Hacker News Reader

Comment by nostromo | Hacker News Reader

The current fatality rate is not 3.7%. The New England Journal of Medicine suggests it may end up being much less than 1%.

https://www.nejm.org/doi/full/10.1056/NEJMe2002387

> it is reasonable to say that we value a human life at about $8 million

We're talking about the very elderly dying a bit earlier than they would have otherwise. Before you accuse me of being heartless, I agree that that is indeed sad and worth mourning. But, no, it is not the same economically as a working-age person or younger dying in large quantities.

And keep in mind it's March, and flu season will naturally be on the decline this month.

View on HN

3.7% is the measured rate in China. Which has the best available data set. Which is why I used it.

The CDC has consistently concluded that the real rate will be somewhere in the 2-4% range.

There are admittedly researchers who think that cases are severely underreported. However data from other countries has been in line with the Chinese experience. If dramatically more get it, those dramatically more seem to also be less infectious.

Given this, it is better to act on the higher end of evidence based estimates and to have possibly overestimated than it is to believe on the low end and find that a higher figure was appropriate.

But even if you disagree with that reasoning, dropping all of my figures by a factor of 4 still comes up with figures that justify truly shocking disruptions to our lives. And the basic facts about epidemiology that I gave are still worth understanding.

You are wrong, that is one of the worst data sets to use, because of so many unidentified cases.

Again, see the NEJM:

https://www.nejm.org/doi/full/10.1056/NEJMe2002387

Or any of the better data sets:

On the Princess Cruise, a naturally controlled study, the death rate was less than 1%, with a cruise full of old people.

Non-Wuhan death rate in China, where more mild cases were identified, ended up being 0.7%.

https://www.livescience.com/is-coronavirus-deadly.html

Goodness, there are so many chicken littles on HN nowadays.

I guess all I can do for this poor comment is vouch and upvote. Sad to see that fearmongering and bad data compounded by bad math is ruling the day.

On the Princess Cruise everyone received top notch care. What if there were no available hospital beds and respirators?

What would be the death rate when the medical system is so overwhelmed none of the acute cases can receive any medical care? That's the number that should influence how far we go in trying to control transmission.

In Israel the single serious case is 38 years old. He is on a respirator. No some 80+. IIRC from the Washington cases there were some in their 20s who are hospitalized. Israel is less likely to have random unknown cases given their stringent border control, quarantines and testing.

There are multiple conflicting reports of the death rate. However, if hospitals are overwhelmed by the serious cares, which I've seen estimated at 15-20% then the death rate, whatever it is, will surely rise as serious cases go unattended.

> [China] has the best available data set

No they don't. They have large biases from the fact that mild and asymptomatic cases are underrepresented. To avoid that you want to look at data sets where they tested 100% of people whether they had symptoms or not.

As far as I know that is the cruise ship and the South Korean church. And they show a lower death rate.

This is interesting. Do you by any chance still have the stats?

https://bnonews.com/index.php/2020/02/the-latest-coronavirus... worth keeping in mind that the numbers are so low that there may be some significant variance at play. But the serious number should be more robust.

Now Diamond Princess let's us estimate Serious / Total, and say Italy let's us estimate Death / Serious (they probably catch all serious cases, and have less variance since they have larger numbers). So it may be possible to combine to get a good estimate of Death / Total.

Possibly significant source of error is age skew.

This study estimated the case fatality rate (CFR, deaths among those presenting symptoms) and infection fatality rate (IFR, deaths among all who tested positive) of infected passengers on the Diamond Princess: https://cmmid.github.io/topics/covid19/severity/diamond_crui...

CFR was estimated at 2.3% and IFR at 1.2%, reflecting that roughly half of infections were asymptomatic. The downside is that the median age of passengers was 58, so it's not truly representative of CFR for all ages. (That's why people are still using data from China; the cruise ship isn't demographically representative.)

You also need to consider that 15% of cases are serious, and may require mechanical ventilation. If medical systems are overwhelmed, it will go up. We saw that in Wuhan.

That study has some non-standard numbers? More deaths (6 vs 7) and fewer cases (696 vs 634). Makes me suspicious.

The age of the South Korean church members was largely in 20s and 30s. The cruise ship crew was a large portion of total affected and few were over 65 where the big risk kicks in, and on top of that immune compromised people don’t go on cruises.

> We're talking about the very elderly dying a bit earlier than they would have otherwise.

The 62 year old still in the workforce, living with managed diabetes.

The 55 year old with asthma.

The active and healthy 72 year old who just retired.

The 32 year old with an autoimmune disorder.

The 43 year old healthcare worker with repeated exposure to the pathogen.

...

Saying it's sad and worth mourning doesn't mean you're not still heartless. People aren't numbers, and turning them into numbers is what makes you heartless for making an argument like this. This is a prime example of a "shit hackernews says" post.

They’re responding to a comment that’s explicitly making an economic argument based on the numbers. I think it’s reasonable to respond to that on its own terms.

While it is tempting to say that "people aren't numbers", a refusal to quantify human impacts just results in poor decision making.

Because we still do make decisions. We still do allocate resources to our own pleasure that could have gone to saving someone else's life. We pass regulations limiting how much people can be exposed to chemicals that impact health. And trade off the convenience of being able to travel in cars for how many of us will die from doing so.

We still make those decisions. The only question is whether we think clearly while making them, or whether we shy away from the ideas that could let us think clearly.

> This is a prime example of a "shit hackernews says" post.

I'd say that applies to your post more than OP. Was that really necessary?

In my family there are four grandparents in their 70's with various health problems, and my two young children. All four grandparents have commented that they're relieved that this virus appears to target old people more than children, and we'd all consider it to be at least less bad news if the numbers confirm that. My father in law, who is a doctor, remarked that he's less concerned about losing his own life because he's "already lived 90% of it".

I'd never want to have to decide whether my parents or children get to live, but if I did, I'd choose my children in an instant.

These are our opinions -- no more or less valid than your own. Do you suppose we, or anyone else for that matter, cares whether someone like yourself thinks we're "heartless"?

If actually making measured, informed decisions without discarding all knowledge of numbers makes me heartless then I readily accept.

Reply on news.ycombinator.com