INTRODUCTION: Infant mortality was studied in an urban area of Southeastern Brazil in the period from 1990 to 1992 using data from death certificates collected at the registry office, by the application of methods for obtaining a collective diagnosis which will assist in the identification and choice of strategies for the control of local problems. MATERIAL AND METHOD: The original data were corrected using documental research into health services and household interviews. Data of the Live Birth Information System (SINASC) was used to study variables such as maternal age and birthweight. The quality of original death certificates was initially analyzed using the amount of information, sensitivity, specificity and Kappa value. RESULTS: The global sensitivity for the underlying cause was 78.84% and Kappa 71.32% for the total of causes. One hundred and eighty-nine deaths occurred, 66.15% of them in the neonatal period, (41.28% during the first day of life) and 33.85% in late childhood. The birthweight of 58.28% of deaths was less than 2,500g. The underlying causes of death were studied the by possibility, of their avoidance (a method developed by Erica Taucher), by a "reduced" group of causes (utilized in International Collaborative Effort (ICE)), multiple causes statistics and geographical distribution. It was observed that in the deaths occurring up to 27 days, 22.23% could have been avoided by adequate care during labour, and 20.64% could have been avoided by early diagnosis and early treatment, 13.75% by good pregnancy care and only 7.94% were unavoidable. Of the deaths occurring in late childhood, 12.17% were classified as of avoidable causes and 4.23% were considered as unavoidable. Using ICE groups, 58.74% died of immaturity or asphyxia, 19.58 of infection and 12.17% from congenital abnormalities. CONCLUSIONS: The results suggest that priority be given to obstetrical care at delivery and during labour and to the pediatric care of low birth weight, among others. The analysis using multiple causes statistics shows that 76.05% of the deaths have underlying causes related to neonatal disorders and confirms the relationship with the weight deficiencies of the newborn. The maternal complications were also related to weight deficiencies. Great differences were identified in infant mortality rates in urban zones not only restricted to the value of the rates but also to the diseases responsible for the occurrence of deaths. We therefore conclude that there is an advantage to be gained in using the four methods, which are complementary, for studying or planning actions with a view to prevent infant mortality.
Infant mortality; Underlying cause of deathunderlying cause of death
