Using AI to support a nursing program for patients with advanced illness
A case study of Houston Methodist Coordinated Care
Overview
Houston Methodist Coordinated Care (HMCC) is the largest Medicare ACO in Houston, Texas with approximately 54,000 attributed Medicare patients.
The Problem
Traditional methods for advanced care planning are often diagnostic-driven or rules-based, random, and reactive, and EHR time-intensive, hampering the care team’s efficiency. Patients in this category represent 4% of the Medicare population but account for 25% of Medicare costs in the last year of life.
AI-Enhanced Solution
HMCC deployed HDAI’s AI-enhanced predictive models, embedded directly in their EHR and web portal, to:
- Predict likelihood of hospice need within 90 days
- Stratify patients by actionability
- Generate targeted chart summaries to reduce clinician burden
- Accelerate referrals to palliative and hospice care
Results and Impact
- Earlier Hospice Referrals: Patients identified earlier, increasing appropriate hospice enrollment
- Cost Reduction: Participants showed lower per-beneficiary costs and lower admission rates compared to non-participants
- Quality Improvement: Higher patient and family engagement, improved satisfaction with end-of-life care
- Efficiency Gains: Reduced clinician EHR time and workload through targeted summaries
Key Takeaways and Future Direction
The AI-enhanced approach delivers lower costs through smarter interventions, better quality with higher patient satisfaction, and stronger provider networks. Future plans include expanding intelligent risk triaging for all populations, embedding real-time risk insights into care workflows, and enhancing chronic disease management through AI integration.
