Wednesday, March 4, 2015

We need to know more about psychological harms of screening

Several years ago, a few colleagues and I performed a systematic evidence review to help update the U.S. Preventive Services Task Force's recommendations on screening for prostate cancer. One of our key questions asked about the harms associated with prostate cancer screening, other than the overdiagnosis (and resulting unnecessary treatment) of clinically insignificant tumors. Since routine prostate-specific antigen screening had been going on for nearly two decades by then, we expected to find plenty of studies measuring anxiety and other mental health changes caused by false positive or indeterminate screening results.

In fact, after sifting through more than four hundred citations, we only found four articles describing three studies. Only one of the studies followed men for as long as one year. Here's what we wrote about that study:

[The authors] compared 167 men who had an abnormal screening result but a benign biopsy specimen with 233 men who had a normal PSA level. After 6 weeks, 49% of men in the biopsy group reported thinking about prostate cancer “a lot” or “some of the time,” compared with 18% of the control group. In addition, 40% of the biopsy group worried “a lot” or “some of the time” about developing prostate cancer compared with 8% of the control group. ... Statistically significant differences between the biopsy and control groups in anxiety related to prostate cancer and perceived prostate cancer risk persisted 6 months and 1 year later.

One might think that men with normal biopsies following an elevated PSA level should have been reassured that they had dodged a bullet and been pronounced prostate cancer-free. In fact, exactly the opposite occurred. And that's hardly surprising, since prostate biopsies, unlike breast biopsies, usually don't aim for a particular location of concern, leaving open the worrisome possibility that the biopsy needle just didn't sample the cancer if it was there. (How uncommonly cancer cells found in the prostate spread and lead to symptoms or death is another issue entirely.)

In the seven years since that review was published, the USPSTF has recommended against PSA-based screening for prostate cancer and recommended for low-dose CT screening for lung cancer in selected patients. There has been a major shift in how scientists view cancer screening and more interest in studying previously undescribed harms. In a review of psychological harms of screening published in the Journal of General Internal Medicine, Dr. Jessica DeFrank and colleagues assessed the literature on the burden or frequency of psychological harm associated with screening for prostate cancer (42 studies), lung cancer (11 studies), osteoporosis (6 studies), abdominal aortic aneurysm (8 studies), and carotid artery stenosis (1 study). They observed that for most screening tests, there remain large gaps in the evidence about the magnitude and frequency of such harms in populations representative of those receiving the tests. (I hasten to add that neither the USPSTF nor any other legitimate medical organization recommends ultrasound screening for carotid artery stenosis.) 

Causing someone needless worry about cancer or another absent health condition can seem trivial compared to the prospect of saving a life. But increasing recognition of the limitations of screening for cancer and disease in general, and the nearly nonexistent effect of these tests on all-cause mortality, have altered the equation. If more than 96 percent of initially positive screens turn out to be false positives (as is the case for lung cancer screening), just how much anxiety and worry are we as a society willing to inflict to merely exchange one cause of death for another?