Duke Research Advantage

by Kevin Schulman

One of the most obscure organizations in the federal government is the U.S. Preventive Services Task Force which began in 1984. While convened by the government, the group is comprised of outside experts in medicine, health services research, and epidemiology.1 The task force’s job is to review the evidence around primary care services and to make recommendations to physicians about the quality of the evidence to support specific recommendations for what services should or should not be included as part of routine clinical practice.

The group’s efforts are very disciplined and scripted. Specific services are assessed on the basis of the risks and benefits to ascertain the potential “net benefit” to patients, with recommendations graded A through D (a grade of I means there is insignificant evidence to make a recommendation). A grade of A indicates high probability there is substantial benefit, B indicates high probability of moderate benefit or moderate probability of substantial benefit, C is a recommendation against a service with at least moderate evidence that the benefit is small, D is a recommendation against a service with a moderate or high evidence that there is no benefit or that the harms outweigh the benefits. The probability estimates in these statements are based on the quality of the evidence in the clinical literature on these topics.2

This group recently performed a periodic update of its 2002 recommendations for breast cancer screening.3 The task force conducted an evidence review to assess all of the new clinical studies on breast cancer screening since its last report, and it commissioned a decision analysis of different screening strategies for breast cancer to assess the net benefit or harm for each strategy.

Findings included: 1) there is convincing evidence that mammography reduces breast cancer mortality in women age 50-74 (with the greatest benefit for age 60-69) with a relative risk of 0.85 (0.75-0.96); less convincing evidence for women age 40-49 (but with the same magnitude of benefit); and evidence of net benefit for women over age 75 is lacking. Evidence was also lacking for teaching breast self-examination, clinical breast self-examination in women of already have mammography screening, and for digital mammography or MRI compared to film mammography.4

In the task force’s decision analysis efforts, it examined the net benefit of mammography screening for women under different strategies. The benefit of mammography screening came from the evidence review, but there were also harms to be considered: 1) radiation exposure and the risk of future breast cancer, 2) false positive test results (1 in every 10 mammograms for women age 40-49 is a false positive), work-up of false positive tests (repeat imaging, biopsies), and overtreatment of “silent” tumors.5 This analysis concluded most of the benefit of annual screening could be retained with bi-annual screening while avoiding substantial false positive exams.

In the end, the controversy resulted from a “C” recommendation against screening women from ages 40-49 since the additional benefit of starting screening at age 40 is small compared to age 50, and that there are harms from screening at any age. The task force suggested decision making for this group be left to “individualized, informed decision making,” or letting women know about the benefits and risks, assessing if there are factors that would influence either assessment, and letting the patients make a choice.

The task force also identified research gaps that require further study, although it has no resources to commission the clinical trials that would be required (and a quick look at women’s health on the Institute of Medicine priority list for comparative effectiveness research did not reveal any studies of breast cancer screening).

So what is the moral of this story? The data the group provided is very clear and transparent for anyone who takes the time to read the material. However, it’s time to include some direct assessment of patient perceptions in assessing the risks and benefits of medical decisions (rather than a paternalistic assessment of physicians). This should be done systematically as part of the review process, or can be done on a case-by-case basis as a less preferred option (as was recommended here). Second, in an era when so much medical care is not beneficial, is there any reason to recommend against beneficial therapy (even if the magnitude of benefit is small and the “net benefit” when subtracting the harms is uncertain)? With these two reservations, we should acknowledge this is exemplary practice of “evidence-based” medicine. We need to gauge whether the public reaction is to the process of evidence synthesis, or to the final recommendations themselves. This determination will be critical to our ability to move forward in making the practice of medicine more transparent.

1http://www.ahrq.gov/clinic/uspstfix.htm

2http://www.ahrq.gov/clinic/uspstf/grades.htm

3U.S. Preventive Services Task Force. Screening for Breast Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med. November 17, 2009 151:716-726

4Heidi D. Nelson et. al., Screening for Breast Cancer: An Update for the U.S. Preventive Services Task Force Ann Intern Med November 17, 2009 151:727-737

5Mandelblatt JS et al for the Breast Cancer Working Group of the Cancer Intervention and Surveillance Modeling Network (CISNET) . Effects of Mammography Screening Under Different Screening Schedules: Model Estimates of Potential Benefits and Harms. Ann Intern Med November 17, 2009 151 : 738 – 747