We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Multi-Level Modeling of Diabetes Disparities in the United States.
- Authors
Schlundt, David G.; Niebler, Sarah E.; Newton, Doshandra
- Abstract
There is a disparity of 1.6 in diabetes prevalence in the United States (US) between African Americans and Whites. Using a national data set, we examine whether this disparity varies by county, and what variables predict county differences in diabetes prevalence and diabetes disparities. Data were obtained from the Behavioral Risk Factor Surveillance System (BRFSS) for 2001-2005. We identified 170 counties in the US for which there were an average of at least 25 Whites and 25 African Americans interviewed each year (n=367,485). Disparities were estimated for each county by taking a ratio of the prevalence in African Americans to the prevalence in Whites. Factor analysis of census variables was used to identify 5 orthogonal sets of variables named: 1) SES Distress, 2) City Size, 3) Educated Elite, 4) Concentrated Black, and 5) Segregation-Crowding. Each county received a score on each set of variables. Multi-level linear models were constructed using the HLM program (SSI, Lincolnwood, IL). The diabetes disparity ranged from 0.46 in Gwinnett County, GA to 5.30 in the District of Columbia. After fitting a level one logistic regression model of diabetes risk using age, gender, race/ethnicity, BMI, education, and income, 2 level two models were examined using the 5 census factor to predict: 1) the county intercept, and 2) the slope for African American ethnicity. Three variables predicted the slope (county prevalence): City Size, Educated Elite, and Segregation-Crowding with all three having an odds-ratio < 1.0. Two variables predicted higher slope for African American: City Size, and Educated Elite. Diabetes prevalence and diabetes disparities differ by county in the US. Two processes contribute to this: 1) factors that increase the prevalence of diabetes which includes living in more rural, and less urban-crowded settings; 2) protecive factors that keep the rate of diabetes lower in Whites which includes higher numbers of educated elites, and living in large urban centers. Large cities with more elites suppress the prevalence of diabetes but increase the importance of race (e.g., Washington D.C., white = 2.5, Black = 13.5). Rural areas without a concentration of highly educated residents have higher rates of diabetes, especially among Blacks (e.g., Ashville, NC, white=6.7, Black=22.1). We conclude that place makes an important contribution to diabetes risk and level of Black-White disparity, and that different strategies may be need to reduce risk and eliminate disparities in rural versus urban areas.
- Subjects
UNITED States; DIABETES; WHITE people; AFRICAN Americans; DISEASE risk factors
- Publication
Diabetes, 2007, Vol 56, pA90
- ISSN
0012-1797
- Publication type
Article