A multilevel analysis of individual, household and community level predictors of child diarrhea in Eswatini

Abstract

The aim of the study was to investigate the effect of individual, household and community level factors on child diarrhea. The study used a pooled secondary data analysis approach from the 2010 and 2014 Eswatini Multiple Indicator Cluster Survey (MICS). Multilevel logistic regression was used for data analysis. Compared to children aged under 6 months, the odds of having diarrhea were higher for children aged 6-11 months, adjusted odds ratio, AOR 5.34(95% Confidence interval):2.39-11.90) and 12-23 months AOR. 3.90 (1.76-8.64).  Reduced odds of diarrhea were observed among children with a normal height for age (AOR0.09(0.01-0.93) and above normal height for age (AOR: 0.07 (0.01-0.85) compared to those with a growth deficit. Children from households with no toilet facility at all had 283% increased risk of having diarrhea compared to those with flush toilets. Regionally, the risk was higher in the Hhohho region (AOR2.59 (1.08-2.65), reference: Lubombo region. Higher odds were observed among communities with a higher and medium proportion of diarrhea AOR9.77(4.59-20.80), AOR 3.60 (2.26-6.20) respectively, compared to communities with a low proportion of diarrhea