Mammography Performance Benchmarks in an Era of Value-based Care

Editor

In the April 2017 issue of Radiology, Dr Lehman and colleagues (1) and Dr Sprague and colleagues (2) reported on updated performance benchmarks for full-field digital mammography with use of data from the Breast Cancer Surveillance Consortium (BCSC). Since 1996, the BCSC has collected breast imaging data and outcomes from a population-based sample of clinical practices in the United States.

In an accompanying editorial, Drs D’Orsi and Sickles (3) discuss benchmark data use for evaluation and improvement of mammography clinical practice. They note that BCSC benchmarks, derived through linkages to tumor registries, are not achievable by many practices by using local biopsy outcomes alone. Instead, they advocate for comparison with performance metrics reported by the American College of Radiology’s National Mammography Database (NMD) (4). Because facilities submit data directly from their local practices to the NMD, the authors write that NMD benchmarks more closely reflect average mammography practices in the United States.

However, cancer ascertainment varies greatly among NMD facilities (4), contributing to a wide range of bias and variability in determination of both false-negative and true-positive cancer rates. This in turn influences estimates of cancer detection rates, sensitivity, and specificity. There is value in the NMD as a benchmarking resource, but currently NMD is missing essential data elements to measure performance, that is, high ascertainment of both screening-detected and interval cancers and their characteristics.

We all agree on the high value of complete cancer capture. We also challenge the assumption that this is not feasible for U.S. practices and recommend collaboration to find strategies for improving cancer capture, such as making linkages to state or regional tumor registries straightforward and financially supported or incentivized. Realizing this goal would provide important and otherwise unavailable information to radiologists to guide performance improvement efforts.

These and other practical approaches to better measure performance and outcomes will go further toward helping us improve the quality of our work than an “either-or” choice between alternative benchmark sources.

The breast imaging community has long been committed to auditing and feedback as means of providing the highest quality care to our patients. This ongoing discussion highlights an opportunity to further raise the quality of auditing and improve patient outcomes in an era of value-based care.