RSNA: Mammo recall rate may be problematic pay-for-performance measure

CHICAGO—The Centers for Medicare & Medicaid Services (CMS) haS designated the screening mammography recall rate a performance measure for radiology. Results of a study presented at the annual meeting of the Radiological Society of North America (RSNA) suggested the practice site is one of the factors that can impact the recall rate, and the authors suggested practice setting be considered when the recall rate is used as a quality measure.

The target recall rate is less than 10 percent in the U.S., and balances sensitivity with the need for additional work-ups, said Ana P. Lourenco, MD, from the diagnostic imaging department of Alpert Medical School at Brown University in Providence, R.I.

Lourenco explained that multiple factors affect the recall rate, including patient population factors, such as age, family history and breast density; radiologist factors, such as years of practice and fellowship training; and system factors, such as batch reading, volume and use of computer-aided detection.

The researchers analyzed screening mammography audit data from May 2008 to September 2011 for a community practice and tertiary hospital. Both practices used digital mammography and read in a batch screening model.

The five radiologists who read at both sites interpreted a total of 73,297 screening mammograms during the study period (36,606 at the community-based site and 36,691 at the tertiary hospital).

Total patients recalled at both sites was 5,799 with an overall recall rate of 7.8 percent. The overall recall rate for the community site was 6.7 percent and for the hospital was 8.6 percent. Recall rates for each radiologist were lower at the community site than at the hospital, reported Lourenco et al.

The researchers attributed the higher recall rate at the tertiary hospital to patient factors, and explained that women who undergo screening mammography at the tertiary hospital site may be more likely to have a surgical, biopsy or cancer history. In addition, they may be less likely to be uninsured and thus prior mammograms may be unavailable.

Based on the finding that mammography recall rates can vary by site for individual readers, Lourenco and colleagues questioned the use of recall rates as a pay-for-performance measure.