Hacker News Reader

Comment by jpxw | Hacker News Reader

> “The way in which we code deaths in our country is very generous in the sense that all the people who die in hospitals with the coronavirus are deemed to be dying of the coronavirus.

How else could deaths be recorded? It would seem negligent to not record someone who died with coronavirus as dying of coronavirus.

View on HN

If 5000 people die a day in average before the virus and 6000 people die a day on average after the virus would you say that the death toll is 1000/day or would you test all 5000 and say that any who had the virus were included in the virus death toll?

Do we have these numbers to compare? How much has the overall death rate increased by?

Some countries might publish sooner, but the most comprehensive database (WHO Mortality Database) has about a three year delay (2016 as the most recent year for many countries).

There is an official European organisation trying to track mortality:

https://www.euromomo.eu/

So far nothing suspicious, but they are always a few weeks behind.

Also note that you can look at subpopulations, e.g. people in Italy of age >= 65

https://www.euromomo.eu/outputs/zscore_country65.html

I think for the purpose of reporting it should be 5000, but those who calculate death rate should have more details. It's just reporting everyone with the virus is the only common way we can use to compare between countries.

Coding for death is something that statisticians do all the time. They have established, well used, processes to assign death to a particular cause.

Covid-19 causes a range of severity of illness.

If someone is hospitalised for covid-19 they are very unwell -- this is because hospitals are overwhelmed by the number of patients and are only admitting those who are very ill, and also because hospitals want infectious people to stay at home out of hospital as long as possible to reduce the amount of infection being spread.

So, when someone with covid-19 dies in hospital it's almost certain that it's the covid-19 that killed them, even if they would have died later that year of something else.

It seems to mean that if someone with coronavirus goes on to have a fatal heart attack, they'd still be recorded as dying "of" coronavirus. (Silly example, of course.)

I'm not sure how much of a blanket rule that is, or to what extent it's masking true numbers, or how it compares with other countries.

That’s exactly how first death in Russia was recorded. That patient died from preexisting condition and her lungs were not affected by corona.

the first death in Romania from coronavirus happened today and was a patient with terminal cancer, taken off chemo weeks ago before infection.

it is still counted as a coronavirus death, but somehow it doesn't seem right...

What’s the most likely place to catch coronavirus? I would posit it’s probably hospitals. If an elderly person needs to go to a hospital for any reason these days, I would guess they get taken to one full of other elderly people with coronavirus. I wonder what effect this might have on the results.

In italy hospitals are right now divided in two parts. One is for covid positives only and the other for covid negatives only. The staff which works in one part does not work in the other and has no contact with any member of the other one. Even ambulances are separated and with independent staff.

Are they testing the negatives? If not, asymptomatic transmission will contaminate the negatives-ward.

This is the most important case for extensive testing.

There is a great podcast from an Italian doctor summarizing the best practices for hospitals dealing with COVID-19 https://www.stemlynsblog.org/covid-19-podcast-from-italy-wit...

The first point being "Divide your department into resp patients and non-resp patients."

But this leads to double (and n-) counting. And that way we don't measure the death rate, but rather the spread of infection in the general population. If that is the case I would expect the death rate nowhere at no point to be much higher than normal or during a heavy flu season. The next weeks will show if the numbers are actually increased.

> If that is the case I would expect the death rate nowhere at no point to be much higher than normal or during a heavy flu season.

In Bergamo, obituaries in the newspaper went from little more than 1 page per day to about 10 [1] (sorry, link in Italian), so yes, it is much higher than normal. And this was about ten days ago, now I guess it is even worse.

[1] https://www.youtube.com/watch?v=iQaQBqH0txs

Yeah I know, and I did some numbers on that particular town. At least for that, the numbers are actually massively higher now, as implied by the video. But even if you look at Lombardy, they are just somewhat higher. And the farther away you look, the more it's just noise. Which is expected of course. But whether this is all worth the consequences we're imposing on ourselves - I'm having doubts.

It depends on your reason for counting.

For a hospital that needs to handle the diseased person, recording the most contagious disease makes sense, and acting as relevantly.

For an actuary looking at deaths after the fact, trying to analyse the lethality of diseases, attributing all deaths to the most contagious disease is less sensible.

I think they're pointing out that some people who got afflicted with Corona virus cannot be completely attributed to the virus alone. There may be underlying conditions which contributed more that caused the demise of the person.

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

Not ‘may’ have underlying problems. No less than 99% have other health problems. Around half of the deaths had 3 underlying problems. But we still credit the virus? Not the fact many of these folks have serious health issues?

”As of March 17, 17 people under 50 had died from the disease. All of Italy’s victims under 40 have been males with serious existing medical conditions.”

I have an eyebrow raised.

What if they were dying anyway? Seems hard to decide how to count that.

I read a story about a guy in Italy, he lost both parents, only his father died in the hospital so his mother was not counted in statistics as having died of coronavirus.

How about reporting ”number of life years lost”?

A death of healthy new born would mean 80 years lost. Death of elderly person maybe 5. You could simple stats by comparing to general life expectancy and more detailed by considering the health condition of the person.

Or report "number of years of experience lost"? I'm disinclined to dismiss the lives of older folk as valueless, as so many are doing these days.

Do they code for covid19 even if someone got hit by a car and just dies in a covid19-treating hospital?

He has to be covid-19 positive to be recorded as such. However, yes, he could die of internal bleeding, but have covid and he will be recorded as a covid-19 death. The article states that 88% of the dead had other pre-morbidity, sometimes two or three.

I think the fact that the hospitals are overwhelmed is killing people a lot quicker than they would've taken.

Reply on news.ycombinator.com