Objectives : Cancer survival information is available only in areas covered by cancer registration. Methods : Five-year breast, prostate, and colorectal cancer relative survival rates from SEER are regressed on socioeconomic, demographic, and health variables at the county level. These models are first validated by comparing the observed rates with projected rates for counties not used in the estimation process. Results : Education was the best indicator of longer cancer survival. Other important predictors of the geographical variability of survival varied by cancer site. Better survival was predicted for breast and prostate than for colorectal cancer.
Projecting SEER cancer survival rates to the US: an ecological regression approach
Surveillance, Epidemiology, and End Results Program
To reflect advances in research and oncology practice, approaches to cancer control are evolving from simply enumerating the development of cancers by organ sites in populations to include monitoring of cancer occurrence by histopathologic and molecular subtype, as defined by driver mutations and other alterations. SEER is an important population-based resource for understanding the implications of pathology diagnoses across demographic groups, geographic regions, and time, and provides unique insights into the practice of oncology in the U. S that are not attainable from other sources. It provides incidence, survival and mortality data for histopathologic cancer subtypes, and data by molecular subtyping is expanding.