ASGE Postgraduate Course at ACG: Innovative Practices for the Endoscopist of Today, Tomorrow and Beyond

7_Fola_May_ASGE_AI_Equity

Pdf Summary

Dr. Folasade P. May's presentation focuses on using artificial intelligence (AI) to address health disparities in gastroenterology. She emphasizes the importance of health equity, defined as providing fair and just opportunities for everyone to be as healthy as possible, and notes that health disparities are closely linked to economic, social, or environmental disadvantages.<br /><br />The presentation addresses biases in AI and machine learning (ML) in gastroenterology and hepatology. Dr. May highlights that biases can arise in various forms, including selection bias, data collection bias, and algorithm development biases. Specific examples mentioned include:<br /><br />1. **Esophageal Cancer**: Research and AI technologies predominantly focus on conditions like Barrett’s esophagus and esophageal adenocarcinoma (EAC), which benefit White populations the most, while esophageal squamous cell carcinoma (ESCC), more common among Black and Asian individuals, is less studied. <br /><br />2. **Inflammatory Bowel Disease (IBD)**: AI/ML algorithms are often based on data from predominantly White populations, though the incidence of IBD is rising among non-White groups, leading to worse management and outcomes for Black and Latino patients.<br /><br />3. **Colorectal Cancer**: AI tools for colonoscopy, like computer-aided detection, need to be equally effective for high-risk polyps in diverse populations. Vulnerable groups often face barriers to accessing such advanced technologies.<br /><br />Dr. May outlines strategies to mitigate these biases:<br />- Involve health equity experts in AI/ML conception, development, and deployment.<br />- Conduct research with diverse study populations and settings.<br />- Implement regulatory measures to ensure equity and fairness.<br />- Mandate pre-deployment and post-deployment audits for algorithmic performance across different subpopulations.<br /><br />Finally, Dr. May stresses that AI has the potential to greatly advance healthcare. However, deliberate efforts are needed to address and prevent AI-driven biases to ensure that health inequities are not exacerbated. The presentation calls for the inclusion of diverse perspectives and expertise in AI development, including increasing the diversity of the workforce in healthcare and research settings.

Keywords

artificial intelligence

health disparities

gastroenterology

health equity

AI biases

machine learning

esophageal cancer

inflammatory bowel disease

colorectal cancer

diverse populations