Study Uses AI to Uncover Disparities in Amputation Rates Among Rural and Minoritized Populations

Recent research has shed light on the persistent disparities in lower limb amputations among rural residents and racial and ethnic minorities suffering from vascular diseases. A groundbreaking study published in Epidemiology employs artificial intelligence (AI) to explore the complex factors behind these disparities, revealing an unaccounted-for element that points to implicit biases influencing clinical decisions.

The study analyzed hospitalizations from 2017 to 2019 across five states—Florida, Georgia, Maryland, Mississippi, and New York—focusing on individuals under 40 diagnosed with Peripheral Artery Disease (PAD) or Chronic Limb-Threatening Ischemia (CLTI). Researchers used AI to assess over 70 variables, including clinical factors such as age and comorbidities, healthcare system capacity, legal and regulatory environments, and social determinants like distance to emergency services and neighborhood income levels.

Findings indicated that, after adjusting for known factors, disparities in amputation rates persisted among Black, Hispanic, Native American, and White populations in rural areas, as well as Black and Native American populations in urban settings. The study suggests that unconscious biases at the physician and hospital levels significantly contribute to these outcomes.

"The AI model allowed us to identify the underlying reasons for higher amputation rates among specific groups, highlighting the likely role of implicit bias in clinical decision-making," said lead author Paula Strassle, an epidemiology professor at UMD. She emphasizes that these biases can influence treatment choices, often resulting in limb loss instead of limb-saving procedures.

PAD affects more than 12 million adults in the US, leading to symptoms like leg pain and numbness, and in severe cases, necessitating amputation. About 10% of PAD patients develop CLTI, requiring urgent intervention such as revascularization, a costly and complex surgical procedure with limited availability due to a shortage of vascular surgeons.

The study underscores the need for evidence-based guidelines that incorporate these findings to help clinicians make more equitable decisions. Senior author Katharine McGinigle advocates for integrating AI tools into clinical practice to mitigate biases and improve patient outcomes.

By identifying overlooked factors and potential biases, this research aims to influence health policies and develop comprehensive guidelines to ensure advanced vascular disease treatments are accessible and equitable for all populations.