Correction for reported cervical cancer mortality data in Brazil, 1996-2005.

OBJECTIVE: To develop a methodology for correction of reported cervical cancer deaths in Brazil. METHODS: Data on 9,607,177 cancer deaths were obtained from the Brazilian National Mortality Database for the period between 1996 and 2005. For correction of underreporting of deaths, factors generated by the Global Burden of Disease Study in Brazil-1998 were used. Proportional distribution was used in order to correct the categories of unknown, incomplete or ill-defined death diagnosis. The corrections were applied to each Brazilian state and the results were presented for Brazil nationwide, macroregions, and geographical areas (capital, other cities of metropolitan areas and interior cities) as percent variability of cervical mortality rates before and after correction. Corrections were analyzed by multivariate linear regression with interaction terms between macroregion and geographical area. RESULTS: After correction, cervical cancer mortality rates showed an increment of 103% nationwide, ranging between 35% (Southern region capitals) and 339% (Northeastern region interior cities). The reallocation of cervical cancer deaths not otherwise specified resulted in greater mortality rate increments. The percent correction by year of death revealed steady trends nationwide. CONCLUSIONS: The study results showed that the proposed methodology was appropriate for the correction of cervical cancer mortality rates in Brazil. It evidenced that cervical cancer mortality is even higher than that reported.