Trend, multivariate decomposition and spatial variations of unintended pregnancy among reproductive-age women in Ethiopia: evidence from demographic and health surveys

BACKGROUND: The magnitude of unintended pregnancy is unacceptably high and more than half of it end up with abortions. This may limit lower and middle-income countries to achieve the sustainable development goal targets of reduction of neonatal and maternal mortalities. Evidence on trends and spatial distribution of unintended pregnancy is limited. Therefore, this study aimed to assess the trend, multivariate decomposition, and spatial variations of unintended pregnancy among reproductive-age women in Ethiopia from 2000 to 2016. METHODS: Ethiopian Demographic and Health Data of 2000 to 2016 were used. A total weighted sample of 30,780 reproductive-age women participated. A multivariate decomposition analysis was employed to identify factors contributing to the change in the rate of unintended pregnancy in Ethiopia for 20 years from (1996 to 2016). The concentration index and graph were used to assess wealth-related inequalities, whereas spatial analysis was done to identify the hotspot of unintended pregnancy in Ethiopia. RESULTS: The 20-year trend analysis showed that the magnitude of unintended pregnancy among reproductive-age women decreased by 13.19 percentage points (from 39.76% in 2000 to 26.57% in 2016 EDHS). About 84.97% of the overall decrement was due to the difference in coefficient of the variables, whereas the remaining 15.03% was due to the difference in composition of the respondent. The differences in coefficient of the variables were decomposed by living metropolitan cities, having previous terminated pregnancy, and not having exposure to media; whereas, the change due to the composition, was expressed by having a household size of 1–3, living in metropolitan cities, being multipara and grand para, being unmarried and having no terminated pregnancy. Moreover, unintended pregnancies were more clustered in Addis Ababa and disproportionately concentrated in the poor groups. CONCLUSIONS: In Ethiopia, a substantial decrement in unintended pregnancy was observed in the past decade. More than four-fifths of this overall decrement was due to the difference in the coefficient of the variables. There was spatial clustering of unintended pregnancy in Ethiopia. A program intervention is needed for high-risk regions such as Addis Ababa. Health education and media campaign should perform for high-risk women such as those having terminated pregnancy, and professing Islam faith. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s41182-022-00440-5.