Evaluating ethnic disparities of diabetes related amputations

Lower extremity amputation (LEA) rates have been historically high among people with diabetes. In particular males and African Americans have been shown to be at much greater risks than females and Whites for having an amputation. The objectives of this study are to determine if there has been any success in reducing the disparity of LEAs in males and African Americans with diabetes over recent years. We obtained the hospital discharge data from the Office of Research and Statistics for the state of South Carolina for the years 1996 through 2003. The dataset included all hospital discharges involving diabetes from all civilian hospitals throughout the state. Individuals with at least one LEA procedure were identified by ICD-9-CM procedure codes 84.10–84.19. Traumatic LEAs, ICD-9-CM 895–897, were not considered as having a diabetes related amputation. We used a linear model to predict LEA rates per 1000 diabetes hospitalizations. Model 1 contained covariates year, race, and a race-year interaction term. Model 2 contained covariates year, gender, and a gender-year interaction term. We tested significance of the interaction effects by using z-tests to determine if changes in LEA rates were equivalent across populations. In 1996 LEA rates for African Americans were 37.4/1000 diabetes hospitalizations compared to 21.3/1000 for Whites. Additionally, rates for males were 35.1 compared to 22.6 for females in 1996. Rates for all populations drastically decreased in 2003 (African Americans 25, Whites 14.2, males 23.5, females 14.7). Both interaction terms were highly significant (p < 0.001), signifying both African American and male rates are dropping quicker than their comparison populations (Whites, females). Although LEA rates are dropping faster for African Americans and males compared to Whites and females, there is still an apparent disparity among these populations. Future studies are needed to pinpoint the underlying mechanisms of such diabetes disparities.