Abstract
OBJECTIVE: To examine the association of household type and household composition with concurrent stunting and overweight in young children from urban and rural Indonesia. DESIGN: This study is a secondary data analysis using a nationally representative cross-sectional survey. Household structure was analysed as household type, household size, number of working adults, number of dependent adults and children, and household head's gender. We defined 'concurrent stunting and overweight' as height-for-age Z-score <-2 and weight-for-height Z-score >+2 based on WHO growth standards. Multivariable logistic regression to test the aforementioned association was performed separately for urban and rural areas. SETTING: Data were from Indonesia Basic Heath Research 2013. PARTICIPANTS: Children aged 2-5 years (n 45 050). RESULTS: The prevalence of concurrent stunting and overweight children was 5·6 %. In rural areas, this prevalence differed significantly by household types and the highest prevalence was among children in nuclear two-parent households (6·8 %). In rural areas, children in extended households had lower odds of concurrent stunting and overweight than those from nuclear households (OR = 0·73, 95 % CI 0·59, 0·92). In urban areas, household size and number of working adults were significantly associated with the decreased odds of concurrent stunting and overweight in children. CONCLUSIONS: Household structure was associated with children's concurrent stunting and overweight in urban and rural regions of Indonesia. The patterns of the association might differ between urban and rural regions, but no significant interaction term was found.