By Katherine Hobson

The government estimates that more than 15,000 women died last year from ovarian cancer.

But routine screening of women with no symptoms isn’t recommended by the United States Preventive Services Task Force, American Congress of Obstetricians and Gynecologists or other professional groups, even for women at high risk for the disease. That’s because the benefits of the available tests — a transvaginal ultrasound and a blood test that detects an antigen called CA-125 — haven’t been shown to outweigh their risks (such as complications from unnecessary surgery), or to reduce the number of deaths.

Despite that evidence, a new study finds that when presented with a scenario of a woman coming to her annual check-up, a significant percentage of physicians surveyed would screen for ovarian cancer.

Some 28.5% of the 1,088 primary-care docs? — OB/GYNs, family physicians and general internists — surveyed said they “sometimes” or “almost always” offered or ordered ovarian-cancer screening tests for low-risk women. when the vignette involved a woman at medium risk of the disease, that proportion jumped to 65.4% of physicians.

About a third of physicians reported believing that ultrasound or CA-125 blood testing is an effective screening test for ovarian cancer, the study found. It’s unclear why that is, and more research is needed to discover why, says Laura-Mae Baldwin, an author of the study and professor of family medicine at the University of Washington.

“We know that in medicine we have an enthusiasm for screening,” she says. “We want to find diseases before they can cause harm.”

The study also found that physicians were more likely to say they’d order screening for patients who requested it — even if the doctor herself didn’t believe screening was effective.

In that scenario, physicians may be trying to maintain a relationship with the patient, or they “may lack confidence in explaining why the test is more harmful than beneficial,” says Baldwin. she says there’s an important role for clinical tools that could accurately illustrate the level of risk faced by a particular woman and also illustrate the harms and benefits of the test.

The study is published in the Annals of Internal Medicine.

There are limitations to the research: doctors were reporting how they’d react to the scenarios presented to them, not discussing actual patients. and while there was a 62% response rate to the initial questionnaire sent to physicians, the results may not generalize to the doctors who didn’t respond, the authors write.