Since the U.S. Preventive Services Task Force
recommended against screening
for prostate cancer, the debate has been furious. In fact, screening rates are likely to remain high, because most urologists disagree
with the recommendations. One argued, "If you don’t do it, it’s negligent." (The debate is not new. Previously
This type of conflict between researchers and practitioners is common. A study of primary care physicians found that a quarter would recommend
[PDF] colorectal cancer screening for a terminally ill 80-year old patient.
Why? Fear of lawsuits contributes: 90 percent of physicians "reported
believing that physicians order more tests and procedures than needed to protect themselves from malpractice suits."
But another reason might be that many doctors just don't understand the numbers. A recent study asked Do Physicians Understand Cancer Screening Statistics?
. Too frequently, the answer is no. In fact, physicians' decisions about whether to order screening tests was more influenced by "irrelevant" than by "relevant" statistics.
This finding is consistent with previous research that too many doctors just aren't very good at the relevant math. As one review
[PDF] concluded in 2008, "Unfortunately, for more than 25 years, the skills of physicians in this area have been shown to be poor." For stat geeks who want to feel superior to (and worried about) doctors' statistical knowledge, some gory details are in Table 3 of this article
. For example, only 37 percent of residents "could interpret an adjusted odds ratio from a multivariate regression analysis."