Author's caveat: "And I recognize just how far my envelope-math is from a real model of health outcome risk difference"