Determinants of stunting in children under 5 years in Madagascar.

Madagascar is among countries where the prevalence of stunting is dramatically high in under 5 years old children. This study investigated the determinants of child stunting based on the UNICEF framework on the causes of malnutrition. A cutoff at 24 months was used to separate the child population into two groups. By using the latest Demographic and Health Survey (2009), logistic regressions were performed to determine the variables associated with stunting. In 2009, 40.1% of the 1,863 children aged 0-23 months and 53.9% of the 2,911 children aged 24-59 months were stunted contributing to the 48.5% overall stunting prevalence in the sample. Girls were less likely to be stunted (adjusted odds ratio with confidence interval [AOR] = 0.69 [0.55-0.88] and 0.84 [0.72-0.97], p < 0.01) than boys; the risk of stunting increased with age. Regarding underlying predictors, increased maternal height was associated with lower odds of stunting in both age groups (AOR = 0.75 [0.68-0.83] and 0.69 [0.61-0.77], p < 0.001). Children living in households using iodized salt (>15 ppm) had lower risk of stunting in the younger group (AOR = 0.76 [0.61-0.94], p < 0.05). Children living in urban areas were less likely to be stunted in both age groups (AOR = 0.67 [0.51-0.88] and 0.73 [0.59-0.90] respectively, p < 0.01]. Region of residence was also a significant basic factor for stunting. This study contributes to the understanding of the determinants of child stunting in Madagascar. The results confirmed the need for specific interventions for each of the two age groups.