Sunday 13 May 2012

Cancer screening part 2: PSA for prostate cancer

For the boys, prostate cancer screening is another case in point. Screening is easy (just a blood test - a PSA) but like all screening programs it is plagued by overdiagnosis, and the harms from treatment for this condition are also quite significant (high rates of incontinence and erectile dysfunction).

The difference between this and breast cancer screening is that there is less general acceptance, and more criticism from official bodies and major academic journals and popular journals, and also from the guy who invented the PSA test in the first place (here).

The large scale randomised trials (the best evidence) on this topic give conflicting results regarding the effectiveness, and again, they concentrate on disease-specific mortality (which always makes screening look better), rather than overall mortality.

For the best explanation of the relative risks and benefits of PSA screening, read this article from some excellent public health academics from Australia, who put the risks and harms into numbers, to make it more digestible. They conclude:

you have to screen 1408 men and treat an additional 48 men to prevent one prostate cancer death over 9 years. In other words, only 1 of those 48 men is going to benefit over the next 9 years; the other 47 … have undergone treatment for no benefit within this period.

So now you can weigh up the risks and benefits yourself, which is the way it should be. Me? I sleep very well at night not knowing my PSA.

Addit 13 Nov 2012: Good overview of history and current thinking for PSA screening here.
Addit 17 Dec 2012: A good link on why screening may not work here.


  1. One factor that is usually ignored is that the success rate from PSA-motivated treatment (1 in 48 or whatever) is never compared with the success rate that could be expected from giving prostatectomies to men selected in some other way or indeed to men in general. Because even if you gave prostatectomies to men in general you will reduce prostate cancer deaths simply because they don't have prostates anymore. This type of preventative strategy is generally not acceptable (even though it would save lives) so if the PSA-motivated prostatectomy strategy is no more successful then that shouldn't be acceptable either, even though it may save lives. Chris O'Neill

    1. Fascinating idea. I wonder if the success of PSA directed treatment would be better than a similar number of prostatectomies done at random. No one would do the study, but a great way of thinking about it. Thanks.


Note: only a member of this blog may post a comment.