AI Chatbot Helps Elderly Patients Set Goals Before Doctor Visits
Researchers at UTHealth Houston have received a $1.3 million grant to develop a conversational AI platform. This tool will help older adults with multiple chronic conditions identify and communicate their health priorities to their physician, making care more personalized and effective.
How AI Helps Elderly Patients Find Their Voice: An Analysis of the MyHealthPriorities AI Project
Introduction
"Patients and caregivers rarely get the opportunity to define their health priorities together with their clinicians. Patients are seldom asked what their health goal is and whether a potential treatment can help them achieve that goal."
This quote comes from Aanand Naik, MD, professor at UTHealth Houston School of Public Health and principal investigator of a new ambitious project. It accurately reflects a systemic problem in modern medicine: given time constraints and increasing specialization, the patient's voice—their true desires and priorities—is often lost. Treatment becomes "by protocol" rather than "by needs."
On April 17, 2026, UTHealth Houston announced a $1.3 million grant from the John A. Hartford Foundation to develop the MyHealthPriorities AI platform—a conversational AI tool that helps older adults with multiple chronic conditions articulate their priorities and communicate them to their doctor.
This project is not just about technology. It is about fundamentally rethinking how healthcare should be delivered, especially for the most vulnerable patient group—older adults with comorbidity.
Event Details and Timeline
On April 14, 2026, the Board of Trustees of the John A. Hartford Foundation approved $1,358,369 for an 18-month project led by the UTHealth Houston team under Dr. Naik. The grant is part of a larger $11.4 million foundation initiative to implement age-friendly approaches in healthcare and public health.
How the MyHealthPriorities AI Platform Works:
- Pre-Visit Preparation: The patient (or their caregiver) interacts with an AI-based conversational agent. The tool uses a structured, guided script to identify the patient's key health priorities.
- Note Generation: The system converts the patient's responses into a concise "health priorities note."
- Integration into the Visit: This note becomes the basis for dialogue between the patient and physician, guiding decision-making and ensuring alignment between the patient's goals and the proposed treatment.
Pilot testing of the platform will be conducted in partnership with America's Physician Groups at two Medicare managed care practices—care delivery models focused on older adults.
This is not UTHealth Houston's first step in this direction. As early as 2024, the university's Institute on Aging worked on improving the Patient Priorities Care website—a previous version of the model that allowed patients to set priorities. Plans at that time included integration with electronic medical records and support for multicultural translation. The current project is a logical evolution: adding AI for more realistic interaction and scalability.
Impact and Significance
For Patients and Their Families
The key problem the technology addresses is fragmentation of care. An older patient with multiple chronic conditions inevitably sees different specialists: a cardiologist treats the heart, a neurologist treats cognitive impairment, an endocrinologist treats diabetes. Each follows their own protocols. The result is a complex, sometimes contradictory treatment plan. But who asks the patient: "What is more important to you—living two years longer or maintaining the ability to walk independently in the morning? Are you willing to tolerate side effects for an extra month of life?"
MyHealthPriorities AI gives the patient time to consider these questions before the visit, in a calm home environment. Early results show impressive effects:
- Patients and their caregivers feel more engaged and heard.
- Patients report a lower treatment burden.
- Treatment aligns better with what truly matters to the patient.
For the Healthcare System and Physicians
From the physician's perspective, the benefits are equally significant. Time constraints are a reality in modern outpatient practice. Discussing priorities requires time, empathy, and facilitation skills that many may lack. The AI agent handles the "preliminary reconnaissance," delivering a ready-made structured document to the physician.
Just as in the MIT Talk2Care study, where an AI assistant for older adults "significantly saved time and effort for healthcare providers," MyHealthPriorities AI frees up the physician for what truly matters: discussion, plan adjustment, and building trust.
The project directly aligns with the principles of Age-Friendly Health Systems, an approach promoted by the John A. Hartford Foundation, based on four key elements (the "4Ms"): What Matters, Medication, Mentation, and Mobility. MyHealthPriorities AI works specifically with the first, most fundamental element—What Matters.
Reactions from Key Stakeholders
The John A. Hartford Foundation, by awarding the grant, bet on a scalable solution. As stated in the grant description, the goal is "to test whether patient goals can be reliably captured and tracked using goal attainment measures" and to assess "real-world feasibility, acceptability, and clinical integration while incorporating robust safeguards for safe and responsible use of AI in the care of older adults."
UTHealth Houston positions the project as part of its broader strategy to use technology for personalized geriatric care. Rafael Samper-Ternent, co-investigator of the project, noted in 2024: "Technology opens new opportunities to provide age-appropriate care, and we want to take advantage of them."
America's Physician Groups (APG) serves as a critical implementation partner. APG will provide access to two real-world Medicare managed care practices where the technology will be tested in actual clinical settings. This transforms the project from a lab development into a pilot implementation with potential for scaling.
Researchers at MIT, who worked on a similar system called Talk2Care, also confirm the relevance of this direction: their interviews with older adults and providers showed that LLMs can "enrich health information collected from older adults" and facilitate asynchronous communication.
Forecast and Conclusions
The MyHealthPriorities AI project marks a shift in how we think about AI in medicine. Instead of focusing on diagnosis (where AI is often compared to physicians) or administrative burden, this tool addresses a more subtle challenge: restoring patient agency.
According to the University of Pennsylvania, about 5% of adult outpatients in the US suffer from diagnostic errors, and older adults are at even higher risk. MyHealthPriorities AI does not diagnose diseases. It diagnoses values. And that is its uniqueness.
Long-term Forecast (3-5 years):
- If pilot testing in two Medicare practices proves successful, the technology will scale rapidly. Given the support from the Hartford Foundation and the approval status, we can expect expansion to dozens and hundreds of geriatric practices nationwide.
- Integration with electronic medical records (EMRs), planned from the outset, means the priorities note will become an integral part of the medical record, accessible to all specialists involved in the patient's care.
- A new quality metric for healthcare may emerge: "goal alignment"—the consistency of treatment with patient priorities. This could change payment models, incentivizing not just adherence to protocols but achievement of outcomes meaningful to the patient.
Key Takeaway: The future of medicine lies not only in gene therapies and breakthrough drugs. It lies in bringing the human face back into a system where technology serves not to replace but to enhance human interaction. MyHealthPriorities AI is an attempt to answer a question rarely asked but meaningful to everyone: "What truly matters to you in your health?"
— Editorial Team
