Can We Rely on GLOBOCAN and GBD Cancer Estimates? Case Study of Lung Cancer Incidence and Mortality Rates and Trends in Iran.

BACKGROUND Around half of input data in the global burden of disease cancer collaboration (GBDCC) and GLOBOCAN projects come from low quality sources, mainly from developing countries. This may lead to loss of precision in estimates. Our question was: Are the absolute values and trends of the GBDCC and GLOBOCAN estimates for lung cancer (LC) in Iran consistent with available statistics?. MATERIALS AND METHODS Incidence and mortality statistics were extracted from national reports (N.IRs and N.MRs) and GBDCC (GBDincidence and mortality) and GLOBOCAN databases for 19902013 where available. Trends were analyzed and absolute values and annual percentage changes (APCs) were estimated and compared. Incompleteness of case ascertainment at the Iranian national cancer registry and Iranian national civil registration was assessed for better understanding. RESULTS Trends of N.IRs were significantly rising for males (APC: 19.4; 95% CI: 12.526.7) and females (23.2; 16.030.8). Trends of GBDincidence were stable for males (0.2; 1.51.1) and females (1.0; 2.30.4). Absolute N.IRs were less than GBDincidence steadily except for 2009. Trend of N.MRs was increasing up to 2004, but stable thereafter. Trends of GBDmortality were also stable. Absolute N.MRs were less than GBDmortality for years up to 2003 and more than GBDmortality since 2005. The estimates of GLOBOCAN were more than N.IRs and N.MRs. CONCLUSIONS The GBDCC and GLOBOCAN values for LC in Iran are underestimates. Generation of data quality indices to present along with country specific estimates is highly recommended.