Aetiology

You've probably seen this announcment trumpeted somewhere: "Cancer deaths fall for first time." I just wanted to post a very brief note on this. First, let's look at their numbers:

The number of cancer deaths dropped to 556,902 in 2003, down from 557,271 the year before, according to a recently completed review of U.S. death certificates by the National Center for Health Statistics.

It's the first annual decrease in total cancer deaths since 1930, when nationwide data began to be compiled.

Now, a decrease in cancer deaths is always a good thing, but talk about making a mountain out of a molehill. That's 369 cases out of well over half a million deaths. So let's look at some more numbers.

In 2003, there were a total of 2,448,288 deaths in the US. WIth 556,902 of them due to cancer, that's 22.7466% of the deaths caused by this disease. In 2002, there were 2,443,387 deaths, with 557,271 due to cancer--22.8073% of them due to cancer. So the difference is .0607%. With such a negligible number, what will happen if they find these deaths rose in 2004?

The CNN article at least noted this in the second-to-last paragraph:

With such a small drop in deaths, it's possible they will rise again when 2004 data is tabulated, said Jack Mandel, chairman of epidemiology at Emory University's Rollins School of Public Health.

Now, in fairness, I'll note that I'm not a cancer epidemiologist, and according to this, there is reason to think this trend is real:

But the American Cancer Society, which conducted the analysis, believes the downward trend is solid, and it is projecting a substantially larger decrease this year.

They also note that while this is the first time the overall number of deaths has gone down, the rate of cancer deaths has been declining for over a decade. So, I'll hope that this trend continues, but you won't see me jumping up and down over a decrease of .06%.

Cancer