Objective: The purpose of this study is to document the prevalence of households with underweight and overweight persons (henceforth referred to as dual burden households) and their association with income and urban residence. The explorations by urban residence and income will test whether dual burden households differ from 'underweight only' and 'overweight only' households, respectively. These comparisons are relevant to differentiating or adapting nutrition-related interventions wherever obesity and undernutrition cluster at the household level. Population: Data analysis is based on national surveys conducted in Brazil, China, Indonesia, the Kyrgyz Republic, Russia, Vietnam and the United States. Methods: All persons were first classified into categories for underweight and overweight, using body mass index (BMI) cutoffs, and then all households were categorized into four types: dual burden, overweight, underweight and normal. Income and urban residence were explored as key risk factors for being a dual burden household, with the effects modeled separately for each country. Multiple logistic regression was used to explore income and urban risk factors, controlling for household size, region of residence and either urban residence or income, as appropriate. Results: In six of the countries studied, 22-66% of households with an underweight person also had an overweight person. Countries with the highest prevalence of dual burden households were those in the middle range of gross national product (GNP). The dual burden household is easily distinguished from the 'underweight only' households in Brazil, China, Indonesia, the United States and Vietnam. In these five countries dual burden households were more likely to be urban and more likely to be among the highest income tertile. There were no significant differences between dual burden and 'underweight only' households in Russia and the Kyrgyz Republic. In contrast, dual burden households were not easily distinguished from the 'overweight only' households in China, Indonesia, the Kyrgyz Republic, the United States and Vietnam. In Brazil and Russia dual burden households were more likely to be lower income and urban than 'overweight only' households. Conclusion: The prevalence of dual burden households presents a significant public health concern, particularly for those countries in the middle range of GNP. In some countries (China, Indonesia, the Kyrgyz Republic, the United States and Vietnam), dual burden households share sociodemographic profiles with overweight households, raising concerns for underweight individuals who may inadvertently become the focus of obesity prevention initiatives. For this reason, obesity prevention efforts should focus on messages that are beneficial to the good health of all, such as increasing fruit and vegetable intake, improving overall diet quality and increasing physical activity.
