Trend, multivariate decomposition and spatial distribution of perinatal mortality in Ethiopia using further analysis of EDHS 2005-2016

Abstract Background Perinatal mortality is the global health problem, especially Ethiopia has the highest perinatal mortality rate. Studies about perinatal mortality were conducted in Ethiopia, but which factors specifically contribute to the change in perinatal mortality across time is unknown. Objectives To assess the trend, multivariate decomposition and spatial analysis of perinatal mortality in Ethiopia using EDHS 2005 to 2016. Methods A community-based cross-sectional study design was used. EDHS 2005-2016 data was used and weighting has been applied to adjust the difference in the probability of selection. Logit based multivariate decomposition analysis was used using STATA version 14.1. Moran’s I statistics using ArcGIS was also used to identify the significant clustering of perinatal mortality. Result The trend of perinatal mortality in Ethiopia decreased from 37 per 1000 births in 2005 to 33 per 1000 births in 2016. About 83.3% of the decrease in perinatal mortality in the survey was attributed to the difference in the endowment (composition) of the women. Among the differences in the endowment, the difference in the composition of ANC visits, take TT vaccine, urban residence, occupation, secondary education, birth attendant significantly decreased the perinatal mortality in the last 10 years. Among the differences in coefficients, skilled birth attendant significantly decreased the perinatal mortality. The spatial distribution of perinatal mortality was randomly distributed. Conclusion The perinatal mortality in Ethiopia has declined over time. Variables like ANC visit, taking TT vaccine, urban residence, have occupation, secondary education and skilled birth attendant reduce perinatal mortality. Perinatal mortality was distributed randomly in Ethiopia. To reduce perinatal mortality more, scaling-up the maternal and newborn health services has a critical role.