Playing with numbers

Townsend Letter for Doctors and Patients,  May, 2006  by Ralph W. Moss

• E-mail
• Print
• Link

If you listened to the mainstream media in February, you might have thought that the end of cancer was finally in sight. The cancer death rates were down at last, we were told, for the first time in over 70 years. "It's a notable milestone," said Dr. Michael Thun, head of epidemiological research for the American Cancer Society (ACS). "That's momentous news," said Andrew C. von Eschenbach, MD, director of the US National Cancer Institute (NCI). "It proves that our expectation of continued progress against cancer is well founded."

When you look more closely at the statistics in question, however, you find that the advance was more symbolic than substantive. A review of US death certificates by the National Center for Health Statistics found that the number of cancer deaths had dropped to 556,902 in 2003. In the previous year, it had been 557,272. That represented a decline of 370 deaths. That's right, just 370 out of 557,272 cases, or around seven-hundredths of one percent (0.066), between 2002 and 2003. At that rate, cancer deaths in the US should be entirely eliminated by the year 3508, which is a little more than 1,500 years from now.

While deaths fell in men by 778, at the same time, they rose by 409 in women. Didn't hear about that, did you? In an alternate universe, the headlines might have read: "Cancer Deaths Among Women on the Rise!"

There was really not much new in this news. In fact, the death rates for several kinds of cancer have been falling slightly for about a decade, a fact that has been easily discernable in the annual graphs published by the ACS. However, apologists for conventional therapy jumped on the latest bandwagon and tried to attribute this minuscule change to improvements in diagnosis and treatment. I will clarify why I do not feel this is a likely explanation.

Changes in Therapy?

The most noticeable change in cancer statistics over the last few decades has been the decline in deaths from stomach cancer (gastric carcinoma). In 1930, the US death rate from gastric carcinoma in men was around 38 per 100,000. Today, it is around 6.9 per 100,000 for men and 3.4 per 100,000 for women. But this dramatic decline had little or nothing to do with improvements in diagnosis or treatment; indeed, outside radical surgery for early stage disease, there are still no effective treatments for most cases of stomach cancer. Adding chemoradiation after surgery will extend life only by an average of nine months (Macdonald 2001).

Experts are unsure of the reason for this demographic shift, but some attribute the falling stomach cancer death rates in the US to a variety of environmental causes, such as better food preservation, better dietary habits, and a downturn (associated with dietary improvements) in Helicobacter pylori infection.

The most precipitous death rate decline in recent years started around 1990, when lung cancer deaths among men began falling. The rate has continued to drop steadily ever since. Many commentators have pointed out that most of this decline can be attributed to a reduction in cigarette smoking. In 1965, 42% of all American adults age 18 years and older smoked; in 2003, only 22% smoked. Although 45 million Americans continue to smoke cigarettes, the decline in smoking has been reflected in a parallel drop, albeit delayed, in lung cancer mortality among women as well as men.

In other words, cancer incidence and deaths rates tend to change over time, and that change can be independent of any new treatments or diagnostic procedures. To suggest, without evidence, that the decline is due to improved medical care is simply unwarranted and smacks of self-promotion.

Prostate Cancer Rates

Prostate cancer deaths have also declined since 1990, but here the reasons are even less clear. I think it is premature to declare that this decline is due to more widespread screening or to more effective treatments. In fact, a recent large study concluded that screening men for elevated levels of prostate specific antigen (PSA) or utilizing the older technique of the digital rectal examination (DRE) does not actually save lives. John Concato, MD, and colleagues at Yale University, New Haven, in a study of 71,000 veterans, failed to show any benefit from either PSA or DRE used as screening method. The paper was published in the Archives of Internal Medicine (an AMA publication) on January 9, 2006.

According to other reports, "one of the most important reasons for the decline in deaths is a huge shift in how technology is helping medical professionals screen for cancer" (Seben 2006). But I repeat: changes in screening and diagnosis have little if anything to do with this decline in cancer deaths. In fact, as has been convincingly shown by William Black, MD, and Gilbert Welch, MD, of Dartmouth Medical School, Hanover, New Hampshire, what we are doing as a society is inflating the number of people diagnosed with cancer through the over-diagnosis of "pseudo-cancers"--that is, the detection of asymptomatic and often totally harmless precancerous lesions that in many cases would never progress to full-blown malignancies (Welch 2005).