Individual level and community level factors affecting exclusive breast feeding among infants under-six months in Ethiopia using multilevel analysis

Background Exclusive breastfeeding (EBF) is the safest and healthiest option of feeding among infants in the first 6 months throughout the world. Thus, promotion of EBF is essential to prevent complex infant health problems even at the adulthood level. But majority of previous studies focused on individual level determinants of EBF by using basic regression models in localized areas. This study aims to identify individual level and community level determinants of EBF which would be helpful to design appropriate strategies in reducing infant mortality and morbidity. Methods It is a secondary data analysis using the 2016 Ethiopian Demographic and Health Survey (EDHS) data. A total of 1185 infants under 6 months of age were included in the analysis. Multilevel logistic regression model was employed to investigate factors significantly associated with EBF among under-six month's infants in Ethiopia. Adjusted odds ratio (AOR) with 95% confidence interval (CI) was used to measure the association of variables whereas Intra cluster correlation (ICC), median odds ratio (MOR), and proportional change in variance (PCV) were used to measure random effects (variation). Result In multilevel logistic regression; 4-5 months age infant (AOR = 0.04, 95%CI:0.02-0.07), female infants (AOR = 2.51, 95%CI:1.61-3.91), infant comorbidities (AOR = 0.35, 95%CI: 0.21-0.57), household wealth index (AOR = 10.34, 95%CI: 3.14-34.03) and antenatal care (AOR = 2.25, 95%CI:1.32-3.82) were determinants of EBF at individual level. Whereas, contextual region (AOR = 0.30, 95% CI: 0.10-0.87), community level of postnatal visit (AOR = 2.77, 95% CI: 1.26-6.58) and community level of maternal employment (AOR = 2.8, 95% CI: 1.21-6.47) were determinants of EBF at community level. The full model showed up with higher PCV; that is, 46.8% of variation of exclusive breastfeeding was explained by the combined factors at the individual and community levels. Similarly, it showed that the variation in EBF across communities remained statistically significant (ICC = 8.77% and variance = 0.32 with P < 0.001). The MOR at final model indicates there was significant cluster difference for EBF indicating the heterogeneity was explained by both individual and community level factors. Conclusion and recommendation Our study showed that both individual and community level determinants were significantly associated with EBF practice. Based on our findings it is strongly recommended to promote and enhance antenatal and postnatal care services utilization of mothers and more emphasis should be given for infants with comorbid conditions and those who live in the pastoralist regions.