I believe that 1 in 10000 number is what you want it to be, the number of deaths per infection in that age group, not the number of deaths per population of tha

I believe that 1 in 10000 number is what you want it to be, the number of deaths per infection in that age group, not the number of deaths per population of that age group. 1 in 10000 infections leading to death gives us a presumably reasonably accurate likelihood of death rates if everyone in that age group were to get infected.