Computable Diagnosis as a Moral Imperative

Dr. Harold H. Hines, Jr Professor, Yale University; Director, Yale New Haven Hospital Center for Outcomes Research and Evaluation; Editor-in-Chief, Journal of the American College of Cardiology (JACC) recently published an article entitled Computable Diagnosis as a Moral Imperative1.

He makes the point that the medical community must transition from relying on individual clinical “hunches” to implementing computable diagnosis—the systematic, population-level use of existing electronic health data to identify undiagnosed patients with treatable diseases.

He goes on to advocate for

• Move from Passive to Real-time Data: Clinical data is currently “passive,” stored in electronic filing cabinets. Computable diagnosis makes this data “active,” scanning years of records to connect distributed findings like EKGs, biomarkers, and prior procedures.

• Move beyond Clinical Decision Support (CDS): Unlike traditional alerts that fire during a patient encounter and are often ignored, computable diagnosis operates at the population level, identifying at-risk individuals continuously, regardless of whether they are currently sitting in a doctor’s office.

• Use the power of AI-Enabled healthcare machine data: Advances in AI can now identify risk from EKGs and ultrasounds months or years before clinical symptoms appear, using patterns invisible to the human eye.

• Eliminate the “Choosing Not to Know” Barrier: Some systems resist these tools because they fear overwhelming existing infrastructure (e.g., echo labs). The author argues that rejecting diagnosis due to operational “inconvenience” is an acceptance of preventable suffering.

• Remove Equity Risk: If these advanced diagnostic tools are only deployed in well-resourced, integrated systems, it will widen the gap in care for patients in safety-net hospitals or fragmented environments.

We could not agree more.

The Pediatric Moonshot mission is to reduce healthcare inequity, lower costs and improve outcomes for children rurally, locally and globally - by creating privacy preserving, real-time AI agents based on access to data in all 1,000,000 healthcare machines in all 500 children’s hospitals in the world. Building accurate real-time AI applications requires large amounts of diverse data.

The traditional method of aggregating the data at a central site will not work in healthcare and life sciences.
The data sizes are much larger (ultrasounds are 1TB), security requirements much stricter, and privacy requirements more stringent. To meet these demands, we have developed a “new rocket”, real-time, privacy-preserving distributed AI cloud infrastructure that moves AI applications to the data-whether that’s in the hospital, research labs, clinics, imaging centers or ambulance.

Based on this infrastructure, we are building disease-precise AI Agents - that enable systematic, population-level use of existing electronic health data to identify undiagnosed patients with treatable diseases – what the author calls computable diagnosis.

We have started our work in nephrology with 5 rare kidney diseases, which is, of course, a small subset. We’ve identified 99 pediatric cancer diseases and 104 pediatric cardiology conditions (which we believe to be complete) and are beginning to turn his vision into a reality.