OBJECTIVES: to quantify inaccurate and missing information and to assess the quality of data related to infant deaths in Blumenau during the year of 1998. METHODS: information was obtained on 125 deaths entered in the civil registration records and of 12 other deaths from hospitals in Blumenau. A new Death Certificate was filled (New DC) for comparison with each original Death Certificate (Original DC) based on hospital records information. New coefficients of child mortality and underlying causes were compared with the original information and the consistency between variables was determined. RESULTS: an inaccuracy of 8,8% of the death records were demonstrated with the following outcomes: official coefficients are Child Mortality Coefficient 16,7; Perinatal Mortality Coefficient 21,0, Neonatal Mortality Coefficient 3,9, and the new ones depicted, 18,0, 23,1, 13,8 and 4,1 respectively. Social and demographic variables of fetal and non-fetal deaths are poorly reported not only in the Death Certificates but in hospital records as well. In general the Original Death Certificates for fetal deaths are worse than those of non-fetal deaths. The Kappa index between Original Death Certificates and New Death Certificates for perinatal causes and congenital malformations was of 0,73 and 0,80 respectively. CONCLUSIONS: notwithstanding the deficiency of information and unsatisfactory quality of the social and demographic data of infant deaths, the causes could be used for health planning considering their consistency.
Child mortality; information systems; basic causes of death; death records