Friday, November 20, 2015

Two perspectives on the PSA screening pendulum

Two research studies published earlier this week in JAMA presented compelling evidence that the 2012 U.S. Preventive Services Task Force recommendation statement that discouraged prostate specific antigen (PSA)-based screening for prostate cancer has had a significant impact on clinical practice. In one study, researchers from the American Cancer Society used data from the Surveillance, Epidemiology, and End Results registries to document an 18% relative decrease (from 37.8% to 30.8%) from 2010 to 2013 in the percentage of men age 50 years and older who reported PSA screening in the previous 12 months. In another study, a separate team of investigators found a similar decline in the prevalence of PSA screening reported in the National Health Interview Survey in men age 50 to 74 years.

For me, and for other proponents of the view that PSA screening is not effective in reducing mortality from prostate cancer and instead leads to substantial psychological and physical harms, this reversal in practice is good news. In an editorial published in the October 15th issue of American Family Physician, Dr. Vinay Prasad argued that family physicians who reduce or discontinue their use of the PSA test for screening are on solid ground:

When it comes to PSA screening, the pendulum has swung. Not only has our understanding of the benefits and harms shifted, as reflected by a continual change in guidelines away from testing, but the burden to justify screening has also swung. For decades, critics of PSA testing have shown the many unintended repercussions of the test, cautioning that our initial widespread adoption was not justified. Moving forward, it must be the proponents of screening who shoulder the burden of proof. Their task will be to show in a future randomized study whether any PSA screening algorithm can improve survival or quality of life compared with what is now the standard of care—no routine screening. Before primary care physicians consider reintroducing the PSA test, they must have proof that it improves outcomes.

In another editorial that accompanied the JAMA studies, Dr. David Penson also described the shift in practice as a "pendulum," but took the position that until a better screening test is developed, "the PSA test can be deployed more effectively (or strategically), maximizing benefit while minimizing harm." If PSA screening does in fact save some lives, Dr. Penson argued, then extending screening intervals and focusing on men who are more likely to develop "high-risk" prostate cancer could be a better approach than not screening at all:

Certainly, physicians have been overly aggressive in their approach to prostate cancer screening and treatment during the past 2 decades, but the pendulum may be swinging back the other way. It is time to accept that prostate cancer screening is not an “all-or-none” proposition and to accelerate development of personalized screening strategies that are tailored to a man’s individual risk and preferences. By doing this, it should be possible to reach some consensus around this vexing problem and ultimately help men by stopping the swinging pendulum somewhere in the middle.

As it turns out, I've met both Dr. Prasad and Dr. Penson. Both men are outstanding researchers and compassionate physicians. Each believes strongly that his position on PSA screening is correct, and it's entirely possible that both could be right to some degree. If you are a clinician who cares for older men, where do you stand on the pendulum of PSA screening? If you are a man over 50 or a loved one, which approach to this test would you prefer?


A slightly different version of this post first appeared on the AFP Community Blog.


  1. About the optimization side: I wonder how come no one researched screening among men with family history of prostate cancer.

  2. I am a 64 year old man and I have read carefully on the issue of PSA testing for some time. I resolved about 10 years ago to never undergo PSA testing. There have been doctors opposed to PSA testing for many years. It seems that the biggest group of doctors who have been for PSA testing are the urologists. I wonder why? $$$$$

  3. One thing I noticed about the Cumulative Hazard of Death from Prostate Cancer curves in the ERSPC results is that the screening group shows no indication at all of death caused by surgery which would occur early on. Should this show up in the curve or is it intentionally (and misleadingly) excluded?