Monday, October 10, 2011

PSA and the Embattled US Preventive Services Task Force

The US Preventive Services Task Force is a teeny tiny little group of researchers, physicians and epidemiologists who can claim the privilege of issuing recommendations on a variety of health-related issues such as screening, counseling, and preventive medication use. They're meant to be independent of the Department of Health and Human Services so as to be as far from the taint of Washington politics as possible, but alas, they've had a habit of getting themselves into the spotlight in the past few years, most recently this past week with some new recommendations on the blood test that screens for prostate cancer known as the PSA (for "Prostate Specific Antigen").

I'm not blaming them for stirring the pot so much, mind you--the USPSTF's job is to evaluate the evidence for a given current health practice and decide whether that practice makes any sense. While that concept sounds simple in theory, it becomes exquisitely difficult to accomplish without wading into dangerous political waters in practice. It was just about two years ago that the USPSTF issued recommendations about mammography as a screening test for breast cancer: they advised that women between ages 50-74 should have mammograms every other year (unlike the then-current annual recommendation) and that women under 50 shouldn't have mammograms at all unless they were in a particularly high-risk group. This fairly understated document generated an enormous backlash (which I've described before here) and caught members of the Task Force by surprise.

But when you look at the numbers, the actual data that formed the basis of the recs, it's not hard to see that the Task Force was if anything being generous about mammography. I don't have the time to review all the data here but one stat may suffice. One typical mathematical model was used by the Task Force to estimate the number of lives saved versus the number of those who would go on to be diagnosed with possible breast cancer based on an erroneous read from a mammogram (these are known as "false positives"). In the model, if you annually screened 1000 women starting at age 40 and did so for 30 years, you would save eight lives. This came at the cost of one hundred fifty-eight false positive diagnoses, at least a group of which, presumably, would progress all the way to mastectomy and possibly even radiation or chemotherapy. If, however, you started the annual screen at age 50, you would save seven lives instead of eight, but you'd reduce the number of false positive mammograms from 158 to 95..."only" 95. (Again, the USPSTF advised against annual screens for women 50-74, and there are data that can be used showing a similar effect in the every-other-year scenario, but I thought these numbers were revealing.)

So when news came this past week of the new recommendations on the PSA screen, I wasn't completely surprised to learn that the panel--which, incidentally, is a different group of doctors than those who issued the mammography guidelines--advised against its use entirely. The evidence has been mounting for several years that PSA is a less than stellar test, and its interpretation can be especially slippery when the value of the test hovers just above the normal range. This leads to many false positive diagnoses with precisely the same problems found in the mammogram. Men with false positives sometimes undergo radical prostatectomy, a surgery that can leave one not only sexually debilitated but incontinent. The test works entirely differently than a mammogram but the principle of test interpretation and the problems of overdiagnosis remain the same.

Likewise I wasn't surprised to hear of a similar backlash against the Task Force and the exchanging of academic insult followed by counterinsult, or more heated comments outside the ivory tower walls. (One advocate for the PSA, the urologist Dr. James Mohler, described the chief medical officer of the American Cancer Society Dr. Otis Webb Brawley, a PSA skeptic, like this: "I have known Otis for over 20 years. He doesn't come off as being ignorant or stupid, but when it comes to prostate-cancer screening, he must not be as intelligent as he seems." That's about as close as one can get to saying, "hey, asshole, fuck you" in the subdued world of academia without actually doing so.) At the blog db's Medical Rants, a fairly innocuous post by db was met with at least one howl of indignation, with commenter Scott Orwig accusing db of being "irresponsible, unprofessional, and unethical". (db's follow-up post is here.)

Caveat emptor: I have not yet read the Task Force report so I don't want to take sides in this post. What I can say is that slogans impress me less than an explanation of complex data, and while the latter is less sexy and the former more emotionally comforting, it's usually an indicator of which argument is more likely to be right. In all the articles I've read so far, all I'm hearing from the advocates are slogans.