Is medical imaging easy pickings for reducing cancer costs?

Medical imaging represents “low-hanging fruit” in the battle to curtail rising cancer costs, according to Dr. Kevin Schulman, director of the Duke Center for Clinical and Genetic Economics at Duke University.

“We need to look at high-growth areas, and obviously imaging is one of those,” said Schulman, whose research documented that imaging expenses rose at twice the rate of other costs associated with the care of Medicare cancer patients from 1999 to 2006.

Exemplifying this trend are the myriad imaging studies typically ordered for lung cancer and lymphoma patients. Schulman and colleagues at Duke University documented that by 2006, the average patient with lung cancer was receiving 11 conventional radiographs, six CTs, a PET scan, a separate nuclear medicine test, an MRI, two echocardiograms, and an ultrasound-all within two years of diagnosis.

Similarly, by 2006, the average patient with lymphoma got eight conventional radiographs, six CTs, a PET scan, a nuclear medicine test, an MRI, three echocardiograms, and three ultrasounds, again within two years of diagnosis.

“It really gives you reason to pause and ask, ‘Are all these things essential? Are they all of value? Is the information really meaningful?” Schulman said. “There is a burden on the patient for going through all this, and there is an obvious a cost to the system.”

The Duke researchers began the study, which will be published in the April 28 issue of JAMA, partly on a hunch that a reduction in Medicare reimbursement for chemotherapy due to the Medicare Prescription Drug, Improvement, and Modernization Act would lead to increased use of imaging procedures after taking effect in 2003. Their research showed evidence of no such cost shift, but rather indicated a steady increase in the use and cost of imaging over the entire period.

The actual impact of this growth on the cost of caring for Medicare cancer patients, however, is open for debate. Despite the rising expense from 1999 to 2006, imaging at the end of this period still accounted for less than 6% of the Medicare costs associated with seven types of cancer. These seven accounted for 57% of all Medicare cancer costs in 2004, according to the researchers.

The Duke study is the first time the types and costs of imaging among Medicare beneficiaries with cancer have been specifically studied, according to the investigators. The analysis focused on a 5% sample of Medicare claims paid from 1999 to 2006. Included were nearly 101,000 cases of breast cancer, colorectal cancer, leukemia, lung cancer, non-Hodgkin’s lymphoma, and prostate cancer.

From these data, Schulman and colleagues singled out PET as an outlier, increasing at an average annual rate of 35.9% to 53.6%. Lung cancer and lymphoma patients showed the largest increase (53.6%) in PET use. This leap in PET imaging implies that the modality is an unreasonable contributor to cost. But the percentages are less damning if you consider that PET entered the clinical mainstream only in 2001 with the commercial introduction of PET/CT and substantially broader Medicare coverage for cancer indications. But Schulman views this recent adoption of PET as more, not less, reason for worry.

“With PET we are watching the adoption of a new technology,” he said. “It is an expensive technology. It does not seem to be replacing other technologies, and I don’t think it has reached saturation yet. You could easily envision a world in which we are using PET like we use CT and MRI today, with multiple numbers of scans per patient.”

But PET might also be skewing the overall data regarding imaging in the Duke study. Percentage gains from an artificially low level in 1999 to mainstream use in 2006 would naturally show a huge increase in costs. Mean annual increases from 1999 to 2006 for other advanced imaging modalities were much lower across the half-dozen–plus types of cancer: from 0.7% to 7.4% for ultrasound, 4.4% to 11.5% for MRI, 5% to 7.8% for echocardiography, and 6.3% to 20% for bone densitometry. Conventional radiography showed steady or decreasing use across the board.

And there are issues with two other modalities, namely echocardiography and bone densitometry. Neither is directly involved in cancer imaging, as echocardiography looks just at the heart and bone densitometry for bone loss is typically associated with advancing age. Schulman argues, however, that there are good reasons for including these in the Duke study.

“In regard to breast cancer, where you’re thinking about preventing fractures, bone densitometry could be supportive care and the same goes for echocardiography, as there is a lot of toxicity [to the heart] from different chemotherapeutic regimens,” he said.

But just as one can maintain that imaging costs might be inflated for Medicare patients, so can one argue that Medicare patients underrepresent the volume of imaging studies prescribed for younger cancer patients. The advanced average age of patients in the Duke study, 76 years old, may actually have limited the number of imaging studies prescribed for them, Schulman said. Younger patients may be receiving more imaging studies.

It is a question only further research can answer.