Óbitos infantis evitáveis em Belo Horizonte: análise de concordância da causa básica, 2010-2011

Abstract

OBJECTIVE: Evaluate the reliability between the underlying cause of preventable infant deaths in the original Death Certificate and the redone one after investigation by the Surveillance Committee for the Prevention of Deaths BH Vida (CPOMFI). METHODS: Cross-sectional population-based study that evaluates infant deaths of residents in BeloHorizonte, occurred in 2010 and 2011, children weighing 1500 grams or more without severe congenital malformation, studied by the CPOMFI. Deaths certificates were redone based on the death surveillance and data from National System Live Birth (SINASC) without prior knowledge of the original death certificate. The agreement between the original and thenew cause of death was evaluated using the kappa statistics and observed agreement after coding the death causes in the new death certificate made by a trained technician using the rules of ICD 10. To select the underlying cause of the new death certificate the SelectionSystem of Underlying Cause of Datasus was used and the underlying cause of the original death certificate was based on the Mortality Information System (SIM). The underlying cause of death in the original and new death certificate classified was also classified according tothe Reduced List Tab of Infant Mortality Causes (LIR-MI). Infant deaths in general and the investigated ones were evaluated according to the distal, intermediate and proximal variables and groups of causes of death, as well as the according to LIR-MI. RESULTS: The study was based on 149 infant deaths, representing 22.4% of the deaths reported in SIM on thestudied period. In both groups deaths took place predominantly within hospitals, were male and black children, were born from a single pregnancy and by vaginal delivery; their mothers had 20 and 35 years of age and eight or more years of schooling, . Most of the deathsoccurred in the early neonatal period (50.7%), were low birthweight children (72.8%) and premature (70.7%). The studied infant deaths occurred predominatly during the post-neonatal period (54.4%), had adequate birthweight (70.5%) and were term babies (74.1%). The maingroup of causes of death according to ICD-10 and LIR -MI were perinatal conditions, both in infant and studied deaths. The concordance between the underlying cause in the original Death Certificate and the new Death Certificate was weak for tested models - the underlying cause with 3 and 4-digit ICD-10 and LIR-MI groups, as well as the low observed agreement. There was an important change in cause of detah on the post surveillance group of death that enhanced the identification of maternal factors and Sudden Death Syndrome in Childhood. CONCLUSION: Investing in the constitution of the Committees for the Death Prevention,training physicians to adequate registration of death certificates and institute a qualified autopsy are strategies to qualify the underlying cause of infant death. Which is essential for a appropriate diagnostic on infant mortality and subsidize public health decision making. The large percentage of deaths due to potentially preventable causes demonstrates the need for improvements in access and / or quality of prenatal, childbirth and child care.