Classification systems for avoidability of infant deaths: different methods, different repercussions?

The objective of this study was to compare the avoidability of infant deaths according to different classification methods. This was a descriptive comparative study from 2006 to 2013 in Espírito Santo State, Brazil, focusing on the classification of 5,316 infant deaths according to five different methods. The methods of the International Collaborative Effort on Infant Mortality (ICE) and the SEADE Foundation correctly classified the highest proportions of deaths as avoidable versus unavoidable (94.6% and 94.4% correct classification, respectively). Most deaths resulted from quality problems in prenatal, childbirth, and postpartum care, regardless of which classification method was used. There were also considerable numbers of deaths from "ill-defined" causes according to all the methods, suggesting difficulty in access or precious care in health services. Avoidability methods provide an important instrument for diagnosis of quality problems in health services performance and orientation of measures to reduce avoidable infant deaths. Thus, strengthening maternal and child care and investment in training and capacity-building for health professionals and services are priorities for public policies to reduce infant mortality.