Addressing Bias in AI Models for Chronic Disease Care: Why Equitable Digital Health Matters
At Adhera Health, we are committed to advancing digital health solutions that serve all families and patients equitably. A new study from our research team, set to be presented this week at the American Medical Informatics Association (AMIA) Informatics Summit in Pittsburgh, PA, reveals significant disparities in machine learning (ML) models used for clinical decision support in diabetes and heart disease management.
AI Bias in Healthcare: The Findings
Our research, Disparate Model Performance and Stability in Machine Learning Clinical Support for Diabetes and Heart Diseases, highlights how AI-driven predictive tools often favor younger and male patients while demonstrating inconsistent performance for older individuals. The study analyzed data from over 25,000 individuals with chronic diseases and found that age- and sex-related biases persist, raising concerns about fairness in AI-powered healthcare.
This finding underscores a crucial issue in healthcare AI: merely training models on diverse data does not automatically result in equitable outcomes. Many existing ML models struggle to accommodate the complexities presented by older patients, potentially leading to misdiagnoses, inappropriate treatments, or gaps in care.
The Need for More Inclusive AI
At Adhera Health, we believe that digital health solutions should work for everyone, not just the groups most commonly represented in training data. That’s why our research introduces a novel analytical framework that assesses model fairness beyond traditional performance metrics. This approach provides a blueprint for developing more reliable and unbiased AI-driven clinical tools.
Adhera Health’s Commitment to Ethical AI
Our family-centric digital companion for pediatric chronic conditions—including Type 1 diabetes, Type 2 diabetes, and childhood obesity—leverages AI-driven personalization while addressing the social, behavioral, and systemic factors that impact health outcomes. We prioritize fairness, inclusivity, and transparency in our AI models to ensure our solutions meet the needs of diverse patient populations.
“We are excited to see these important discussions take center stage at AMIA,” said Ricardo C. Berrios, CEO and Co-Founder of Adhera Health and contributor to the published research. “This study underscores the need for digital health companies to prioritize fairness and inclusivity in AI-driven tools. By leveraging new research and ethical AI principles, we can develop more effective digital health solutions that truly improve outcomes for all family members.”
Join the Conversation at AMIA 2025
If you're attending the AMIA Informatics Summit 2025, join us for a discussion on how to create more equitable AI-driven healthcare solutions. Our findings will be presented during Session S25 | Toward Implementation: Addressing Real-World Deployments on March 12 from 1:30 PM – 3:00 PM at the Omni William Penn Hotel in Pittsburgh, PA. Ioannis Bilionis, Data Scientist at Adhera Health, will be leading the presentation.
We look forward to engaging with industry leaders, clinicians, and researchers to drive meaningful advancements in ethical AI for healthcare. Stay connected with us for updates on our research and initiatives to promote equitable digital health for all.
This blog article is based on the press release - https://www.einpresswire.com/article/792388479/adhera-health-research-reveals-bias-in-ai-models-for-chronic-disease-care