I don't think this has to be incorrect. Or that your "unless the age distribution...." needs to be true either. Everyone eventually dies. When that happens, as

I don't think this has to be incorrect. Or that your "unless the age distribution...." needs to be true either. Everyone eventually dies. When that happens, as things currently stand, it will be a 2.5% chance the cause will be the flu. You might die young or old. Across those possibilities, it is 2.5% That can be consistent with the idea that if you were to die next year, it probably (i.e., much less than 2.5%) won't be the flu because you are a young person.