3 ms·

> "What % of the population will die of this disease?" stays the same That doesn’t sound right — you can’t know the population death rate definitely without kn

by setr 6y ago

> "What % of the population will die of this disease?" stays the same That doesn’t sound right — you can’t know the population death rate definitely without knowing the infection rate; If no one gets the disease in your sample population of 100, your death rate says 0. If 1 person is infected and 1 dies, your population death rate says 1%. If 80 are infected, and 10 dies, your population death says 10%. The “correct” number should be stable, but the population death rate should be changing as the disease spreads, approaching the correct number (but this is also true of liklihood of dying to COVID) Your population death rate would only be static if your sample population is 100% infected, which is the same as “How likely am I to die if I get COVID”

IgorPartola 6y ago
You are getting a bit carried away. You aren’t wrong, the number of people who already got sick and either recovered or died affects future infection rates. So do other things. But I’ll use the example I used in another comment here: if you have 7.5 billion people in the world, and a rare disease that requires surgery where 100 people per year get sick, and all 100 undergo the surgery, and 15 of them die due to post surgery infections, is the mortality rate of the surgery 15/100 or 15/7.5 billion? If the latter, why? As another example: lots of people have the tuberculosis infection which lies dormant in their lungs. TB infection is distinct from TB disease which is where you get symptoms. The former is only dangerous in that it can develop into the latter. The latter can kill you. How should you look at these numbers? Is it worthwhile to drop the number of people with TB disease and just divide deaths by TB infections or is it more informative to derive mortality as deaths / TB diseases while also using TB infections and the conversion rate from those to TB disease to calculate risk and transmissibility? I argue the latter is much more relevant to decision making.

by setr 6y ago

IgorPartola 6y ago

You are getting a bit carried away. You aren’t wrong, the number of people who already got sick and either recovered or died affects future infection rates. So do other things. But I’ll use the example I used in another comment here: if you have 7.5 billion people in the world, and a rare disease that requires surgery where 100 people per year get sick, and all 100 undergo the surgery, and 15 of them die due to post surgery infections, is the mortality rate of the surgery 15/100 or 15/7.5 billion? If the latter, why? As another example: lots of people have the tuberculosis infection which lies dormant in their lungs. TB infection is distinct from TB disease which is where you get symptoms. The former is only dangerous in that it can develop into the latter. The latter can kill you. How should you look at these numbers? Is it worthwhile to drop the number of people with TB disease and just divide deaths by TB infections or is it more informative to derive mortality as deaths / TB diseases while also using TB infections and the conversion rate from those to TB disease to calculate risk and transmissibility? I argue the latter is much more relevant to decision making.