[aub3bhat]

Merely looking at numbers without understanding the underlying statistical process that generates those numbers is not how you do risk estimation.

[mlindner]

Please illuminate me then. What is wrong with looking at pure causes of death?