What AMIA 2025 Taught Adhera Health About Fair, Responsible AI in Healthcare

Insights from the AMIA Informatics Summit that are helping guide our approach to using AI and digital tools to better support families managing chronic health conditions.

Earlier this month, I had the opportunity to attend and present at the AMIA Informatics Summit 2025, one of the most influential conferences in health informatics. The event brought together leading researchers, clinicians, and industry innovators to explore the latest advancements in AI, machine learning, and digital health solutions.

Attending this summit was essential for Adhera Health as we continue advancing our mission in digital health innovation. Our goal is to develop Precision Digital Companions that leverage AI, machine learning, and Social Determinants of Health (SDoH) to improve health outcomes—especially for families managing pediatric chronic conditions, a group often overlooked in digital health innovation. The discussions at AMIA reinforced the importance of fairness in AI, ethical deployment of machine learning models, and the critical role of SDoH in shaping equitable healthcare solutions.

Key Takeaways from the Summit

1. Fairness in AI: Addressing Bias in Machine Learning for Chronic Disease Care

One of the most pressing challenges in AI-driven healthcare is the issue of bias in machine learning models. During my presentation, I highlighted how current AI models often favor younger, male patients, leading to disparities in chronic disease care. These biases, embedded in training data, can result in less accurate predictions for older adults, women, and underrepresented populations—widening healthcare inequalities.

The session sparked thought-provoking discussions with attendees on ways to mitigate these biases, including:

1.         Rebalancing training datasets to ensure diverse representation.

2.         Developing fairness-aware algorithms that actively correct disparities.

3.         Integrating SDoH data to capture a more holistic view of patient health.

We also discussed how fairness in AI extends beyond individual patients to their families—particularly in pediatrics, where caregivers play a central role in daily disease management. Supporting equitable outcomes means considering not just clinical risk factors, but also the emotional, logistical, and behavioral burden placed on families.

2. Engaging with Leading Experts in Biomedical Informatics

One of the most valuable aspects of AMIA is the opportunity to connect with top researchers and thought leaders. I had insightful conversations with experts from UCSF, Stanford, NYU Langone, and Mayo Clinic on cutting-edge developments in AI and health informatics, particularly in:

1.         The potential and risks of Large Language Models (LLMs) in healthcare. While LLMs hold promise for patient communication and clinical decision support, concerns around hallucinations and bias remain critical.

2.         Ethical considerations in AI deployment. How can hospitals ensure that AI-driven decisions align with patient-centric care?

3.         Advancing NLP and machine learning in EHRs. How can we leverage AI to extract meaningful insights from electronic health records while maintaining privacy and fairness?

These discussions reinforced the importance of responsible AI development and the need for cross-sector collaboration to ensure these technologies truly benefit patients.

3. Spotlight on SDoH and AI Integration

A key theme throughout the summit was the integration of SDoH in AI models. In a discussion with a researcher from a leading children's hospital, we explored the role of SDoH data in family-centered precision digital health solutions for pediatric chronic conditions—a critical yet often overlooked aspect of AI in healthcare. These conversations reinforced the importance of designing tools that recognize caregivers as active participants in care and address the real-world challenges families face, such as housing instability, food insecurity, and access to transportation.

By incorporating factors like socioeconomic status, access to care, and environmental conditions, AI-driven tools can offer more personalized and equitable healthcare recommendations. This aligns directly with Adhera Health’s mission to create digital companion solutions that account for real-world patient contexts beyond clinical data alone.

Looking Ahead: Next Steps for Adhera Health

Insights from AMIA will directly inform the evolution of Adhera Health’s Precision Digital Companion programs. Specifically, we are focused on:

1.         Enhancing AI fairness in our predictive models to reduce bias in chronic disease care.

2.         Expanding partnerships with healthcare institutions and researchers to drive responsible AI innovation.

3.         Continuing to refine our support for family dyads—both the child and caregiver—to ensure our digital companions foster confidence, resilience, and health literacy at home.

4.         Further integrating SDoH data to improve personalization and patient engagement in digital health solutions.

We are excited about upcoming opportunities to collaborate with leaders in AI fairness, biomedical informatics, and digital health to ensure these innovations lead to more just and effective healthcare solutions.

Conclusion

Attending AMIA Informatics Summit 2025 was a powerful reminder of the potential—and responsibility—of AI in healthcare. From addressing bias in machine learning to leveraging SDoH for more equitable family-centered care, the discussions and collaborations at this event will shape the next phase of digital health innovation.

At Adhera Health, we are committed to leading this change. If you're interested in collaborating on AI fairness, SDoH integration, or digital health solutions, let’s connect!

Stay tuned for an upcoming research publication where we dive deeper into these topics. In the meantime, follow Adhera Health for more updates on our work in responsible AI and precision digital health.