Outcomes Analysis Shows Variations in Breast Cancer Care

FORT LAUDERDALE, Fla—With the first results from its Oncology Outcomes Database, the National Comprehensive Cancer Network (NCCN) has “taken a step to move beyond the guideline development process into a process of connection,” William McGivney, PhD, CEO of the network, said at its Fourth Annual Conference. “That critical step is the implementation of guidelines through performance measurement.”

Dr. McGivney introduced Jane Weeks, MD, of Dana-Farber Cancer Institute, who is the principal investigator for the NCCN’s outcomes and data project.

The NCCN began the project with a breast cancer pilot study, involving five member centers, to get “proof of concept” and to develop and test techniques for data collection, aggregation, and analysis, Dr. Weeks said.

Patients eligible for the pilot are those with newly diagnosed breast cancer receiving some or all of their primary oncologic care in an NCCN institution. “We’re collecting sociodemographic data, clinical data, and outcomes data,” Dr. Weeks said. “Everybody in our database gets followed in perpetuity.”

Data collection began on July 1, 1997, and the database now includes 1,542 cases. Most patients presented with early-stage disease. “There were only six patients in the whole database we were unable to stage,” Dr. Weeks said. “This compares very favorably with typical tumor registry data and, I think, reflects the very aggressive training of our data collection staff at the sites, and the coordination to ensure its accuracy.”

The data show an “extremely large spread among institutions in the proportion of patients who are participating in clinical trials,” she said. For stage II patients, for example, it ranges from 9% to 53%. “This is important data to feed back to the institutions to give them some sense of how their trial participation rates compare with those of other similar institutions,” she commented.

The first patterns of care analysis of the dataset involved the use of breast-conserving surgery by stage. “We found that among women with stage I disease, 63% received breast-conserving surgery as opposed to mastectomy. The range is quite wide, indicating clear differences in patterns of care within the NCCN institutions,” she said.

The guidelines allow either type of surgery, reflecting the view that survival outcomes are comparable and that it is appropriate to tailor therapy to patients’ preferences. “Do these differences across institutions reflect patient preferences or physician preferences and local patterns of care?” she asked. “We’re interested in feeding this data back and making sure that all institutions are really relying on patients’ preferences to choose between these equally efficacious treatments.”

The guidelines recommend use of axillary lymph node surgery for full staging in all patients. In stage 1, the rate is 86%, and in stage 2, it is 96%. “So this is routinely being done in our institutions, although not absolutely in all patients,” she said.

A Provocative Finding

The data also showed a significant association between age and use of lymph node surgery. “Older women are less likely to get axillary procedures,” she said, “and this is provocative.” It means one of two things, she said: That physicians are using legitimate discretion in not performing these procedures in older women, “in which case that ought to be reflected in our guidelines,” or that physicians may be undertreating some elderly women.

“That’s a key difference,” Dr. Weeks said, “and I think it’s a nice example of how very quickly we’re in a position to provide feedback to the guideline writers and work with them to make sure that the guidelines really reflect optimal care.”

Among women treated with breast-conserving surgery, 89% received radiation therapy, again a measure of guideline adherence, with a range across institutions from 80% to 100%.

Within the mastectomy group overall, 12% of patients received postmastectomy radiation, ranging from 6% to 18%. “The critical question here would be whether the high-risk patients identified by the guidelines as needing radiation therapy after mastectomy were, in fact, the 12% that got it,” Dr. Weeks said, adding that such an analysis is planned for the future.

Data on adjuvant chemotherapy show different choices (doxorubicin or no doxorubicin) being made based on stage and whether the chemotherapy was administered within the NCCN or outside. “Many of us have long suspected that there might be more use of Adriamycin-containing regimens in the centers than in the communities, which are still a bit more reliant on CMF; it will be interesting to see how that evolves over time,” she said.

Dr. Weeks next presented some data looking at actual adherence to guidelines for adjuvant therapy by selectively looking at the groups of patients for whom there are specific recommendations that do not involve patient or physician preferences. Among stage I patients with small tumors, the guidelines specifically recommend no adjuvant therapy (or enrollment in a clinical trial). In this subset, 58 of 71 patients (82%) had care that was adherent with the guidelines.

A Good Number or a Bad Number?

“What do we make of that?” Dr. Weeks asked. “Is that a good number or a bad number? The simple answer is that we don’t know because nobody else is doing this. It is really critical to start collecting this sort of data so we can establish an appropriate level of adherence.”

The appropriate level, she continued, “almost certainly, is not 100%. There needs to be some allowance for patient preferences and special circumstances.” She added that the NCCN does not require physicians to record reasons for deviations from the guidelines.

In general, for a first-time look at adherence with no prior feedback, Dr. Weeks called the 82% compliance figure “remarkably high.” In other subsets, adherence to adjuvant therapy guidelines was even higher, ranging up to 93%.

To sum up, she said that it is “reassuring to know that we can get these data, look at it, analyze it, and present it back to our institutions.” The analyses to date show that most care is adherent to the guidelines, she said, “but there is some room for improvement, which suggests that there is a value in doing this. Within the leeway provided by the guidelines, institutions are making some different choices. The database will allow us to begin to look at how those choices translate into differences in outcomes.”

Dr. Weeks noted that the NCCN member institutions have been very interested in receiving feedback on their patterns of care. “They’re not frightened by these data, not threatened by them. They really want to see what they’re doing,” she said.

The NCCN plans to continue to expand the pilot program to include all NCCN institutions and has begun work on the next outcomes database, for non-Hodgkin’s lymphoma. Finally, as the database grows, the network hopes to establish new partnerships with entities likely to have an interest in the data, such as regulatory and accrediting bodies, the pharmaceutical and biotech industries, and insurers.