Are Primary Health Care Features Associated with Reduced Late Neonatal Mortality in Brazil? An Ecological Study.

Abstract

To analyze the effect and efficiency of the characteristics of PHC facilities' structures and the work process of PHC teams on late neonatal mortality (LNM). This ecological time-series study adopted 3.764 Brazilian municipalities as analysis units. The independent variables were sorted into three hierarchical levels and four blocks. The distal level consisted of economic and demographic variables; the intermediate level comprised health coverage and demand for services; and the proximal level included structure and work process. The dependent variable was LNM. A linear mixed-effects regression analysis with a hierarchical approach was performed, estimating the crude (β) and adjusted (alpha = 5%) regression coefficients. Data involution analysis and municipalities were the decision-making unit according to their strata. LNM was directly associated with the number of live births and unemployment rate. LNM was inversely associated with the year, per capita income, the community health worker's strategy coverage, vaginal delivery, household visits, and available vaccines. In the 2002-2014 period, the number of municipalities efficient in reducing LNM dropped from 38 to 27. In 2014, a more significant investment occurred in the number of vaginal deliveries in almost all strata to make inefficient municipalities efficient. The deaths of children aged 7-28days are affected by the characteristics of the PHC structure and work process.