cause such a severe spike in mortality statistics was the devastating 1918 Spanish flu

The second wave of the 1918 flu killed between 17 to 50 million people, perhaps as high as 100 million people, during the fall of 1918, with a higher than expected mortality rate for the young and healthy. It killed between 1 and 6 percent of the world's population.

So far 200,000 people have died from the coronavirus.

The current coronavirus is 1/100 as deadly as the flu of 1918.

Your argument assumes that coronavirus deaths will stop tomorrow, and also that mitigation measures used to reduce deaths from coronavirus via social distancing don’t affect the evaluation of whether it is “deadly”.

It feels like you didn’t engage with the data I presented at all. (Edit: removed sentences that would only apply if you had read my other post)

Edit 2: if your only argument is that COVID-19 is not worse than Spanish flu, that’s something that is reasonable. It’s worse than any pandemic flu since Spanish flu though, this is a once in a century occurrence.

I apologize, there's a lot of numbers flying around and I probably did miss the thrust of what you were saying. I shouldn't have focused on the comparison with the 1918 flu.

Plus, as someone pointed out, drawing conclusions from the numbers we have so far in comparison with the 1918 flu is difficult.

I've been working with NYC numbers. The link you gave was for NY State deaths. NYC has reported about 12.3k deaths [1], has a population of about 8.4 million [2] and 21% of people appear to have antibodies at this time.

Doing some math with the above numbers I end up with .14% dead so far with 1/5th of people exposed. From there I assume the amount of dead would quadruple before we reach herd immunity. I also assume that antibodies provide at least a handful of years of immunity, just like for other coronaviruses, reducing the severity of future infections. A likely outcome is that this becomes another variant of several coronaviruses that circulate with the common cold.

I guess even for those who might agree with all of the above there still exists differences in opinion on what actions should be taken. I personally understand some people's comparison with the yearly flu, but realize that is most definately not an apples to apples comparison.

[1] https://www1.nyc.gov/site/doh/covid/covid-19-data.page [2] https://en.wikipedia.org/wiki/New_York_City [3] https://www.nytimes.com/2020/04/23/nyregion/coronavirus-new-...

Thanks.

Yes I believe your rough math is matching others that are reported by the scientific literature which is estimating around 0.7% infection fatality ratio (IFR) for this virus, 4*0.14 would be on the low end of estimates but within the consensus estimate. Seasonal Flu IFR is estimated to be between 0.04% and 0.1%, so by your numbers the coronavirus is, at a minimum, 5 times more deadly than seasonal flu.

I also agree this virus will become a part of life in the future. Although some Asia-Pacific countries seem to have successfully controlled the spread (Vietnam, Taiwan, etc) the only way for them to keep it from spreading will be to maintain a border closure until a vaccine can be developed to safely give their populations herd immunity.

In the west, we let the virus run free for two months and it’s probably too late to get it under control no matter what we do now.

> The current coronavirus is 1/100 as deadly as the flu of 1918.

Wow: you are saying the final death rate from Covid19 will be from 0.01% to 0.06%.

You are making irrelevant comparisons: we don’t yet know how many people will die from Covid19; we have no idea of numbers of deaths in many poorer countries where deaths can’t even be counted; comparing number of deaths in a world with 7 billion people against number of deaths for 1918 population is pointless.

I agree with what you are saying, the numbers I presented in response to a comparison with the 1918 flu are not good to draw conclusions from.

In fact I will admit that thinking further about the numbers we have so far I am sobered by how they may add up across the world.

But back on topic, talking about lockdowns: The stated purpose of the lockdown was to flatten the curve so that hospitals are not overwhelmed, which has been successful so far. Do you think lockdowns lower the number of deaths when all is said and done? How long can lockdowns be sustained? Would it be more effective for vulnerable folks to self-isolate? Unless we can mass produce a vaccine in the next couple of months won't this continue moving through the population regardless of what we do?

> So far 200,000 people have died from the coronavirus.

No. We don't know the numbers because most countries don't have the numbers yet. This 200,000 number includes data from countries like the UK or US who are only counting confirmed covid-19 deaths. We know that deaths outside hospital make up about half the total, so we'd expect about 400,000 deaths.

The death counts rise extremely quickly, but also drop quickly as well. The total amount of deaths is represented by the area of under the graph, not the maximum. This is a tall, but narrow spike.

Also, bear in mind that you're being very selective with your data: you're picking out one of (the?) most densely populated centers, which also happens to be an international travel hub

There isn’t any reason to believe the virus kills a resident of NYC any easier than it kills a resident of Kansas if either is infected with it.

NYC is an excellent example to look at precisely because of what you said - it’s a dense population center and international travel hub, so it was seeded with quite a few coronavirus cases early on and unfortunately the virus seems to have infected up to 20% of the residents of the city (as shown by a new serology survey), despite social distancing having begun mid-March.

Other regions outside NYC weren’t as affected in terms of mortality, and we have no indication that’s because of anything other than that not as many people outside NYC fell ill with the virus.

The drop in deaths after the spike is because of data reporting (deaths are commonly revised upward as the data trickles in) and also because 6 weeks of social distancing and school closures would have greatly reduced the spread of the virus, this is just common sense as we know the virus spreads through physical proximity of an infected person with non-infected persons, and social distancing has greatly reduced the number of those situations.

The infection rate relevant insofar as it means the population will approach 100% infection [1],- and subsequent immunity or death - faster than other regions. Densely populated areas like New York will have a sharp spike and sharp decline. Loosely populated areas like Kansas will see a slower, steadier rise but also a slower and steadier decline. Social distancing makes areas behave more like rural areas, by virtue of limiting contact. But the same number of people are going to get infected eventually until herd immunity is achieved.

Faster infection doesn't directly result in higher total fatalities, unless medical care lacks capacity. This is a bigger problem for places like New York where the infection rates did put significant burden on healthcare facilities. On the flip side, places less dense than NYC can go without restrictions and this will not impact the overall health outcomes.

Deaths nation-wide have peaked more than a week ago. Deaths among developed countries have also mostly peaked over the last few weeks. Testing almost always under counts, both by virtue of not counting asymptomatic people, and by delays in forming antibodies in the infected. The latter also affects random studies. New York's 20% indicated infection rate likely means that the infection rate today is approaching the 50% or more required to achieve herd immunity.

1. To be more specific, it's estimated that the virus will spread to 50-70% of people before herd would diminish the ability of the virus to spread much further.

> New York's 20% indicated infection rate likely means that the infection rate today is approaching the 50% or more required to achieve herd immunity.

Yes, if by approaching you mean at least twice the number of people who have already died from the virus would have to die to achieve this, since 20% is less than half of 50%. I’ve seen scientific experts estimating that with this specific virus, due to its high R0, 50-70% is not the threshold for herd immunity anyway and it could be up to 84% or higher (Polio = 80%, Measles = 95% thresholds, for comparison). That could mean up to 4x the amount of deaths we’ve already seen.

The total number of people who would be affected by a herd immunity strategy could also be changed by a social distancing intervention around the time the threshold is reached. Otherwise, models indicate there would be an overshoot and a higher % of people would end up being infected than is necessary for herd immunity. No country or region seems to have reached this threshold yet though since they all engaged in social distancing first. https://twitter.com/CT_Bergstrom/status/1252078649827667968

The other option would have been to do what several countries in the Asia Pacific region did - Taiwan, Vietnam, and apparently China, for instance. Instead of allowing the virus to run free, they have worked toward eliminating it within their borders. Many lives will be saved if that plan can succeed and they can avoid a “herd immunity” result until a safe vaccine is available to provide the herd immunity.

The R0 of Measels is 10-3x as high as COVID-19, at 12-18 as compared to 1.7 to 5. The total number of people affected to achieve herd immunity is fixed. Distancing reduced the rate at which people get infected. And contrary to what you claim, several countries are not engaging in social distancing. And they're not experiencing death rates several times higher than the US.

Where much of the western world is seeing cases peak, China has been seeing them rise again [1]. The idea that the virus has been eliminated is incorrect. China aggressively flattened its curve, but concluding that they will eliminate infections without herd immunity is not a certainty at this point.

1. https://www.npr.org/sections/coronavirus-live-updates/2020/0...

One of the comments made in the video (referenced in the article as well) is

"We always have between 37,000 and 60,000 deaths in the United States, every single year. No pandemic talk. No shelter in place. No shutting down businesses."

Since early Feb we've seen ~54K deaths from Coronavirus alone (excluding influenza deaths) [0]. The argument that this is just part of the normal flu cycle doesn't make any sense.

[0] https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/case....

The data regarding number of deaths is severely tainted though. Many doctors have claimed(this is also mentioned in the OP video) that their administration pressures them for including COVID-19 in causes of death.

But in reality, the number of excess deaths compared to the average year, or compared to last year, is well above the counted number of COVID-19 deaths. Anecdotal claims about deaths being miscounted as COVID-19 don’t really stand up to this.

https://www.nytimes.com/interactive/2020/04/27/upshot/corona...

https://twitter.com/lymanstoneky/status/1252616296891465729

The level of deaths/week we are seeing is like nothing that has occurred in decades of recorded vital statistics. (The 1968 flu pandemic shows on the 2nd chart, but is small beans compared to COVID).

For those that are paywalled, in New York the excess deaths of 6 times normal. You can't make that go away by claiming normal deaths are being misclassified.

There are likely deaths that are collateral damage. Fewer people are going to hospitals for heart attacks and strokes. It seems unlikely that we actually have fewer heart attacks, but people are scared of the hospital - so don't go and are more likely to die.

I am in the end the lockdown camp; but even if we allow businesses to open, many people will still be legitimately worried or scared and stay home even when it is something important like medical care.

> There are likely deaths that are collateral damage. Fewer people are going to hospitals for heart attacks and strokes.

And on the flip side, there are people who might have died of things like car accidents if the pandemic and subsequent lockdown had not taken place.

On the flip side I saw a short interview with a UCSF doctor that went to help out in New York where mentioned the hospital converted their cafeteria into a COVID19 ward. Not just in case but because it was full of patients on vents.

No amount of Hacker News Mental Gymnastics(tm) is going to explain that away.

Well, I also saw a medical worker saying that if someone dies before testing, that will be counted as a COVID death[1].

No amount of Hacker News Mental Gymnastics(tm) is going to explain that away.

[1]:https://twitter.com/BraveTheWorld/status/1254232498122428416

Considering mortality is 6x normal, that is a sane default under lack of resources for testing considering at least 80% probability.

> I also saw a medical worker saying that if someone dies before testing, that will be counted as a COVID death[1]

But this is what happens for flu too. You want to count them using the same methodology. You either use confirmed cases (Covid-19 kills more people than a bad flu year, and yes we do test for flu); or you use confirmed and suspected cases (again, covid-19 kills more than a bad flu year); or you use excess mortality (and this year we have huge increase in excess mortality).

There's no way to look at the data and come up with anything other than "covid-19 kills a lot of people".

Reply on news.ycombinator.com