From Scratch To Custom EHR Extensions: How Dr. Kenneth Qiu Vibe Coded A Direct Primary Care Mega-App

When Dr. Kenneth Qiu opened his Direct Primary Care (DPC) practice straight out of residency five years ago, he quickly learned that commercial software vendors rarely deliver the precise tools that nimble independent clinics need. For years, he asked for features that never materialized.

Then, he discovered vibe coding. Despite having zero formal programming background, Kenneth managed to build a fully functioning application ecosystem that integrates directly with his clinical tech stack. As Dr. Qiu shared, "I'm not a programmer. The closest I ever got to software was in the 10th grade we spent a week doing HTML on Notepad and ever since I took some YouTube classes to try to learn Python or whatever and quit within 30 minutes."

Resources:

• 📑 Get the slides
• 🎥 Watch the recording

Identifying the Friction Points

Every single tool Kenneth built started from a simple place: a daily workplace annoyance or a persistent technical limitation. Instead of waiting on corporate tech updates, Kenneth opened up an AI development tool and began to describe his solutions. As Dr. Qiu explained, "These things I've asked vendors for for years I'm just creating in a weekend."

Over a series of weekends, he successfully designed, tested, and implemented four clinical tools:

• A Transparent Pricing Calculator: DPC models rely heavily on transparent, upfront pricing. Kenneth took an Excel spreadsheet containing his practice's top 25 most common laboratory tests and built an interactive public widget for his website. He achieved this by feeding the details into the AI model and explaining that he was working within WordPress.
• A Unified Patient Timeline: DPC medicine is built around continuous, long-term patient relationships rather than isolated, episodic insurance billing. Kenneth was frustrated by having to dig through separate communication logs and chart notes just to piece together a patient's history. His vibe-coded timeline pulls from his messaging application and his electronic medical records to display all communications, office visits, prescriptions, and lab orders in a single interface.
• AI-Driven Problem Lists & Summaries: Kenneth connected his unified database into Google’s Gemini API. The app automatically reviews the entire timeline history to generate an active patient problem list complete with context-specific citations.
• Automated Pre-Visit Summaries: To eliminate the burden of pre-charting before annual physicals, his application automatically drafts a clinical note based on the updated problem list. It then produces a clean, printable patient handout complete with a color-coded health dashboard mapping cardiovascular and metabolic risk factors.

Testing and Scaling Up

Kenneth learned to navigate his tools entirely by using the AI as its own interactive manual, describing it as an automated instruction manual and built-in customer support channel. After mastering local development, Kenneth reached a massive milestone right before the webinar: he asked the AI to convert his single-user local app into a secure, multi-user, multi-computer program that his entire clinical team could access across multiple locations. This rapid progression felt entirely magical, demonstrating how quickly a non-technical clinician can scale custom software infrastructure.