Using cancer registries to assess the accuracy of primary liver or intrahepatic bile duct cancer as the underlying cause of death, 1999-2010.

Inaccuracies in primary liver cancer (ie, excluding intrahepatic bile duct [IHBD]) or IHBD cancer as the underlying cause of death on the death certificate vs the cancer site in a cancer registry should be considered in surveillance of mortality rates in the population. Concordance between cancer site on the death record (1999-2010) and diagnosis (1973-2010) in the database for 9 cancer registries of the Surveillance, Epidemiology, and End Results (SEER) Program was examined for decedents with only 1 cancer recorded. Overreporting of deaths coded to liver cancer (ie, lack of confirmation in SEER) was largely balanced by underreporting (ie, a cancer site other than liver cancer in SEER). For IHBD cancer, overreporting was much more frequent than underreporting. Using modified rates, based on the most accurate numerators available, had little impact on trends for liver cancer in the SEER population, which were similar to trends for the entire US population based on routine statistics. An increase in the death rate for IHBD cancer, however, was no longer evident after modification. The findings support the use of routine data on underlying cause of death for surveillance of trends in death rates for liver cancer but not for IHBD cancer. Additional population-based cancer registries could potentially be used for surveillance of recent and future trends in mortality rates from these cancers.