Denise Cummins Ph.D.

Good Thinking

Mammograms and PSA Tests: What Your Doctor Needs to Tell You

Patients are suspicious of evidence-based medicine--and so are their doctors.

Posted May 24, 2012

"I'm going to predict that people will continue to be unscientific”, said Dr. Otis Brawley of the American Cancer Society to NPR reporters. This dour prediction was based on the highly publicized controversies surrounding breast cancer and prostate cancer screening. To the average person—and his or her doctor—this seems like a no-brainer: The more screening, the better, right? After all, early cancers are more treatable than advanced cancers. So why, then, are panels of medical experts advising against routine screening for these kinds of cancers?

The simple answer is this: Because the tests are not fool-proof, and the treatments carry non-trivial medical risks. Here is what you need to know:

Let’s take a look at breast cancer first. According to the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Cancer Statistics report (, one out of eight women born today will be diagnosed with cancer of the breast at some time during their lifetime. This is called the lifetime risk of developing cancer, and it is based on the cancer incidence rate, that is, the number of new breast cancers that occurred in the U.S. during a year. This is different from prevalence; prevalence is the number of people in a specific population that have a certain type of cancer at a specific point in time. So incidence tells you the estimated number of new cases of a cancer, while prevalence tells you the number of all cases.

The most important thing to keep in mind is that breast cancer is extremely rare. Only about 20 out of 1,000 women have breast cancer right now, meaning 980 out of 1,000 do NOT. But incidence and prevalence are strongly related to patient age. Take a look at the numbers yourself (numbers are rounded):

 Age        Prevalence             Incidence

30-39        2 out of 1,000        4 out of 1,000

40-49        9 out of 1,000      14 out of 1,000

50-59      22 out of 1,000      24 out of 1,000

60-69      38 out of 1,000      35 out of 1,000

So, you might be saying to yourself, even though the numbers are small, doesn’t it make sense to screen anyway? Better safe than sorry, right?

Well, here is the problem: Reading a mammogram isn’t like looking at a photograph. You can’t just look at mammograms and “see cancer”. The films must be interpreted by trained radiologists, and there is a non-trivial error rate associated with those interpretations. According to the 2009 Breast Cancer Consortium report, about 1 in 12 40-year-old women getting a mammogram for the first time may be told they have cancer when they don’t. The error rate for 50 year-old women is about one in eight women! So every time you have a mammogram, you run the risk of getting a false positive result. In fact, if you start having mammograms at age 40, you have almost a 50-50 chance of being told you have breast cancer when you don’t by the time you are 50! You will then not only suffer a good deal of needless anxiety, but you may undergo biopsies and other invasive tests and treatments that carry medical risks, such as infection. Because of this, the U.S. Preventive Services Task Force issued new guidelines in 2009 recommending that screening should begin at age 50 (rather than 40), and that women should have mammograms every two years rather than every year.

So let’s get back to the original question, the reason your doctor ordered the mammogram in the first place: Do you have cancer? This requires a calculation using Bayes rule. Here it is in easy format:

Imagine 1000 women at age 40 who are undergoing their first mammogram screening for breast cancer.

9 of them will have breast cancer and will get a positive mammogram.

178 of them will noy have cancer and will get a positive result anyway.

What are the chances that a woman who has a positive mammogram result actually has cancer?

That’s easy: 9 out of 178 (which is equivalent to 1 out of 20). A positive mammogram at age 40 means you have a 1 in 20 chance of actually having breast cancer, and 19 out of 20 chances that you don’t. Your doctor should take this into consideration, along with other important factors such as genes, smoking, whether you have had other types of cancer, and so on. So if you have a positive mammogram, do not panic.

The same kind of controversy has exploded over prostate cancer screening. The most prevalent test used is called the Prostate Specific Antigen test, or PSA. Both normal and cancerous glands produce PSA. As men age, both benign prostate conditions and prostate cancer become more common. As a result, interpreting a rise in PSA in terms of benign or cancerous conditions is not an easy task. Doctors used to begin PSA screening at age 40. But the probability that a 40-50 year old man has prostate cancer given that he has had a positive PSA test (PSA score of 4.1 or higher) is 1 in 20! On top of that, the results of a long-term study involving more than 75,000 found little difference in prostate cancer death rates between men who has PSA screening and those that did not. The U.S. Preventive Services Task Force examined all the evidence, and found little if any reduction in deaths from routine PSA screening. Instead, too many non-fatal tumors had been discovered, followed by aggressive treatment that yielded serious side effects. In fact, according to the Task Force, up to one-third will end up with urinary incontinence, impotence or bowel problems. Based on results like these, they issued guidelines recommending that routine PSA screening for prostate cancer cease.

So far, the vast majority of physicians have chosen to ignore mammogram and PSA guidelines, continuing to routinely order these cancer screens just as they did before. In fact, the American Urological Association declared itself "outraged" by the guidelines. And a recent investigation by Johns Hopkins researchers found that fewer than two percent said they would no longer order PSA tests. When asked why, the most frequent reasons given were that patients expected the screening to continue, it would take too much time to explain, and fears of malpractice suits. This essentially means that, in the doctor’s office, medical decisions are dictated more by popular opinion and fear of lawsuits than scientific evidence.

About the Author

Denise Dellarosa Cummins, Ph.D., is the author of Good Thinking, The Historical Foundations of Cognitive Science, and Evolution of Mind.

More Posts