How Edge Medicine Works: Open-Source Health AI Infrastructure

Today’s healthcare remains siloed, fragmented, episodic, and reactive.

75% of Americans are living with at least one chronic condition, such as heart disease, cancer, diabetes, or kidney disease. These conditions are 24/7, occurring where people live, work, and play.

We should be preventing chronic disease, yet most of our effort goes toward treating it after the fact. Chronic conditions drive 90% of the U.S.'s $5 trillion in annual healthcare costs, yet billions of dollars continue to go toward apps, companies, and technologies that address one disease at a time, building isolated solutions. But patients rarely live with just one condition. Research has consistently shown that the greatest impact on health outcomes comes from shifting care upstream to prioritize prevention and management of chronic diseases outside traditional clinical settings (Schroeder).

Edge Medicine is designed to be

01Human centered and privacy preserving
02Powered by streamlined data and compute
03Clinically validated and built into care workflows

Breaking Down Silos with Open Source Infrastructure

Through JupyterHealth, we are bringing the proven Jupyter ecosystem, used by millions of scientists worldwide, into everyday clinical care at the Edge. JupyterHealth combines data interoperability with the analytics power of Project Jupyter to create a scalable, interactive foundation for healthcare transformation.

JupyterHealth is an open source technology stack that connects patient generated health data from multiple sources directly with the health system to support care. In place of fragmented, single purpose solutions, it offers open, standardized infrastructure that lets data and AI move freely from home to health system and back. The stack provides the core tools developers need to create Edge Medicine solutions, enabling teams to concentrate on clinical impact rather than redundant infrastructure development.

Explore JupyterHealth documentation and code

Across the edge, and back

Diagram: signals from connected devices flow up through JupyterHealth and Patient-facing Apps / Clinician-facing Portals to provide coordinated, continuous, contextual care provided at the Edge

Signals from the Edge

Devices collecting data are inputs that provide health signals through continuous data streams, allowing for a complete picture of health to be a click away for health teams.

JupyterHealth

Infrastructure layer that makes Edge Medicine possible. Facilitates data integration from multiple sources (e.g., wearable devices), Epic integration, analytic tooling and algorithms to support standardization and clinical insights. Learn more about JupyterHealth

Patient-facing Apps

Holistic single view of personal health history (e.g., Medical records, Labs, Current medications) integrated with health systems and healthcare providers.

Clinician-facing Portals

Provider and patient generated cohesive view allowing for cross-specialty coordination and real-time monitoring of condition progression.

The Edge

The true frontline of care where health is lived continuously in everyday life.

Alex is prompted to download the UCSF EdgeMed app, and his medical records from other institutions sync automatically.

An AI model trained on millions of electronic health records assigns Alex a risk score, and the system automatically orders his labs.

A UCSF clinical team member sets Alex up with a continuous glucose monitor and a contactless, smartphone-based blood pressure sensor. His Oura Ring data syncs as well.

UCSF's AI model projects a concerning health trajectory.

Instead of waiting five months for a first primary care visit and another three months for a follow-up, Alex is seen within a week. His care team can review his sensor data, and AI helps triage him to subspecialty care.

The iCKM collaboration portal lets specialists and primary care physicians model and discuss treatment options together.

A case across the Edge Medicine continuum

Meet Alex. He is 37 and for years had rising blood sugar, high blood pressure and weight gain. His warning signs were missed in today’s system. Alex has a heart attack at age 48. He has Cardio-Kidney-Metabolic syndrome, or CKM, which is now the single biggest driver of chronic disease burden.

Click through Alex's story (on the right)(below) to see how his care changes with Edge Medicine, built on JupyterHealth.

Next steps

Get involved

Developers & researchers Build on JupyterHealth Open-source infrastructure you can deploy, extend, and contribute to. Explore the docs Students, founders & innovators Bring your idea Pitch how you would use JupyterHealth to change care at the Edge. Applications are due November 8. Innovation Challenge Health systems, clinicians & funders Partner with us Pilot Edge Medicine with your patients, or help fund what comes next. Get in touch

Built by a team with deep expertise across medicine, software, and computation. Meet the team