Racial Disparities in Amputation Rates Among Native Americans with Peripheral Artery Disease: Analysis of the Health Care Cost and Utilization Project Database


      While studies have documented disparities in amputation rates for Native American patients with peripheral artery disease (PAD), the importance of specific factors has not been quantified. This study seeks to quantify these factors, and to determine how much of the difference reflects observable versus unexplained characteristics.


      This study used the nationally-representative HCUP inpatient database from 2006-2013 for Native American and Caucasian patients with a primary diagnosis of PAD. Amputation rates were calculated for any leg amputation [above or below-the-knee] and separately for below-the-knee only. Logistic regression models were estimated to identify the determinants of amputation rates for Native Americans compared to Caucasians. All models were adjusted for patient demographics, income strata, comorbidities, PAD risk factors, and hospital characteristics. The Blinder-Oaxaca decomposition method was employed to ascertain and quantify factors that contribute to the differences in amputation rates between Native Americans and Caucasians.


      Treatment patterns are dramatically different for Native Americans residing in the West Census Region relative to the other regions. Multivariable results reveal that Native Americans in the West are more than twice as likely to be amputated as are Caucasians. Decomposition results show that observed factors collectively explain just 43-62% of the variation in amputation rates. Individual observed factors contribute little to explaining these disparities. In contrast, results from the other three Census Regions reveal that amputation rates are not statistically different between Native Americans and Caucasians.


      Native Americans with PAD who reside in the West Census Region are substantially more likely to be amputated than are Caucasians. Most of this variation appears to reflect differences in treatment patterns that are unrelated to illness severity, demographic factors, hospital characteristics, and other factors. These substantial unexplained disparities may reflect systematic differences in treatment patterns by race.