Use of Computer-Aided Mammography Soared After Mandated Medicare Coverage

Roxanne Nelson

June 23, 2010

June 23, 2010 — The use of a largely unproven mammography screening device has risen since Medicare began covering its cost, according to a report published in the June 14 issue of the Archives of Internal Medicine.

The prevalence of computer-aided detection (CAD) rose from 5% in 2001, when Medicare first began coverage, to 27% in 2003, the most recent year for which data are available. Supplemental fees for CAD use, the authors note, cost Medicare an estimated $19.5 million in 2003.

"This illustrates how industry and government interact to determine the course of healthcare practice. It's not really guided by science," said lead author Joshua Fenton, MD, MPH, assistant professor of family and community medicine at the University of California, Davis.

"This is a case in which expensive technology gets widely adopted in clinical practice before it is proven effective," explained Dr. Fenton in a release.

The authors note that 10 years ago, Congress established national Medicare coverage for the use of digital mammography technologies, including CAD. At that time, there was promising, albeit limited, evidence to support the ability of CAD to produce better clinical outcomes than routine film mammography.

Since that time, systematic analyses continue to reveal "persistent uncertainty" about whether CAD has a clinically important impact on key breast cancer outcomes, the authors write. However, despite the lack of supportive evidence, "a heavily lobbied Congress mandated that Medicare pay a premium for use of CAD above the average national reimbursement for screening mammography," they note.

Previous Study Fails to Show Superiority

In fact, in a previous study that involved 222,135 women and a total of 429,345 mammograms, Dr. Fenton and colleagues reported that the use of CAD was associated with reduced accuracy in interpreting screening mammograms (N Engl J Med. 2007;356:1399-1409). Rather than improve the interpretation of mammograms, CAD use was associated with significantly higher false positive, recall, and biopsy rates, and "with significantly lower overall accuracy in screening mammography than was nonuse." The higher biopsy rate was also not clearly associated with better detection of invasive breast cancer.

Newer Technologies Not Necessarily Better

Newer technologies are not necessarily better than those currently in use, and healthcare providers and individuals should not presume that they are in the absence of actual data confirming that, says Karla Kerlikowske, MD, in an accompanying editorial.

"Healthcare providers should not adopt new technologies without first demanding scientific evidence beyond that required for [US Food and Drug Administration] approval," writes Dr. Kerlikowske, who is from the San Francisco Veterans Affairs Medical Center in California. "They need to ask how strong and consistent is the evidence for new technologies and whether the evidence shows an important clinical benefit, whether there are important harms, and whether the benefits outweigh the harms."

To be responsible advocates for high-quality medical care, she concludes, "our enthusiasm for new technologies should not replace strong, consistent evidence that the benefits of the new technology outweigh the harms in a clinically important way."

Prior to Medicare coverage, the use of CAD was limited in the United States. This was partially because of the high monetary outlay — a cost in excess of $100,000 — to install the system, explain the authors. In addition, many radiology practices were already finding that mammography was not a profitable venture; in fact, it was a "money-loser."

But the Congressional mandate abruptly changed this paradigm, and the use of CAD soared, they note in their paper. In fact, manufacturers made reimbursement a central theme when marketing CAD and targeted facilities performing mammography.

Rooted in Politics Rather Than Evidence

Using linked Surveillance, Epidemiology, and End Results (SEER)–Medicare data, the authors estimated the annual prevalence of CAD use during screening mammography among Medicare enrollees 68 to 89 years of age from 2001 to 2003. They identified 112,195 screening mammograms that were performed on 66,125 eligible Medicare enrollees.

However, CAD use was uneven across the United States. even while its overall prevalence increased. For example, its use ranged from a low of 5.1% in Atlanta, Georgia, to 40.8% in Seattle, Washington.

Although these findings suggest that Medicare coverage can be associated with rapid diffusion of preventive technologies, they note, the data do not provide any insight as to why the use of CAD was so much higher in some regions of the nation.

One hypothesis is that regional variations might be viewed as a form of discretionary care, in which a service of uncertain effectiveness was adopted by some, but not all, healthcare providers, they explain.

"In addition, this case study illustrates the potential for political and economic interests to trump science in the development of Medicare coverage policy," they write.

Congress recently empowered the Centers for Medicare and Medicaid Services to add Medicare coverage for preventive services under specific conditions, and if there is strong evidence supporting its use. But the authors note that Congress retains the authority to add benefits for preventive services, and Congressional decisions "may be rooted in politics rather than in science and evidence."

Because Medicare coverage can exert a substantial influence on the increased use of new technologies, rigorous studies are needed to demonstrate a technology's effectiveness and safety among older Americans in real-world practices, the authors conclude. "If widely disseminated, unproven services may provide little value at substantial cost to the Medicare program."

Dr. Fenton was supported by an American Cancer Society Mentored Research Scholars grant. His coauthors and the editorialist have disclosed no relevant financial relationships.

Arch Intern Med. 2010;170:987-989. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: