FORT LAUDERDALE, Fla--The much-anticipated, prospective, comprehensive outcomes data base of the National Comprehensive Cancer Network (NCCN), a consortium of 16 US cancer centers, is now up and running, and includes data on almost 400 patients with newly diagnosed breast cancer collected at five NCCN sites over 3 months.
"We project this will grow easily to 12,000 cases over the next 3 years as we expand and add in all of the NCCN sites," Jane Weeks, MD, of the Dana-Farber Cancer Institute, told a large audience at the third annual NCCN conference.
Initiation of actual data collection represents the culmination of a 2-year study aimed at pilot testing the NCCNs data collection methods and systems. Five institutions were involved in that pilot study--City of Hope National Medical Center, Dana-Farber Cancer Institute, Fox Chase Cancer Center, M.D. Anderson Cancer Center, and Roswell Park Cancer Institute.
Three more institutions--the Arthur G. James Cancer Hospital and Research Institute at Ohio State University, Northwestern University/Lurie Comprehensive Cancer Center, and the University of Michigan Comprehensive Cancer Center--are poised to begin data collection soon, and the remaining eight NCCN sites will come "on-line" by fall 1998.
The ultimate goal of the NCCN outcomes data base project, Dr. Weeks said, is "to generate uniform and comprehensive sociodemographic, clinical, treatment, outcome, and cost data on all patients with the common cancers treated in the NCCN."
The data base will serve various purposes, Dr. Weeks said. "First and foremost, our goal is to establish the network capability to select, analyze, and report patterns of care and outcomes."
A second, related goal is the establishment of a true, data-based continuous quality improvement program. Dissemination of the collected data to the member institutions will allow them to assess their compliance with the NCCN clinical practice guidelines (which now cover the treatment of 90% of all adult cancers), as well as to benchmark their performance against the rest of the NCCN.
The data base also can be used to support clinical disease-oriented research, as well as methodologic studies.
Lastly, the data base will provide the NCCN with a vehicle for forging partnerships with various sectors of the health care arena. "We are really excited about the opportunity to establish partnerships with managed care organizations and other insurers to provide data in return for assurance of enhanced patient flow," Dr. Weeks said.
The NCCN also envisions the creation of potential partnerships with the pharmaceutical industry, wherein industry would provide general financial support for the outcomes data base project, as well as collaborate with the NCCN in conducting specific studies employing the data base.
Finally, partnerships with regulatory and accrediting bodies are a distinct possibility, Dr. Weeks said.
The NCCNs Breast Cancer Pilot Study began in 1996 with the creation of a data dictionary--a document that specifies exactly which data are to be collected and how each variable is defined.
After the data dictionary was tested for validity and modified, the next steps were to: (1) construct institutional data bases, as well as a central data base to serve as a repository for shared treatment and outcomes data; (2) create institutional and web-based forms for data collection; (3) develop an operations manual, which sets forth patient inclusion criteria and strict rules for data confidentiality; (4) create a web site; (5) train all those involved in the project; and (6) establish procedures for data submission.
NCCN institutions can submit data to the central data repository at the City of Hope in one of two ways, said Joyce Niland, PhD, director of the Department of Biostatistics at City of Hope.
If the required data have already been collected by the institution for some other purpose, a file containing those data can simply be transmitted to the City of Hope. Alternatively, she said, the institution can enter data directly by accessing the NCCNs web site.
Emphasizing that the breast cancer data collected thus far are "extraordinarily preliminary," Dr. Weeks gave the audience a "taste" for the types of analyses that the NCCN will be able to generate.
The data base will allow the NCCN to track the age, race, and ethnicity of breast cancer patients treated at member institutions and compare these data with published data on breast cancer patients registered in the National Cancer Data Bank (NCDB). Among other variables that will be tracked are type of health insurance, stage at diagnosis, and treatment status.
The ability to monitor treatment status sets the NCCN outcomes data base apart from clinical trial and tumor registries, Dr. Weeks emphasized. "This data base is modeled to capture the true disease trajectory of patients with breast cancer, following them through the courses of treatment and the stages that occur in the progression of the natural history of the disease."
Non-Hodgkins lymphoma is the next disease to be added to the outcomes data base, Dr. Weeks said. Plans also are in the works to expand eligibility criteria so as to encompass all of the patients with a specific cancer receiving their care in member institutions, not just newly diagnosed patients, as was done for the breast cancer study.
"We would also like to expand the data elements," Dr. Weeks said. "Were very interested in complications, patients self-reported quality of life, patient satisfaction, and cost." A cost model currently under development will, in the future, allow costs to be incorporated into the outcomes data base (see box on page 24).
Another future goal is to begin collecting data in selected affiliates of NCCN institutions, Dr. Weeks said. This will enable comparisons to be made between NCCN institutions and their community-based affiliates.