Lies, damned lies, and statistics - Part I

In a National Cancer Institute report dated 1998, kidney cancer accounted for approximately 28,000 of new cancer diagnoses per year. By 2003, that number had risen to approximately 32,000. Now, in 2006, the estimated number of kidney cancer diagnoses is 38,890. [You can find a dizzying array of statistics at the NCI SEER site] In 1998, kidney cancer was considered a "man's disease" with a 2:1 occurrence rate. Now, the incidence among women is rising, with estimated new cases for 2006 at a 1.73:1 ratio.

Is the incidence of kidney cancer really climbing, or are diagnostic techniques simply getting better? Are more women developing the disease, or is the medical community simply more aware of kidney cancer and more willing to consider it as a possible diagnosis? What can statistics really tell us?

Statistics, like the Bible, can be used to argue or justify nearly any position. An excellent review of the impact and meaning of statistics in a cancer diagnosis can be found at Steve Dunn's Cancerguide. Evolutionary biologist Stephen Jay Gould wrote a fabulous article on statistics, "The Median Isn't The Message", that both demystifies the statistical process and provides a stubborn message of hope.

Well, I'm stubborn, and I'm still here 3 years after the second diagnosis. I haven't bothered to look at the statistics on that yet. I'm too busy knitting to bother (finished a pair of socks yesterday; 4 pairs to go) ...