How are cancer deaths defined?

“Cancer Facts & Figures” 2012 by the American Cancer Society has been published. It continues to naively perpetuate the fictions of cancer statistics, which are many.

If you go looking for cancer, you will find it more often than if you don’t.

First of all is the continuing promotion of mammography for early breast cancer detection, despite data that have not convincingly showed that it saves lives (Welch, 2004). This is not really surprising when radiologists who benefit financially from mammography do it, but it is unconscionable that the ACS has actively chosen to denigrate and ignore the science used to change the recommendations of the U.S. Preventative Services Task Force task force over two years ago, based on data that the risks of routine screening for women in their 40s probably outweigh the benefits:

Early detection: Mammography can often detect breast cancer at an early stage, when treatment is more effective and a cure is more likely. Numerous studies have shown that early detection saves lives and increases treatment options. Steady declines in breast cancer mortality among women since 1990 have been attributed to a combination of early detection and improvements in treatment.

There are layers of misconceptions in these sentences, which reflect dogma more than actual data. This is at best unbecoming for an organization which as one of the main sources of information for cancer patients should to be on the forefront of scientific and medical progress in cancer.

What is the risk of death from cancer?

One of the main misunderstandings about cancer rates and risk of death is that the trends in these can be objectively produced in order to support the idea that we are doing a better job of early detection and treatment:

The 5-year relative survival rate for all cancers diagnosed between 2001 and 2007 is 67%, up from 49% in 1975-1977. The improvement in survival reflects both progress in diagnosing certain cancers at an earlier stage and improvements in treatment.

However, anyone with any understanding of probability can see that as imaging detects smaller and smaller cancerous masses, many of which will never become life-threatening (or indeed present any symptoms at all), then a smaller and smaller percentage of cancer diagnoses will ultimately result in death, without treatment playing any role whatsoever. The fact is that data of cancer incidence from 2011 are simply not comparable to data from 1977, and what “progress in diagnosing” really means is we now have the ability to find lots of little cancers that we never would have known about.

For example, “63,300 new cases of in situ [noninvasive] breast cancer are expected to occur among women in 2012. Of these, approximately 85% will be ductal carcinoma in situ (DCIS).” The report does not mention that in many cases, DCIS will never become invasive. And yet, all the positive screens which result in risky and painful treatment for many women are always spun as a success story, because the cancers were “caught early.”

Even worse is that the term “cancer deaths” continues to be poorly defined; in fact, the ACS report doesn’t bother to define it at all. To be fair, this is probably because in general the reporting of exact cause of death is so poor, and it is likely nearly impossible for the ACS to gather accurate data. The problem is that they do not state this caveat explicitly, but instead give the impression that these are good, solid, real data.

Why is it difficult to define “cancer death”?

Cancer certainly kills many people; if it did not, we would not spend the money and time we do trying to treat it. But what often goes undiscussed is the degree to which medical treatment for cancer hastens death, or even worse, causes deaths that might have been avoided without treatment.

It should be noted that the way that cancer deaths are counted is both underestimated and overestimated, but it is not clear whether or not these biases cancel each other out. The underestimation of cancer deaths happens when patients receive a treatment (usually surgical) that causes a secondary complication resulting in death that is not immediate (Welch and Black 2002). Welch and Black argue that these deaths should be included in cancer death statistics, but often are not.

But this can also be looked at from another point of view. Many people who develop complications from chemotherapy or surgery and die from them quickly are classified as having died from cancer, even when, as in the above case, they died of treatment rather than disease. A recent example of this is the experience of Joe Paterno, who clearly died from the effects of chemotherapy sooner than he would have of his cancer itself.

It is certainly true that many of the people who die from cancer treatment would have died from the cancer had they not been treated. But what remains undiscussed is that there are also certainly people who die from treatment who would not have died from their cancer. Especially for common cancers such as early stage breast and prostate cancer, patients often go into treatment perfectly healthy, with no symptoms at all. (Although this can also be true when cancer has metastasized [has reached stage IV], the probability that one will die from stage IV cancer is obviously considerably higher than from stage I-II cancer.) Some of those patients will end up dying of surgical and chemotherapeutic complications. Some of those deaths will be classified as cancer deaths, which is not quite right, and some of them will be classified as death from another cause, which is not quite right either.

The worst kind of cancer-treatment death is from complications due to diagnosis and staging. For example, lung biopsies are notoriously risky, but only now are researchers attempting to quantify that risk (Wiener et. al, 2011). Although on the surface most of the risks are sublethal, the chance of a collapsed lung was found to be 15 percent. Treatment for a collapsed lung is an incredibly painful procedure that adds to the risk of infection, respiratory failure and/or death. Other types of biopsies carry risks that are routinely disregarded in our blinding quest to wage our endless war on cancer.

So, we are left with a hodge-podge in defining cancer deaths, and this is important given how statistics are used to lobby for more research funds, for insurance coverage, etc. It is bordering on ridiculous that a country with the resources of the U.S. has not made an effort to increase the accuracy of our data collection on our second-largest cause of death. We need to be identifying more precisely causes of death that occur during medical treatment, and categorizing them for what they really are when we talk about cancer statistics. This will not only give us better information about the true trajectory of cancer deaths, but more broadly will allow epidemiologists and other researchers to identify areas in which deaths could be prevented, not to mention money saved from the treatment of unnecessary complications. Resistance from doctors to this would seem inevitable, in that this would undoubtedly cause more doctors to have to take more responsibility for treatment deaths But just like weaning doctors off of profits from using their fancy machines, this it is one more necessary piece of the puzzle to counter the spiraling trajectory of health care costs which will inevitably result in fewer and fewer Americans being able to afford insurance at all. And some of those left behind really do die from their cancer, because they can’t afford any treatment at all, let alone the spurious kind.

References

Welch, HG, and Black, WC, 2002. Are deaths within 1 month of cancer-directed surgery attributed to cancer? J Natl Cancer Inst. 94(14):1066-70.

Welch, H.G., 2004. Should I Be Tested for Cancer?: Maybe Not and Here’s Why University of California Press, Berkeley, CA.

Wiener RS, Schwartz LM, Woloshin S, and Welch HG. 2011. Population-based risk for complications after transthoracic needle lung biopsy of a pulmonary nodule: an analysis of discharge records. Ann Intern Med. 155(3):137-44.

• Related Posts
• Bookmark
• Tags