Innovative programs cut through mammography's crisis

2002 09 30 15 13 07 706

Many radiologists and healthcare planners believe that screening mammography in the U.S. is in crisis, and that measures must be taken to maintain widespread access to screening and to maintain public confidence in mammographers and screening centers.

A recent New York Times series questioned the competence of many radiologists who read mammograms, suggesting that too many doctors read too few studies to maintain their critical edge. At the same time, many screening facilities are closing or giving up screening mammography because reimbursement is so low it doesn’t pay to continue.

Teaching facilities report that they have unfilled fellowships, and that few residents plan to specialize in mammography. Lawsuits are frequent, and malpractice insurance has become difficult to obtain. To many radiologists, mammography is simply a nuisance.

While the scenario might seem bleak, there are signs that better organization and increased use of new technology can revive mammography. The specialty seems to be more lucrative in Canada and Europe, and U.S. healthcare experts are taking a close look at how these countries manage.

Fewer, more efficient centers?

The Mammography Quality Standards Act (MQSA) was adopted by Congress in 1992 to guarantee access for all women to quality mammography to detect breast cancer at its earliest stages. Regulations were written to cover facilities and equipment, as well as technologist and radiologist competence.

Although there are about 9,500 MQSA-certified screening centers, more than 650 have closed their doors in the past year because they couldn’t break even. But reducing the numbers of screening centers may be a blessing: Fewer centers may result in a higher skill level at remaining facilities; tighter scheduling and higher throughput can help facilities do better financially.

Current Medicare reimbursement for screening mammography is $81.81, which includes both the professional fee ($35.48) and the technical component ($46.33). If computer-assisted detection (CAD) is used with screening, the government pays an additional $17.74. Private insurance generally follows Medicare’s example, although CAD reimbursement remains spotty.

The MQSA requires that radiologists read 960 mammograms every two years -- just 480 a year -- to maintain certification. Canada and the U.K. require a minimum of 2,500 a year, and studies have shown that radiologists’ error rates diminish in inverse proportion to the number of mammograms they read. Some breast specialists, like radiologists Dr. Ken Heilbrunn in Seattle and Dr. Linda Warren in Vancouver, BC, personally read more than 15,000 screening studies each year.

Europe's hub-and-spoke model

What has come to be known as the "Swedish model" is taking hold in the European Union and throughout most of Canada, and shows promise for the U.S. In Sweden, women have their screening tests at conveniently located sites or mobile units where the images are obtained by skilled technologists. Films are sent to a central site where they are interpreted by radiologists who are dedicated to reading mammograms. If the patient needs additional evaluation, she goes to a central diagnostic center.

Europeans are also leading the way in using full-field digital mammography (FFDM) and CAD. The main obstacle to adoption of these modalities is cost. While a film-screen mammography unit may cost from $80,000 to $150,000, FFDM machines run from $250,000 upwards to $500,000. Discounts may be available to medical centers for research studies, but that usually doesn’t help patients and radiologists in rural settings.

A European study, called the Soft-Copy Reading Environment (SCREEN) project, was recently completed. Its main objective was to develop a soft-copy reading system that would meet the demands of high-volume screening, such as is practiced in European screening mammography programs and in the larger breast care centers in the U.S. One of its principal investigators, Dr. Carl Evertsz in Bremen, Germany, said his team worked with R2 Technology of Sunnyvale, CA, to adapt CAD technology to the SCREEN project.

SCREEN’s main goal was the creation of an efficient workflow where digital mammograms (including priors) can be read without loss of quality and with the use of CAD. Researchers developed and evaluated a workstation that can be used in digital screening mammography. The European Union is funding a follow-up project to test this system with demonstration sites in four EU countries.

Evertsz said high volume makes FFDM economical. He said that with screening units that process at least 50 cases a day, cost is not an obstacle, because more-efficient workflow and elimination of film results in significant savings over time. Reading can be centralized, reducing the number of reviewing workstations, and archiving can also be centralized. His group concluded that screening should not be done in low-volume clinics for economic reasons, as well as due to the difficulty in maintaining skills in such environments.

Making mammography pay

Seattle’s Heilbrunn makes mammography pay, even though he doesn’t use FFDM or CAD. His company, Imaging Associates, has no fixed sites, but runs three mobile units and visits 40 clinics in western Washington. Heilbrunn reads nearly 18,000 mammograms a year, and is the only radiologist on staff. He believes CAD is too slow and costly, and that his false-negative rate is so low that CAD probably wouldn’t be advantageous.

Heilbrunn recently bought three new film-based machines for his practice. At the time there were no mobile digital machines on the market, although GE Medical Systems of Waukesha, WI, has just released a mobile version of its Senographe 2000D FFDM system. Heilbrunn noted that the recent trend toward consolidation of reading centers across great distances have made him reconsider his aversion to digital.

Ninety-five percent of the 70 screening units in the Netherlands are mobile. Interpretation is done in about 25 reading centers. In Europe overall, mobile units make up around 60% of the installed base. SCREEN researcher Dr. Nico Karssemeijer of the University Medical Center in Nijmegen in the Netherlands said that in order to reach women in sparsely populated areas, mobile units and satellite communication are a must.

Digital images, of course, can be sent electronically from remote to central locations in the U.S as well, sometimes across state lines. However, a radiologist who reads images taken in another state must be licensed in that state.

Another barrier to FFDM in the U.S. is variation in insurance reimbursement. With Medicare and the majority of private insurers paying at the same rate as film-screen mammography, it is often impractical to go digital.

Radiologist Dr. Tim Freer in Plano, TX, is one entrepreneur whose group, Women’s Diagnostic of Texas, manages to stay in the black and offer CAD to his patients through what he calls "a classic hub-and-spoke model." His office reads 85,000 mammograms a year. Although Freer and his five fellowship-trained radiologists haven’t been able to afford digital hardware, they do make money by running relatively large patient volumes.

Freer’s group has a $20 surcharge for CAD, and if the patient’s insurance covers it, the surcharge is refunded. Eighty percent of his patients opt for the added security of this technology, he said. Freer considers himself a CAD evangelist, and he and a colleague last year published a key prospective study of CAD (Radiology, September 2001, Vol.220:3, pp. 781-786).

Another study just published also concluded that "computer-aided diagnosis can potentially help radiologists improve their diagnostic accuracy in the task of differentiating between benign and malignant masses seen on mammograms (Radiology, August 2002, Vol.224:2 pp.560-568).

There is not enough margin for many radiology practices to absorb the cost of CAD, which can be up to $150,000, but Freer offers it to patients as an add-on. "Our research has shown that it improves detection as much as 20%," he said, and his group hopes reimbursement reaches a level where they can offer CAD to all their patients. For now, Freer points out that patients choose different levels of care all the time, something that bothers many physicians.

Alternative modalities on the horizon

Other technologies are popping up as alternatives to x-ray mammograms. A recent study by Dr. Thomas Kolb and colleagues at Columbia-Presbyterian Medical Center in New York City combined conventional mammography with ultrasound for imaging women with dense breasts. The group found that the addition of screening ultrasound increased the detection sensitivity to 94% in those with the highest density (Radiology, October 2002, Vol.225:1, pp. 165-175).

Dr. Etta Pisano, chief of breast imaging at the University of North Carolina School of Medicine in Chapel Hill, and co-authors studied breast PET to determine whether this modality could aid in differential diagnoses of benign from malignant lesions. They concluded that there is not enough available data to answer the question (Academic Radiology, July 2002, Vol. 9:7, pp. 773-784).

Pisano told that she believes there will be a role for breast MRI. A small pilot study is underway by the Cancer Genetics Network and the National Breast MRI Consortium to screen high-risk women with both MRI and ultrasound.

Also being studied are technologies such as thermal imaging, laser imaging, and impedance imaging. None of these involve ionizing radiation.

Should women be advised to have breast screening with other modalities? Pisano argued against this practice: "If you did, you would have a biopsy every year. The more tests you have, the more chances you’re going to have a (false) positive test."

Battling lawsuits

Radiology has become one of the most-sued medical specialties in the U.S., and the delay in diagnosing breast cancer is the primary reason. A woman who had a screening mammogram and is later diagnosed with breast cancer is likely to sue for missed diagnosis. This trend is reflected in the availability and cost of malpractice insurance, and is one of the main reasons young radiologists avoid the field.

Some practices take extraordinary measures to limit malpractice exposure. In the second installment of the New York Times series on mammography, Dr. Stephen Kim of Kaiser Permanente Colorado in Denver reported on his facility’s rigorous quality-assurance program in screening mammography that sometimes resulted in staff radiologists being fired. Other physicians feel any proficiency monitoring beyond MQSA regulations violates professional peer-review standards (New York Times, June 28, 2002).

Freer’s group in Texas believes that enlisting technologies like CAD produces a significant improvement in early detection -- and thus decreases lawsuit vulnerability.

"We take a lot of measures to limit malpractice exposure," Freer said. "We have an extremely rigid patient follow-up system and an ongoing collection of data to measure our performance parameters. Rating radiologists is required under the MQSA regs, but we even go beyond that. We monitor recall rate, our positive predictive values for biopsies, and our cancer detection rate. This is important to us because we all read many thousands of mammograms each year."

One of the specialists in mammography performance is Dr. Edward Sickles at the University of California, San Francisco. He and his colleagues just published a study showing that radiologists who read mammograms exclusively have better scores than those who do so only part time. The study concluded that "specialist radiologists detect more cancers and more early-stage cancers, recommend more biopsies, and have lower recall rates than general radiologists" (Radiology, September 2002, Vol.224:3, pp. 861-869).

Ultimately, it may take federal legislation to protect radiologists from an avalanche of lawsuits over missed cancers in screening mammograms. Robert Smith, Ph.D., director of cancer screening for the American Cancer Society in Atlanta, advocates screening mammography of healthy women as a public health program, much like mass vaccination of school children.

In 1998, Congress established the national Vaccine Injury Compensation Program (VICP), a federal "no-fault" system designed to compensate individuals, or families of individuals, who have been injured by childhood vaccines administered in the public or private sector. The act requires parties to file under VICP before they can pursue civil litigation. A federal fund was set up to provide awards for justified complaints.

"In a sense, mammography is similar (to VICP)," Smith said. "It is the one radiological exam in which we screen people in good health to determine whether or not they have breast cancer." He said it would require new standards beyond MQSA, and that radiologists and screening centers would have to have high volumes and a demonstrated level of ability to find small cancers.

"There would have to be good-faith criteria that when errors occur, they are likely to be human errors," Smith said, "as opposed to errors where you were delivering a service even though you weren’t competent to do it."

By Robert Bruce contributing writer
October 17, 2002

Related Reading

Multiple-reading protocol bolsters interpretation for non-mammographers, September 6, 2002

Ultrasonography improves cancer screening of dense breasts, September 20, 2002

Radiologists slammed by malpractice insurance crisis, April 23, 2002

AMA delegates support higher payments for breast imaging, July 1, 2002

Mammographers question newspaper’s ‘crusade’ against breast imaging, June 28, 2002

New York Times takes mammographers to task, June 27, 2002

More cancer screening linked to earlier detection, May 9, 2002

Mammography reader volume directly linked to diagnostic accuracy, March 6, 2002

Copyright © 2002

Page 1 of 569
Next Page