Objective: We describe differences in linear growth as measured by height-for-age z-score (HAZ) between children from Bengali and tribal populations in the Chittagong Hill Tracts of Bangladesh and examine factors associated with HAZ in both groups.
Design: We used nutritional surveillance data collected in Bangladesh from 2003 to 2006 to analyze HAZ among 12,006 children aged 6-23 months and conducted multivariate linear regression and Blinder-Oaxaca decomposition to quantify the relative contribution of independent variables to differences in HAZ between Bengali and tribal children.
Results: Mean HAZ was stable for Bengali children (-1.88 in 2003 to -1.90 in 2006) but improved for tribal children (-1.87 in 2003 to -1.68 in 2006). The difference between groups was significant across the study period (p = 0.008). Among Bengali children, HAZ was positively associated with father's schooling (β = 0.221; 95% CI: 0.079-0.363). Among tribal children, HAZ was positively associated with consumption of animal source foods (β = 0.073; 95% CI: 0.051-0.096) and goat ownership (β = 0.240; 95% CI: 0.025-0.454). Results of a likelihood ratio test indicated that distance to a health center was inversely associated with HAZ among tribal children (p < 0.001). Blinder-Oaxaca decomposition demonstrated a difference in pooled coefficients between groups (p = 0.004), explained primarily by differences in coefficients for paternal education (p = 0.001) and village-level prevalence of open defecation (p = 0.004).
Conclusions: Different responses among Bengali and tribal children to village-level open defecation are an explanatory factor for the difference in HAZ between Bengali and tribal populations. Open defecation may also act a proxy for unmeasured factors such as household environmental conditions and food hygiene.
Abbreviations: Chittagong Hill Tracts (CHT); height-for-age z-scores (HAZ); Nutrition Surveillance Project (NSP); World Health Organization (WHO).
Keywords: Bangladesh; nutrition; sanitation; stunting; tribal.