[Cardiovascular mortality and impact of corrective techniques for dealing with underreported and ill-defined deaths]

Identify the impact of corrective techniques in the mortality rate trends for cardiovascular disease, ischemic heart diseases, and cerebrovascular diseases, in the City of Manaus, State of Amazonas, Brazil, between 1980 and 2007.

Data were obtained from the Unified Health System's Information Technology Department and the following steps undertaken: (1) proportional redistribution of death records, but without taking into account age and/or sex; (2) redistribution of ill-defined deaths among those whose cause is known; (3) treatment and reallocation of so-called "junk codes" in cardiology; and (4) correction of underrecordings of deaths by indirect techniques. Records treated in steps 1 and 2, and steps 3 and 4 together, generated base rates, adjusted rates, and adjusted and corrected rates, respectively, which were analyzed according to sex and age cohort; subsequently, they were standardized by the direct method. Simple linear regression was used for trend analysis.

The Mortality Information System's raw data underestimated death from heart diseases, mainly starting in the year 2000. With regard to the trend over time, the adjusted and corrected rates, compared to the base rates, pointed to the following: attenuation of the downward trend for cardiovascular diseases as a whole; accentuation of the downward trend for ischemic heart diseases; and stability in the rates for cerebrovascular diseases during the period.

This correction made it possible to identify an excess of deaths that had not been computed, generating more reliable adjusted and corrected death rates.