Hacker News Reader

Hacker News

Comment by kasperni | Hacker News Reader

Full thread

kasperni

> the real fatality rate could in fact be closer to 0.06%

If you take a look at Diamond Princess.

~ 4000 on board

* 712 cases

* 10 deaths

Gives a case fatality rate of 1.4%.

The average age onboard was ~60 years (don't know if it includes staff) which is definitely higher than average. However, the two numbers (1.4% and 0.06%) sounds very far apart.

View on HN

Another factor is viral load. More evidence is coming out that the total exposure can be a big factor in severity of symptoms and risk of death. This is likely partly why healthcare workers, including young ones, are often becoming grievously ill and sometimes dying.

In a confined space that everyone is stuck in for an extended period of time, people are continuously getting exposed and re-exposed to the virus, and those people are getting infected and continuously re-exposing everyone else. The ship eventually becomes a frothing virus stew.

Maybe one exposure ends up with a bit in your mouth from a nearby cough, and then another from your hand touching a surface and then wiping your nose or eyes with the back of your hand, etc., and then those keep happening over and over. Without proper ventilation, everyone will be getting exposed from breathing the same air in certain rooms, too. The frequent exposures give it more numbers that need to be fended off and more footholds to gain in different internal areas.

Then combine that with the high average age, lack of adequate medical care on the ship, general ignorance over the virus at that point, maybe re-use of unwashed clothes or hats.

Extrapolating from an environment like a ship or a hospital should be done carefully.

That's kind of right. The shear number of viruses that begin the infection, and also possibly over the first week or so, affect how hard it hits you. After the first week, your body is producing huge amounts of antibodies against the virus and it's not likely a factor anymore. Someone posted a link in another thread about a commercially available antibody test kit. The product page showed a nice graph of the timing of the antibody response (IgM shows up after the first week and IgG later). Now I can't find it. Anyone have that link?

Thanks, this makes sense. I imagine cruise passengers and hospital workers are pretty likely to get a lot of initial exposure in a short period of time, but people who get occasionally re-exposed through day-to-day life are probably at much less risk.

US population median age is 39. Fatality rate for people under 40 is <0.2%. Fatality rate of people aged 60+ is >3.6% [1]

So, no, 1.4% and 0.06% sound about right given the age difference.

[1] https://ourworldindata.org/coronavirus

It is impossible to calculate this from just the medians. John Ioannidis estimates a range of 0.025% to 0.625% by comparing the age distributions. However, there is so little data that any small systematic error could completely invalidate this.

Except that the US isn't composed of 327 million 39 year olds, it's composed of a fairly even mix of people of various ages. If you multiply the case fatality rates from your link [1] by the population in each age bracket [2], you come up with a projected CFR of about 1.8% in the US, assuming that a similar fraction of people in each age range are infected.

       Age  | US Pop |  CFR  | Est Deaths if
            |        |       | 100% infected
    --------+--------+-------+--------------
     0 -  9 | 40.01M |  0.0% | 0.00M
    10 - 19 | 41.97M |  0.2% | 0.08M
    20 - 29 | 45.43M |  0.2% | 0.09M
    30 - 39 | 43.63M |  0.2% | 0.09M
    40 - 49 | 40.46M |  0.4% | 0.16M
    50 - 59 | 42.83M |  1.3% | 0.56M
    60 - 69 | 37.41M |  3.6% | 1.35M
    70 - 79 | 22.66M |  8.0% | 1.81M
    80+     | 12.68M | 14.8% | 1.88M
    --------+--------+-------+--------------
      Total |327.08M |  1.8% | 6.02M

1.4% is believable with those numbers. 0.06% is not plausible at all in the context of those numbers.

[1] https://ourworldindata.org/coronavirus [2] https://www.statista.com/statistics/241488/population-of-the...

The passengers also saw themselves fit for travel and had full access to health care which points in the other direction.

However it bothers me that there are no solid numbers at this point that properly controls for age and prior conditions.

It really should be possible now and we shouldn’t be discussing whether this is real or overblown.

Now imagine a population that was averaged in the 30's years. What do you imagine the death rate to be then? it wouldn't be half, as it scales non linearly the younger the person gets.

As you say, the average age is high, which makes a big difference, but some people think there were more cases than the confirmed ones on the boat. That could explain the difference.

The age is a monumental factor in this, I don't think it's wise to just toss it aside like that.

Weren't the passengers strictly quarantined soon after the first cases occurred?

Exactly. It's naive to automatically assume all 4000 people were exposed.

Conditions like obesity and diabetes seem to increase risk dramatically. Not just plain age.

...and smoking is a big factor too.

If it can't really kill anyone under 60, then you have to divide by 5 (less than 20% of the population is > 60)

= .3%

If it can't really kill anyone under 70 or 80 (Seems kinda right) you might have to look at the demographics deeper.

You need more numbers.

It kills people under 60.

It will kill many more people under 60 where we run out of hospital beds and ventilators.

Reply on news.ycombinator.com