OBJECTVES: to evaluate predictors of neonatal death among live births and the quality of use of the institutional Live Birth Information System register (Sinasc). METHODS: a time serie based on data from the Sinasc of the Instituto de Medicina Integral Prof. Fernando Figueira (1995-2006) in the city Recife, Pernambuco, Brazil. The following were evaluated: the completeness of the forms and the tendency regarding risk factors for infant mortality (low birth weight; anoxia; premature birth; caesarian birth; lack of prenatal care; adolescent pregnancy; illiteracy; and having had no live or still born child). Statistical significance was tested using Student's t test with p<0.05 in a linear regression model. RESULTS: 58,689 live births occurred with a continued rise from 2002 onwards; low birth weight, 22.8%; Apgar <7 1º minute 15.3%; premature 22.4%; Caesarian birth 38,2%; adolescent mothers 27.2%; illiteracy 2.7% and 89% with no live or still born child. One percent of variables were not recorded. The following variables increased (p<0.05): Apgar in the 1st minute, low birth weight, premature birth and caesarian birth; and the following decreased (p<0,05): no pre-natal consultation, adolescent pregnancy, illiteracy and having had no live or still born child. CONCLUSIONS: increasing percentages for the predictors of infant mortality suggest problems with the quality of pre-natal care. With only 1% of information overlooked, the quality of the keeping of SINASC records at the institution was found to be excellent. Routine evaluation of births allows for agility of information and appropriate intervention to prevent neonatal mortality at local level.
Health evaluation; Neonatal mortality; Information system
