ACS projects > 33,000 prostate cancer deaths in US in 2011

The annual cancer statistics report from the American Cancer Society (Cancer Facts & Figures 2011) has just been released. The report is projecting 240,890 new cases of prostate cancer diagnosis and 33,720 prostate cancer-specific deaths in 2011. These numbers are up, yet again, from the projections for 2010.

While we recognize that the initiation of the retirement of the “baby boom” generation is inevitably increasing the raw numbers of men of 65 and older who are at significant risk for a diagnosis of prostate cancer, we have to say that we are more than mildly surprised by this increase in the projected number of prostate cancer-specific deaths, when you compare this projection to the actual (not projected) number of deaths, which has shown a continuous decline from 1991 to 2007 (according to data on the SEER web site).

The “New” Prostate Cancer InfoLink is increasingly of the opinion that the details of the statistical analyses that are beloved by epidemiologists may not be helping us all to really understand whether we are significantly impacting risk for prostate cancer in general — let alone the risk for prostate cancer mortality for undiagnosed and newly diagnosed men. It appears that one can now find data to prove almost any point you want to make if you simply take the time to search for the supportive statistical methodology and assumptions.

Who are men at risk and patients diagnosed with this disorder supposed to believe?

The pro-screening lobby will use the new ACS data to argue that these data “prove” that screening is necessary. The anti-screening lobby will argue just as forcefully that they demonstrate that screening is having no impact at all on prostate cancer-specific mortality (which the ACS estimated at a low of 27,050 for 2007).

We don’t think the table below will help much, but what is does is provide a summary of the estimated numbers of new cases and deaths projected each year by the American Cancer Society since 2001 and compare these to the supposedly accurate “actual” incidence and mortality data calculated using the SEER data. (If you click on the table you can see a larger version.)

In looking at this table, it is important to understand that changes in the way the calculations have been carried out over time definitely impact the data offered by the American Cancer Society and the SEER calculations too. The SEER data are age-adjusted compared to the population data based on the US census for 2000. In all honesty, the only thing that is clear to us from these data is that the SEER data are showing a consistent decline in the age-adjusted mortality rate from prostate cancer per 100,000 men. All the other numbers seem to “bounce around” without showing any time-consistent trend. When one looks at these data in the cold light of day, it is hard to tell just how much difference all of the diagnosis and treatment has made on a population basis over the past decade.

4 Responses

  1. Any numbers using crude data are subject to dramatic errors. Men who died of prostate cancer in the past few years were most likely born in the 1930s, a period of dramatic decline in births, for obvious reasons. This is reflected a lifetime later in death statistics, and throughout their lives by all other crude data, if not corrected by statistical standards. The huge increase in births beginning in 1946 will soon show an increasing curve in EVERY end of life statistic. (You are too late to invest in funeral homes and cemeteries.)

    SEER uses a standardization model.

    I spent many years working in schools, and the realities of demographics can be crushing if not recognized, anticipated, and planned for.

  2. Absolutely true … which is why (to be of any practical use to anyone except an expert in epidemiology and statistics) someone needs to re-think how these data are packaged for general public consumption.

  3. Another moment of confusing data from the ACS. The let’s screen less mantra does not go well with the climbing casualty assessments they portray. …

  4. When it comes to clinical and funding policy the numbers that count are the rates (“real” or projected), not counts — and certainly not estimates of counts by an advocacy organization that believes its mission is furthered by inflating the problem. The incidence and mortality rates are both generally trending down — not up. The 2005 drop in incidence rate is almost certainly a statistical artefact.

    With the exception of the 2005 anomaly, the SEER incidence rates between 2003 and 2007 are remarkably unremarkable — flat, likely within the MOE except possibly the drop in incidence between 2003 and 2004 — particularly in the context of the culture of prostate cancer in the early 2000s when many clinicians still truly believed that nerve-sparing techniques might authentically lower the risk and intensity of side effects.

    The SEER mortality rates per 100k have been steadily but marginally declining, and are almost certainly within the MOE between any given year for the years 2003 and 2007. But generally, 5 fewer per 100,000 men who died, i.e., died of prostate cancer in 2007 compared to 2001.

    Regardless, epidemiological trends on chronic conditions are understandable only through long-range analysis that takes into account policy impacts, adoption of emerging technologies and protocols, and a general understanding of basic population demographics.

    As the boomer cohort starts/continues to develop disease and conditions related to aging and the overall population of the US continues to grow, the raw numbers of men diagnosed will also increase. However, our understanding of the effectiveness of screening protocols and standards of practice can only be understood relative to the proportional likelihood of prostate cancer diagnosis and outcomes in the overall population.

    Further, the fact that both the incidence and mortality are generally trending down indicates two very possible likelihoods: (1) concerns about over-diagnosis and over-treatment are valid, and (2) Mark Twain was right when he complained, “There are three kinds of lies: lies, damned lies and statistics.”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: