4 ms·

There is no way the fatality rate of someone with my health profile or the health profile of the average HN user has a 2.5% chance to die from the flu this year

There is no way the fatality rate of someone with my health profile or the health profile of the average HN user has a 2.5% chance to die from the flu this year. That must be including sick and/or elderly.

saalweachter 2y ago
Most people don't die every year; 2.5% [well, 1.5% last year] of the people who die, die of the flu.
- DiggyJohnson 2y ago
  Thanks, apologies for the misunderstanding. I wonder what the number is for pneumonia.
- triceratops 2y ago
  > Most people don't die every year I don't think anyone dies every year. Just once. /s
- haswell 2y ago
  2.5% of deaths caused by flu is not the same as a 2.5% chance of dying of the flu, though.
  - marcosdumay 2y ago
    Unless you do something unusual, that's almost exactly what it means. (Almost exactly because new causes appear, and those chances won't stay constant during your life.)
    - PaulDavisThe1st 2y ago
      No. US population: 330M Deaths per year (approx): 3.46M (call it 1% of the population) 2.5% of deaths: 86,500 people Number of flu infections, per year: 9M .. 41M 2.5% chance of dying of flu: 225,000 .. 1,025,000 people Number of deaths due to flu, per year: 4,900 .. 51,000 Actual chance of dying of flu (2010-2022): 0.01% - 0.5% [0] [0] min ((4900/41000000), 4900/9000000)) .. max (51000/41000000, 51000/9000000)
      - marcosdumay 2y ago
        The OP didn't say "per year". You added that one gratuitously.
        PaulDavisThe1st 2y ago
        It makes no difference at all. But whatever set of statistics you use has to represent the same period. We could use daily deaths from flu, or decadal deaths from flu, or forthnightly deaths from flu .. as long as the total deaths we're comparing with cover the same period.
        DiggyJohnson 2y ago
        Thanks for this breakdown, I was too tired to protest but I completely agree with your interdiction.
        saalweachter 2y ago
        I mean, I began by saying: > the IFR of the seasonal flu is around 0.1% We're all going to die some day. Based on current causes of death, there's a 20% chance, give or take, that it will be heart disease. A 20% chance it will be cancer. A 2% chance one of us will die from kidney failure. And around a 1.5% chance it will be from the seasonal flu, which has an IFR of 0.1% or less. Anyway, bringing this back around to the Spanish flu, it's fascinating, to me, that if you were alive in 1918, there was around a 1% chance you would die, that year, of the Spanish flu, and around a 2% chance you would die, years or decades later, of a seasonal flu. Pandemics get you upfront, all at once, epidemics get us eventually, over time.
    - yumong 2y ago
      "x% chance dying of the flu" has two vastly different meanings. "If you die, then the cause of death is to 2.5% the flu" is one meaning. Likely close to correct. "With 2.5% you will die this year a death caused by the flu" is another. Unlikely to be correct.
- PaulDavisThe1st 2y ago
  This doesn't seem correct. Deaths per year are around 3.5M; deaths from flu vary between 5k and 50k.
  - saalweachter 2y ago
    TBH I'm not sure where I'm getting the 2.5% number from; it's the one that stuck in my memory from the last time I dug into this data. It's entirely possible that my memory is wrong or the data I was looking at represented an outlier in flu deaths. For the 2023 / 1.5%, I referred to the provisional vital statistics for 2023: https://wonder.cdc.gov/controller/datarequest/D176;jsessionid=46AC35A44FC3A6AF54FC1F82FE4B https://wonder.cdc.gov/controller/datarequest/D176;jsessioni... Looking at past years (https://www.cdc.gov/nchs/data/databriefs/db355_tables-508.pdf#page=2 https://www.cdc.gov/nchs/data/databriefs/db355_tables-508.pd...), I get more like 2%, even for bad flu years like 2018. So that 2.5% number is probably either completely misremembered or based on provisional data that has since been revised downward.
    - PaulDavisThe1st 2y ago
      Yeah, 1 - 1.5% seems inline with most data. That is: about 1.5% of all deaths are deaths from flu. However, deaths among those infected by flu are much, much smaller than this (because the number infected is much, much bigger than the total number of deaths)

saalweachter 2y ago

Most people don't die every year; 2.5% [well, 1.5% last year] of the people who die, die of the flu.

DiggyJohnson 2y ago
Thanks, apologies for the misunderstanding. I wonder what the number is for pneumonia.
triceratops 2y ago
> Most people don't die every year I don't think anyone dies every year. Just once. /s
haswell 2y ago
2.5% of deaths caused by flu is not the same as a 2.5% chance of dying of the flu, though.
- marcosdumay 2y ago
  Unless you do something unusual, that's almost exactly what it means. (Almost exactly because new causes appear, and those chances won't stay constant during your life.)
  - PaulDavisThe1st 2y ago
    No. US population: 330M Deaths per year (approx): 3.46M (call it 1% of the population) 2.5% of deaths: 86,500 people Number of flu infections, per year: 9M .. 41M 2.5% chance of dying of flu: 225,000 .. 1,025,000 people Number of deaths due to flu, per year: 4,900 .. 51,000 Actual chance of dying of flu (2010-2022): 0.01% - 0.5% [0] [0] min ((4900/41000000), 4900/9000000)) .. max (51000/41000000, 51000/9000000)
    - marcosdumay 2y ago
      The OP didn't say "per year". You added that one gratuitously.
      - PaulDavisThe1st 2y ago
        It makes no difference at all. But whatever set of statistics you use has to represent the same period. We could use daily deaths from flu, or decadal deaths from flu, or forthnightly deaths from flu .. as long as the total deaths we're comparing with cover the same period.
        DiggyJohnson 2y ago
        Thanks for this breakdown, I was too tired to protest but I completely agree with your interdiction.
        saalweachter 2y ago
        I mean, I began by saying: > the IFR of the seasonal flu is around 0.1% We're all going to die some day. Based on current causes of death, there's a 20% chance, give or take, that it will be heart disease. A 20% chance it will be cancer. A 2% chance one of us will die from kidney failure. And around a 1.5% chance it will be from the seasonal flu, which has an IFR of 0.1% or less. Anyway, bringing this back around to the Spanish flu, it's fascinating, to me, that if you were alive in 1918, there was around a 1% chance you would die, that year, of the Spanish flu, and around a 2% chance you would die, years or decades later, of a seasonal flu. Pandemics get you upfront, all at once, epidemics get us eventually, over time.
  - yumong 2y ago
    "x% chance dying of the flu" has two vastly different meanings. "If you die, then the cause of death is to 2.5% the flu" is one meaning. Likely close to correct. "With 2.5% you will die this year a death caused by the flu" is another. Unlikely to be correct.
PaulDavisThe1st 2y ago
This doesn't seem correct. Deaths per year are around 3.5M; deaths from flu vary between 5k and 50k.
- saalweachter 2y ago
  TBH I'm not sure where I'm getting the 2.5% number from; it's the one that stuck in my memory from the last time I dug into this data. It's entirely possible that my memory is wrong or the data I was looking at represented an outlier in flu deaths. For the 2023 / 1.5%, I referred to the provisional vital statistics for 2023: https://wonder.cdc.gov/controller/datarequest/D176;jsessionid=46AC35A44FC3A6AF54FC1F82FE4B https://wonder.cdc.gov/controller/datarequest/D176;jsessioni... Looking at past years (https://www.cdc.gov/nchs/data/databriefs/db355_tables-508.pdf#page=2 https://www.cdc.gov/nchs/data/databriefs/db355_tables-508.pd...), I get more like 2%, even for bad flu years like 2018. So that 2.5% number is probably either completely misremembered or based on provisional data that has since been revised downward.
  - PaulDavisThe1st 2y ago
    Yeah, 1 - 1.5% seems inline with most data. That is: about 1.5% of all deaths are deaths from flu. However, deaths among those infected by flu are much, much smaller than this (because the number infected is much, much bigger than the total number of deaths)