NCCN Takes First Steps in Creating Outcomes Database

March 1, 1997

FORT LAUDERDALE, Fla--Preliminary data on the treatment of breast and colorectal cancer patients at National Comprehensive Cancer Network (NCCN) member institutions show that the members are in compliance with the network's guidelines, Jane Weeks, MD, MSc, of the Dana-Farber Cancer Institute, said at the second annual NCCN conference.

FORT LAUDERDALE, Fla--Preliminary data on the treatment of breast andcolorectal cancer patients at National Comprehensive Cancer Network (NCCN)member institutions show that the members are in compliance with the network'sguidelines, Jane Weeks, MD, MSc, of the Dana-Farber Cancer Institute, saidat the second annual NCCN conference.

In addition to its cancer guidelines initiative (eight guidelines werepresented at last year's meeting and eight preliminary guidelines at thisyear's conference), the NCCN has taken the first steps toward the creationof a comprehensive, uniform, prospective outcomes database, Dr. Weeks said.

The data will be used to monitor adherence with guidelines at NCCN institutionsand will ultimately help define optimal strategies for managing commoncancers.

Dr. Weeks described early results of two ongoing projects of the NCCN'sOutcomes Committee. These projects are laying important groundwork forthe outcomes effort. First, a tumor registry project is using existingdata to look at patterns of care within the network's member institutionsand comparing them to national benchmark data. Second, a pilot study isunderway that will prospectively collect data on key quality-based outcomeindicators in breast cancer patients at five NCCN institutions.

Care Surpasses National Norms

The tumor registry project collected data on the treatment of breastand colorectal cancer patients submitted by nine member institutions during1993, 1994, and 1995 to the National Cancer Data Bank (NCDB), a nationaltumor registry compiled by the American College of Surgeons. These datawere then compared with national data published by the NCDB.

"We were gratified to see that the care being delivered in NCCNinstitutions, at least in the areas we looked at, appeared to be closeto the levels that we would expect," Dr. Weeks told Oncology NewsInternational. Also encouraging was the fact that the patterns of carein those institutions were in compliance with the NCCN guidelines.

Specifically, these very preliminary data showed that a greater percentageof patients with stage II breast cancer treated at NCCN institutions receivedadjuvant chemotherapy, compared with the national data (90% versus 58%).Similarly, the rate of breast-conserving surgery in patients with stageI or II breast cancer was about 10% higher in NCCN institutions than inthe nation as a whole (52% versus 42%).

Likewise, the use of adjuvant chemotherapy in patients with stage IIor III colon cancer occurred at a much higher rate at NCCN institutions,compared with national benchmark data.

The analysis further showed that, at NCCN institutions, the majorityof patients with rectal cancer underwent sphincter-sparing surgery. Nonational data are available for comparison.

There are several drawbacks of the tumor registry project, Dr. Weekspointed out: The NCDB data are incomplete and are not collected uniformlyeverywhere. Also, no information on such factors as comorbidity is included.

Breast Cancer Pilot Study

The prospective outcomes database that the NCCN is poised to amass willnot be subject to these drawbacks, she said. Ultimately, this project willprovide data on a wide variety of outcomes, such as response, recurrence,survival, hospital days, employment status, days lost from work, and performancestatus.

The pilot study involves five NCCN institutions: City of Hope, Dana-Farber,Fox Chase, M.D. Anderson, and Roswell Park. Together, these centers expectto provide data on approximately 4,000 to 5,000 breast cancer cases peryear.

Before accumulating those data, the NCCN institutions had to agree onwhich data to collect. The result was the creation of the data dictionary--avoluminous document that specifies exactly which data to collect and howto define every variable. "The stakes are high," Dr. Weeks said,"because once an undertaking of this magnitude is begun and data arecollected, it is exceedingly difficult to change the system."

The second task was to pilot test strategies for data collection andanalysis. The NCCN encouraged the five institutions involved in the pilotstudy to use whatever data collection strategy was already in place attheir centers. However, the key was to make sure that those existing datacollection mechanisms could be modified in such a way as to lend themselvesto the contribution of data to a single database, she said.

The third and final phase, which is slated for completion in July, isthe actual collection of data on all breast cancer cases in the five centers.In the future, similar outcomes data will be accumulated for other commonsolid tumors.