METHODS Background

(NAACCR) sought objective estimates of the completeness of cancer registration among member registries, to complement self-reported estimates of completeness. Ratios of incidence to mortality had been used previously to estimate the completeness of cancer registration across cancer types in a single state registry. The technique was applied1 to data from U.S. member registries of NAACCR and evaluated for its validity in estimating the relative completeness of case registration across registries. A simpler method, which does not require mortality rates, but relies exclusively on comparing local cancer incidence rates to SEER incidence rates (LI/SI), was also evaluated using data from both U.S. and Canadian member registries