Abstracts Health Sciences

Assessing the relationship between community poverty and quality ofcare for ESRD patients with consideration of case-mixcharacteristics

by Caitlin Maureen Casey

In order to assess the relationship between community poverty and quality of care for ESRD patients with adjustment for relevant case-mix characteristics,a cross-sectional analysis was conducted using facilities-level data on patients 18-90 years old from ESRD regional surveillance networks and county-level economic data from the national census in the US and its territories from 2005 to 2010. The exposure and outcome studied were community poverty and prevalent AVF use rate. AVF use rates were compared across a gradient of increasing community poverty in the context of case-mix characteristics. Differences in crude AVF prevalence rates and other differences in characteristics of communities and treatment centers were tested using linear regression, ANOVA tests and odds ratio tests of association. Correlations were examined using Pearson's correlation coefficient. Linear regression models were used to examine the association between the degree of community poverty and the prevalent AVF use rate in the context of case-mix characteristics. The multivariable model with the best fit was selected based on a maximum R2 statistic and the model with the most variables that significantly contributed to the model with a p-value <0.05. This study found that there is a statistically strong relationship between county poverty and prevalent AVF rate, such that as county poverty increases, the use of AVF in prevalent cases of ESRD decreases. The final multivariable model indicates that the prevalent AVF rate is not sensitive to county poverty alone, but also to a variety of demographic and clinical case-mix factors, including black ethnicity, diabetic etiology of ESRD and comorbid HTN and CHF. Amputation and unemployment in the patient population were also predictors included in the final multivariable model, although they did not significantly contribute to the model. Although case-mix factors significantly contribute to prevalent rates of AVF use, they do not significantly confound the relationship between county poverty and prevalent AVF use rate. Increasing AVF use in facilities in indigent areas would both benefit individual health outcomes and would help to minimize racial disparities, but improving healthcare access and education in impoverished areas may be equally or more important. Table of Contents Introduction Page 1 Methods Page 4 Results Page 7 Discussion Page 11 Conclusion Page 14 References Page 16 Tables and Figures Page 18 Advisors/Committee Members: McClellan, William (Thesis Advisor).