Cancerous Statistics

An op-ed in the Sunday L.A. Times (registration required), authored by Drs. Samuel S. Epstein and Quentin D. Young, argues for increasing government spending on cancer research, especially research focused on prevention rather than cures. In order to prove the current approach (which already involves extensive government spending) is inadequate, they note the following:

A recent government analysis of leading causes of mortality in the U.S. from 1973 to 1999 revealed that, although the percentage of the population dying from heart disease decreased by 21 percentage points during the period, cancer deaths increased by 30 percentage points.

These statistics do not even begin to show that cancer is a growing problem in the United States. The numbers provided would even be consistent with a decreasing cancer threat, so long as the threat of heart disease declined even faster. And as a matter of fact, that’s almost exactly what has happened.

If you download the SEER report cited by the authors (in pdf format) and look at Figure I-2 (page 42), you’ll see that cancer was responsible for 23.0% of deaths in 1999 versus 17.7% in 1973 -- a 30% increase in the percentage (yes, that’s a percentage of a percentage), as the authors indicate. But if you go to the very next page (Figure I-3), you’ll see that the number of cancer deaths per 100,000 actually went down by about 20% for people under the age of 65 during the 1973-99 period, while rising by just over 10% for people over 65. You’ll also notice that in both pictures, the number of deaths from heart disease per 100,000 fell dramatically over the same period. (Note on interpreting the graphs: “neoplasms” is another word for tumors -- in other words, cancer.)

Of course, old people die a lot more than young people, so the rate for the whole population could have increased. But if you look at Table I-2 (page 18), you’ll see that the U.S. cancer death rate rose by just a tenth of a percent annually from 1975 to 1999, which is not nearly as scary as the authors imply in the article. (The figures are age-adjusted to take into account changing demographics, so the small increase is presumably not attributable to the aging of the population. It might, however, be attributable to the fact that some of the elderly people who would have died of other illnesses died of cancer instead -- which again does not indicate an increase in the cancer threat.) It also turns out that cancer death rates have been falling since the early 1990s (new document, Table II-3, page 3).