Health Disparities
Peripheral arterial disease and lower extremity amputation: Chicago, 2014-2024 Sandra Tilmon* Sandra Tilmon Rishab Bhatt Sam Volchenboum Trissa Babrowski
A manifestation of severe atherosclerosis, peripheral arterial disease (PAD) is strongly associated with lower extremity amputation (LEA) and mortality. Affecting over 8 million people in the U.S., health inequities exacerbate PAD, with non-Hispanic Black/African-American patients (NH Black) facing the highest LEA rates.
This study uses survival analysis to predict LEA with data from an urban research hospital with a historically underserved patient population. Analysis began at the time of entry into an “at risk” category between 2014 and 2024. K-means, hierarchical, and DBScan clustering were employed for unsupervised exploration. “Late presenters” were defined as patients with amputations within 1 month; these patients are excluded from survival incidence. A multivariate Cox Proportional Hazards Model (Cox PH) was performed and the proportional hazards assumption was tested; any variables in violation were modeled as time-varying.
The cohort included 8,784 patients of which 863 patients underwent LEA. Late presenters removed 538 cases from further analysis, leaving 7,196 patients at risk. 3,927 patients identified as NH Black, 2,131 as NH White, and 280 as Hispanic/Latino. 3,285 were male, most had Medicaid or Medicare insurance (3,280 patients), and the average age was 67 years.
The concordance index was strong at 0.84. Increased hazards were found with being male (1.56, 1.25-1.94) and with being non-Hispanic Black (2.03, 1.44-2.86). A strong increased hazard was present for our K-Means Cluster 3 (5.78, 2.54-13.64). Cluster 3 was strongly associated with being more NH Black, having Medicare insurance, and higher levels of congestive heart failure and hypertension. Prescriptions of antiplatelets, statins, and beta blockers were more common.
Although our study corroborated the health disparities documented in the literature, future efforts include spatially-based social determinants of health analysis to delineate particular factors associated with LEA.