Every 20 seconds, someone in the world loses a limb to diabetes. In the United States alone, more than 150,000 diabetes-related amputations happen every year — a number that has climbed steadily even as our ability to prevent it has improved. Five-year mortality after a major amputation tops 50%, rivaling most cancers. And rates vary as much as tenfold across U.S. regions — driven not by how sick people are, but by systems of care: access to vascular and podiatric services, food security, and the density of primary care.
The evidence has been sitting in our journals for over a decade. The problem was never the data. It was getting that data in front of the congressional staffers writing health policy, the administrators allocating resources, and the community health workers guiding patients through fragmented systems.
So Caitlin Hicks and I tried something — and it is now published in Surgery.
In under a week, much of it between patients in clinic and the operating room, I used a conversational AI system to build three interactive, data-driven web platforms that translate 15 years of published evidence on amputation into something the public can actually explore. No development team. No technology grant. No formal training in software engineering. The method now has a name — “vibe coding”: you describe what you want in plain language and iterate with the model until the output does the job. My role was not writing code. It was editorial direction — deciding what to emphasize, what to caveat, and what to make navigable.
The three platforms span scales
- The ZIP Code Lottery — a long-form data-journalism narrative showing how limb-loss risk tracks your address more than your disease, with interactive maps of Los Angeles County.
- The National Amputation Heat Map — amputation rates by U.S. state and congressional district, with a “Find Your Representative” tool that shortens the distance between seeing a number and doing something about it.
- Every 20 Seconds: The Global Limb Threat Index — a composite index across more than 130 countries, surfacing that systems of care, not disease severity, drive much of the variation in who keeps their leg.
What still stops me is the compression. Work like this used to require a funded grant, a team, and a timeline of months. We built three public platforms in under a week at essentially no direct cost — months to days, tens of thousands of dollars to roughly zero, a development team to one clinician and an AI. If a surgeon at a safety-net hospital can do that, what about the thousands of clinicians sitting on actionable data that never reaches the people who need to see it?
We are candid about the limitations. AI-generated code is not always optimal, anything touching patient-level data needs serious security review, and the very speed of building persuasive visualizations raises the stakes for rigorous validation. The bottleneck in health communication may no longer be technical capability — it may be editorial judgment. Which is exactly why the people who understand both the data and what it means for real patients should be the ones building these tools.
Vibe coding does not solve the amputation crisis. But it removes one of the most stubborn barriers between the people who understand a health emergency and the people with the power to act on it.
Read the full article authors.elsevier.com/c/1nOm-90L-76CN
Armstrong DG, Hicks CW. Vibe coding a silent crisis: How conversational artificial intelligence built 3 health advocacy platforms in under a week. Surgery. 2026;110397. doi:10.1016/j.surg.2026.110397
#ActAgainstAmputation #VibeCoding #AmputationPrevention #LimbPreservation #DiabeticFoot #ConversationalAI @CaitlinWHicks @KeckSchool_USC @ALPSLimb
Leave a Reply