Analysis of Cancer Registry Data
- Costs for data access
- Staff time
- Data use agreement and other registry requirements
- Barriers to accessing data
- Wait time for accessing data
- Estimated incidence and burden of hereditary cancers
- Data for program planning
Analyzing cancer registry data can help states estimate the incidence and burden of cancers, including those that may increase risk of carrying a pathogenic mutation. State health departments can develop criteria to query their registry data using national recommendations, such as those outlined by the United States Preventive Services Task Force (USPSTF)external icon and the National Comprehensive Cancer Network (NCCN).external icon These guidelines provide personal and family history criteria that can help states estimate the number of newly diagnosed cancer patients who may be at increased risk for carrying a pathogenic genetic mutation and may have benefitted from genetic counseling and testing as part of their diagnostic process. Findings can be used in health education materials, for policy and systems change activities, or to tailor activities to the needs of each state.
The resources needed to analyze cancer registry data may depend on states’ policies for access to and analysis of data. Some cancer registries charge fees for data access and analysis, require data use agreements, or have other requirements (for example, the submission and approval of data analysis plans). Some have specific requirements for accessing, using, and presenting cancer registry data. For example, cancer registries may have policies based on level of geography, suppression of small cell sizes, stratification by key variables such as race or ethnicity, and presentation of results.
Michigan analyzed its state cancer registry data annually to monitor annual rates and trends of cancer incidence and deaths that may be associated with Hereditary Breast and Ovarian Cancer syndrome (HBOC) and Lynch syndrome (LS) based on NCCN guidelines. These analyses allowed Michigan to understand its state’s hereditary cancer burden, assess disparities, and measure the program’s impact over time. The state genomics program paid a nominal fee for data access and submitted an annual data use agreement. The approval process took 6 to 8 weeks. Epidemiologist and cancer registry staff time were necessary to clean and analyze data and disseminate results.
Each year Michigan analyzed the data, created a report pdf icon[PDF-1.1MB]external icon and shared it broadly with internal and external partners. Through presentations at scientific conferences and meetings, the program shared its findings with broader audiences. The data were also used for program planning. For example, findings were used to target educational activities to populations living in geographic areas that had the highest HBOC and LS burdens.
Colorado developed a SAS software program to analyze data to determine which patients met NCCN guidelinesexternal icon for referral to genetic counseling based on risk criteria for HBOC and LS (see criteria below). Required resources included staff time to consult with clinical experts, review NCCN guidelines, develop and run the program, and analyze the data. Staff reviewed the NCCN guidelines with clinical experts periodically to ensure the program was using the most up-to-date information. The state worked with genetic counselors, and genomics and cancer registry staff at the health department to complete this effort.
Criteria for HBOC
- Breast cancer ≤50
- Two breast cancer primaries
- Breast cancer ≤60 that is triple negative for EP/PR/HER2
- Male breast cancer
- Ovarian cancer at any age (epithelial non-mucinous)
- Metastatic prostate cancer (Summary Stage 4–7)
- Ashkenazi Jewish descent with breast, ovarian, or pancreatic cancer at any age
- Breast and pancreatic cancer at any age
Criteria for LS
- Colorectal cancer (CRC) or endometrial cancer <50
- CRC or endometrial cancer at any age that is MSI unstable or MMR gene deficient
- CRC or endometrial cancer with metachronous or synchronous LS associated cancer*
*CRC, endometrial, gastric, ovarian, pancreas, ureter and renal pelvis, brain (usually glioblastoma), small intestine, as well as sebaceous adenoma, sebaceous carcinoma and keratoacanthomas as seen in Muir-Torre Syndrome.
Since the Colorado Cancer Genomics Program is housed within the state cancer registry at the Colorado Department of Public Health and Environment, it did not have any challenges in accessing the data. Analyzing state registry data using the program helped inform Colorado’s activities and approaches to educating providers and creating systems change to improve access to genetic counseling and testing. Findings were included in presentations and educational resources, such as the number of breast, ovarian, colorectal, and endometrial cancers diagnosed each year and the number expected to be associated with an inherited mutation. Data were also shared with reporting health systems to encourage systems change in referral practices through feedback interventions.