Spatial analysis of provincial and district trends in stunting among children under five years in Nepal from 2001 to 2016

Background The average prevalence of stunting reported by the Nepal Demographic Health Survey from 2001 to 2016 only reports the prevalence of stunting at the national level and provincial and district level information is missing. Also, no previous study has reported a provincial trend in stunting from 2001 to 2016 in Nepal. This study for the first time presents the spatial trend of stunting among children under five years for 7 provinces and 77 districts of Nepal over 15 years using Demographic Health Survey Global Positioning System coordinates, Demographic Health Survey indicators, and geospatial covariates. Methods This is a secondary analysis of data from Nepal Demographic Health Survey from 2001 to 2016. The study population was children under five years. The outcome variable was stunting, which was analyzed as per districts and provinces. Sample weight was applied to calculate the percentage of stunting and 95% confidence interval for all survey years. The geographic dataset was used to provide information about the latitude and longitude of the survey cluster. Poisson-based model was used during the purely spatial analysis in SatScan for identification of clusters with stunting caseload. Results The reduction in the prevalence of stunting among children under five years has not been equal when disaggregated for district and provincial level data. In 2001, 57 districts had a prevalence of stunting among children above or equal to 50%, which has reduced over time except for districts in Karnali province. In 2016, 16 districts had a prevalence of stunting above or equal to 50%. Jumla (91.7%) and Kalikot (77.8%) still had the highest prevalence of stunting as of 2001. Among 7 provinces, the prevalence of stunting is found highest in Karnali for all subsequent survey years. Sudurpaschim and Madhesh provinces also had a high proportion of stunted children. The highest reduction in the prevalence of stunting was noted for Province Bagmati (by 30%) and Gandaki (by 28%). Conclusion The inequalities in childhood stunting persisted at the district and provincial levels although a good decline was noted at the national level. This calls for rigorous attention to be provided to districts and provinces with a high prevalence of stunting, and being prioritized for a targeted intervention.