Spatial patterns and determinants of undernutrition among late-adolescent girls in Ethiopia by using Ethiopian demographic and health surveys, 2000, 2005, 2011 and 2016: a spatial and multilevel analysis

Background Undernutrition among late-adolescent girls (15-19 years) in Ethiopia is the highest among Southern and Eastern African countries. However, the spatial variation of undernutrition as a national context is not well understood in Ethiopia. This study aimed at the spatial patterns and determinants of undernutrition among late-adolescent girls in Ethiopia. Methods Secondary data analysis was conducted from women's data of four consecutive Ethiopian Demographic and Health Surveys (EDHS) from 2000 to 2016. A total of 12,056 late-adolescent girls were included in this study. The global spatial autocorrelation was assessed using the Global Moran's I autocorrelation to evaluate the presence of geographical clustering and variability of undernutrition. SaTScan cluster analysis by using the Bernoulli model to detect most likely SaTScan cluster areas of significant high-rate and low-rate of undernutrition was explored. A Multilevel binary logistic regression model with cluster-level random effects was fitted to determine factors associated with undernutrition among late-adolescent girls in Ethiopia. Results Undernutrition was clustered nationally during each survey (Global Moran's I = 0.009-0.045, Z-score = 5.55-27.24, p-value < 0.001). In the final model, individual and community level factors accounted for about 31.02% of the regional variations for undernutrition. The odds of undernutrition among 18-19 years of adolescent girls, was 57% (AOR = 0.43; 95% CI: (0.35-0.53) lower than those 15-17 years old. Late-adolescent girls with higher educational status were 4.40 times (AOR = 4.40; 95% CI: (1.64-11.76) more likely to be undernourished than those with no educational status. The odds of undernutrition among late-adolescent girls, with the occupation of sales, was 40% (AOR = 0.60; 95% CI: 0.43-0.84) lower than those with not working adolescents. The odds of undernutrition, among late-adolescent girls, having an unimproved latrine type, was 1.79 times (AOR = 1.79; 95% CI: 1.15-2.79) higher than those participants with improved latrine type. The odds of undernutrition among late-adolescent girls with rural residents was 2.33 times higher (AOR = 2.33; 95% CI: 1.29-4.22) than those with urban residents. Conclusion Undernutrition among late-adolescent girls was spatially clustered in Ethiopia. The local significant clusters with high prevalence of undernutrition was observed in Northern and Eastern Ethiopia. Those regions with a high prevalence of undernutrition should design interventions to combat undernutrition.