Objective: To evaluate the automated processing and the epidemiological potential of multiple-cause-of-death data listed on death certificates in Brazil in 2003.
Methods: Data were obtained from the Brazilian Mortality Information System. The distribution of the number of causes per death certificate was processed using the Multiple Cause of Death Tabulator software and expressed as a percentage of the total deaths in federation units, regions, and in the country as a whole. For the main causes of death identified in Brazil in 2003, we calculated the rate at which each cause was listed as the "underlying cause of death" in relation to the total times the cause was mentioned.
Results: The mean number of causes of death per certificate ranged from 2.07 in the state of Maranhão to 3.15 in the state of São Paulo, and from 2.45 in the Northeast to 2.99 in the Mid-West. For the entire country, the mean number of causes of death per certificate was 2.81. External causes and neoplasms, unattended deaths, human immunodeficiency virus, and alcoholic liver disease were listed as the underlying cause more than 90% of the times they were mentioned; neoplasms, Chagas' disease, congenital malformations, ischemic heart diseases, and cerebrovascular diseases, between 70% to less than 90% of the times they were mentioned; and pneumonias, heart failure, alcohol dependence, malnutrition, hypertensive diseases, anemias, ill-defined causes, prematurity, septicemias, and respiratory failure, less than 40% of the times they were mentioned.
Conclusions: Although underlying causes of death are still essential to analyze historical trends, compare countries, and to guide the prevention of death, multiple-cause-of-death-data offer a new insight into the study of mortality. The combination of the two methodologies is more useful than the isolated use of either approach.