Rural vs. urban differences in body mass index of Hispanic elementary age children
Abstract
The purpose of this research was to evaluate Hispanic children in rural and urban environments in a north-central Indiana region to determine if geographical residency was an additional risk factor for overweight/obesity in this sub-group of the population. Data from a north-central health department in Indiana that was collected during the 2010-2011 school year on children (5-12 years of age), was utilized to address this question. The epidemiologist and other trained staff collected weight, height, age, sex, and race/ethnicity for 471 children at 4 elementary schools, 3 in an urban environment (n=253) and 1 in a rural environment (n=218). Centers for Disease Control and Prevention (CDC) sex-specific percent BMI for children were used for weight classification. Schools were classified as rural or urban based on Rural Urban Continuum Code (RUCC) for the county in which they resided. A linear mixed model showed no significant association between BMI and geographical residency. However, this model did show a possible relationship between BMI and age and sex (p-value <0.0001 and 0.03 respectively). This suggests that BMI classification or obesity as depicted by this model may be significantly related to the child's age and sex but not geographical residency. An ordinal logistic regression of the data showed a protective effect against obesity for rural residing Hispanic children (OR 0.57); meaning rural dwelling Hispanic children were 0.57 times as likely as their urban counterparts to obese or about half as likely. Further testing was conducted by way of a binary logistic regression to compare obesity against all other BMI categories suggested geography (OR 0.51, p-value 0.001) as well as sex (OR 0.63, p-value 0.02) were significant predictors of obesity. These results suggest that rural dwelling Hispanic children are 0.51 times as likely as their urban dwelling counterparts to be classified as obese and regardless of geographical residency, female Hispanic children are 0.63 times as likely as males to be obese. Therefore both the linear mixed model and the ordinal logistic regression suggest that female Hispanic children are at reduced risk of obesity compared to male Hispanic children. Binary logistic regression was conducted to evaluate overweight and obesity combined compared to all other BMI categories. No significant results at the level of p-value 0.05 were observed for geographical residency. However, a p-value of 0.095, suggest that further investigation of the relationship of geography to the combined overweight and obese classification is warranted. The results of this study are contradictory to previous research on the impact of rural vs. urban environments on the prevalence of obesity in general populations. Thus, further study is needed to determine if these results hold true in larger Hispanic samples and in different areas of the country.
Degree
M.P.H.
Advisors
Lyle, Purdue University.
Subject Area
Public health
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.