There is a high level of agreement between SEER and Medicare reporting of radiation treatments after a cancer diagnosis, suggesting that either source can be used to assess radiation-related treatment patterns. However, for the most complete ascertainment of radiation use, both SEER and Medicare sources should be used.

Users of the linked SEER-Medicare database commonly perform analyses that focus on the complex interactions among patient characteristics, cancer treatments, and outcomes. The authors review the source and scope of the patient-specific data elements, with a focus on three domains--demographic characteristics, socioeconomic characteristics, and survival status. They offer some concrete recommendations regarding the use of these data elements. In particular, they describe analyses that provide an estimate of the accuracy of the sex and age variables, and raise some cautionary notes about race and ethnicity variables. The authors describe the available measures of socioeconomic status, and recommend, with some caveats, the use of median income measures as a proxy for socioeconomic status. Finally, they describe the available data on date of death, and explain why confidence in these measures is justified.

The agreement of SEER and Medicare data appears to be good for major surgical procedures and for excluding persons who did not undergo cancer-directed surgery. Both the SEER and the Medicare data captured a small number of surgeries not reported in the other file. Therefore, where possible, using both data sources will enhance identification of surgeries. Because the analysis was performed with linked data, the accuracy of surgical claims in Medicare data alone cannot be assessed.

The 21-gene Recurrence Score assay is validated to predict recurrence risk and chemotherapy benefit in hormone-receptor-positive (HR+) invasive breast cancer. To determine prospective breast-cancer-specific mortality (BCSM) outcomes by baseline Recurrence Score results and clinical covariates, the National Cancer Institute collaborated with Genomic Health and 14 population-based registries in the the Surveillance, Epidemiology, and End Results (SEER) Program to electronically supplement cancer surveillance data with Recurrence Score results. The prespecified primary analysis cohort was 40–84 years of age, and had node-negative, HR+, HER2-negative, nonmetastatic disease diagnosed between January 2004 and December 2011 in the entire SEER population, and Recurrence Score results (N=38,568). Unadjusted 5-year BCSM were 0.4% (n=21,023; 95% confidence interval (CI), 0.3–0.6%), 1.4% (n=14,494; 95% CI, 1.1–1.7%), and 4.4% (n=3,051; 95% CI, 3.4–5.6%) for Recurrence Score <18, 18–30, and ⩾31 groups, respectively (P<0.001). In multivariable analysis adjusted for age, tumor size, grade, and race, the Recurrence Score result predicted BCSM (P<0.001). Among patients with node-positive disease (micrometastases and up to three positive nodes; N=4,691), 5-year BCSM (unadjusted) was 1.0% (n=2,694; 95% CI, 0.5–2.0%), 2.3% (n=1,669; 95% CI, 1.3–4.1%), and 14.3% (n=328; 95% CI, 8.4–23.8%) for Recurrence Score <18, 18–30, ⩾31 groups, respectively (P<0.001). Five-year BCSM by Recurrence Score group are reported for important patient subgroups, including age, race, tumor size, grade, and socioeconomic status. This SEER study represents the largest report of prospective BCSM outcomes based on Recurrence Score results for patients with HR+, HER2-negative, node-negative, or node-positive breast cancer, including subgroups often under-represented in clinical trials.

Challenges persist in measuring screening use, and innovative approaches are required to distinguish screening tests from diagnostic and follow-up evaluations.