Cancer risk and Life Expectancy

Bluestem Associates (bluestem@webserf.net)
Sat, 18 Dec 1999 03:04:21

On Fri, 17 Dec 1999 13:22:00 -0500, Hugh Lovel wrote:

>Are people living longer? I'm not so sure of this either as I haven't heard
>that life expectancy figures have been changing significantly in the US in
>recent decades,

Life expectancy is an oft-misused statistic. Reminder: life-expectancy
is simply the *average* age of all people who die in a given year. High
infant mortality significantly lowers life expectancy numbers.

By themselves, life-expectancy numbers are almost meaningless as
anything but an indirect indicator of infant mortality rates, which are
available directly in any case.

Cancer people, BTW, play a version of this game with their so-called
5-year survival rates. The statement "Early detection saves lives" is
only partly true because it is based on 5-year survival. If a given
cancer will kill you in 8 years and is discovered at year 4, you count
as a 'death.' If early detection discovers it at year 2, you now
survive at least 5 years and count as a survivor. Equally dead in year
8, but statistically in two different categories.

To Unsubscribe: Email majordomo@ces.ncsu.edu with the command
"unsubscribe sanet-mg". If you receive the digest format, use the command
"unsubscribe sanet-mg-digest".
To Subscribe to Digest: Email majordomo@ces.ncsu.edu with the command
"subscribe sanet-mg-digest".

All messages to sanet-mg are archived at:
http://www.sare.org/htdocs/hypermail