Ines Gonzalez-Casanova1, Olga Lucia Sarmiento2, Michael Pratt3, Julie A Gazmararian4, Reynaldo Martorell4, Solveig A Cunningham4, Aryeh Stein4. 1. Hubert Department of Global Health, Rollins School of Public Health, Nutrition and Health Sciences Program, Emory University, 1518 Clifton Rd NE, 7000H, Atlanta, GA 30322. E-mail: igonza2@emory.edu. 2. Facultad de Medicina, Universidad de los Andes, Bogotá, Colombia. 3. Rollins School of Public Health, Emory University, Facultad de Medicina, Universidad de los Andes, Bogotá, Colombia, and Centers for Disease Control and Prevention, Atlanta, Georgia. 4. Rollins School of Public Health, Emory University, Atlanta, Georgia.
Abstract
INTRODUCTION: Information from high-income countries is often used to design childhood obesity prevention interventions in low- and middle-income countries, even though determinants may differ greatly between settings. METHODS: We examined the associations of individual, family (household), and community (municipality) characteristics with body mass index (BMI) z scores and likelihood of overweight among children aged 5 to 18 years measured for the Colombian National Nutrition surveys of 2005 (n = 9,119) and 2010 (n = 21,520). We used 3-level hierarchical linear models with child as level 1, household as level 2, and municipality as level 3. RESULTS: The prevalence of combined overweight and obesity among Colombian children and adolescents was 15.7% in 2005 and 16.6% in 2010. The household level explained 40% in 2005 and 31% in 2010 of the variability in BMI z scores. Wealth was positively associated with BMI in 2005 (0.09 increase in z score per wealth quintile) and 2010 (0.13 increase in z score per wealth quintile) (P < .01). Children and adolescents from extended families had higher BMI z scores than those from nuclear families; BMI z scores were inversely associated with the number of family members living in the same household. The municipality level explained only between 2% and 3% of the variability in BMI. Income inequality was positively associated with BMI z scores in 2010. CONCLUSION: These patterns differ from those commonly described in high-income countries and suggest more appropriate opportunities for interventions to prevent child and adolescent obesity in Colombia and other Latin American settings and populations.
INTRODUCTION: Information from high-income countries is often used to design childhood obesity prevention interventions in low- and middle-income countries, even though determinants may differ greatly between settings. METHODS: We examined the associations of individual, family (household), and community (municipality) characteristics with body mass index (BMI) z scores and likelihood of overweight among children aged 5 to 18 years measured for the Colombian National Nutrition surveys of 2005 (n = 9,119) and 2010 (n = 21,520). We used 3-level hierarchical linear models with child as level 1, household as level 2, and municipality as level 3. RESULTS: The prevalence of combined overweight and obesity among Colombian children and adolescents was 15.7% in 2005 and 16.6% in 2010. The household level explained 40% in 2005 and 31% in 2010 of the variability in BMI z scores. Wealth was positively associated with BMI in 2005 (0.09 increase in z score per wealth quintile) and 2010 (0.13 increase in z score per wealth quintile) (P < .01). Children and adolescents from extended families had higher BMI z scores than those from nuclear families; BMI z scores were inversely associated with the number of family members living in the same household. The municipality level explained only between 2% and 3% of the variability in BMI. Income inequality was positively associated with BMI z scores in 2010. CONCLUSION: These patterns differ from those commonly described in high-income countries and suggest more appropriate opportunities for interventions to prevent child and adolescent obesity in Colombia and other Latin American settings and populations.
Authors: Jinan C Banna; Opal Vanessa Buchthal; Treena Delormier; Hilary M Creed-Kanashiro; Mary E Penny Journal: BMC Public Health Date: 2016-01-15 Impact factor: 3.295