AI in Healthcare
Bridging Healthcare Accessibility Through Intelligent Doctor Matching and Community-Centered AI
Abstract
This research explores the rapid development of "Baymax," an AI-powered healthcare accessibility system designed to bridge critical gaps in doctor-patient connections within underserved communities. Built over a single weekend (Friday night through Sunday), this project demonstrates the potential for AI to transform healthcare navigation through intelligent doctor matching using the National Provider Identifier (NPI) system.
Despite having to halt development due to potential competition with existing work initiatives, the project successfully produced a working demo that validated the core concept. The system allows cities to add their preferred hospital URLs, uses NPI data to find and pull doctor profiles, enabling users to call or book appointments directly—solving the simple but critical problem of avoiding emergency room redirects.
⚡ Weekend Build: This functional healthcare navigation system was conceptualized, developed, and tested in just one weekend, demonstrating the rapid potential of AI-driven healthcare solutions when properly focused.
1. The Spark: A Mayor's Question
After showcasing AI-Powered Community Assistants to the mayor of Mount Vernon, he paused and asked:"Can this be done for healthcare to help people find doctors?"
This simple question sparked an intensive weekend development sprint. From Friday night through Sunday, we built a working prototype that addressed critical gaps in healthcare accessibility. The demo successfully validated our approach, showing how AI could help community members navigate complex healthcare systems, overcome language barriers, and find appropriate medical care.
Project Status: Development was halted when leadership identified potential competition with existing organizational initiatives. However, the working demo proved the concept's viability and highlighted the transformative potential of dedicated engineering resources.
Weekend Challenge
- • 48-hour development window
- • Solo development with limited resources
- • Rapid prototyping and testing approach
- • Proof-of-concept validation requirements
Demo Success
- • Working AI doctor matching system
- • Functional user interface
- • Real-time healthcare provider search
- • Proof-of-concept validation
Lessons Learned
- • AI can rapidly solve complex healthcare navigation
- • Community-centered design is achievable at scale
- • Weekend builds can validate major concepts
- • Full team potential is exponentially greater
2. The Real Problem: Healthcare is Intimidating
To be completely honest: I don't even know what insurance I have.
I have a pile of 8 cards that look very similar, and I have no idea which one is the most recent. As a software engineer—someone who builds complex systems for a living—I found the insurance terminology and healthcare system so intimidating that I joked I might need a 2-year degree just to understand it all.
The Insurance Problem: If understanding insurance is this difficult for a software engineer, imagine how overwhelming it is for everyone else. That's a whole other problem that needs to be solved with AI—but it wasn't what we built this weekend.
What We Actually Solved
Instead of tackling the insurance complexity nightmare, we focused on a simpler but critical problem:How do you find the right doctor without ending up in the emergency room only to get redirected somewhere else?
The Solution: NPI-Based Doctor Discovery
- • Cities add their preferred hospital URLs - Local governments can input their trusted healthcare facilities
- • NPI (National Provider Identifier) integration - We use the official NPI system to find and pull doctor profiles
- • Direct contact information - Users can call or book appointments directly with the right provider
- • Avoid the ER redirect loop - Find the appropriate care provider before making the trip
Note on Taxonomy: We're still working to fully understand medical taxonomy classifications. It's complex, but the NPI system provides enough structure to match patients with appropriate providers based on their needs.
3. Baymax: Weekend Prototype to Production Vision
Named after the healthcare companion from Big Hero 6, Baymax began as a weekend experiment that quickly evolved into a functional prototype. The system leverages the National Provider Identifier (NPI) database to help users find appropriate healthcare providers, pulling real doctor profiles and contact information so people can directly call or book appointments.
Weekend Development Timeline
Friday Night
- • Concept validation and research
- • Core architecture planning
- • Initial AI model integration
Saturday
- • Doctor matching algorithm development
- • Multilingual interface implementation
- • Healthcare provider database integration
Sunday
- • Demo preparation and testing
- • User interface refinement
- • Proof-of-concept validation
Core System Features
NPI-Based Doctor Discovery
- • National Provider Identifier (NPI) database integration
- • City-managed hospital URL preferences
- • Doctor profile extraction and display
- • Direct contact information for calling or booking
- • Specialty matching based on medical needs
Location & Accessibility
- • Geospatial analysis for proximity matching
- • Distance-based provider sorting
- • Location mapping and visualization
- • Geographic search radius customization
Baymax System Capabilities
Patient-Centered Features
- • Natural language symptom description
- • NPI-based provider search and discovery
- • Location-based recommendations
- • Direct access to doctor contact information
- • User-friendly interface design
City & Provider Integration
- • City-managed hospital URL preferences
- • NPI database connectivity
- • Specialty and credential display from NPI data
- • Contact information for direct booking
- • Provider profile presentation
4. System Architecture & Workflow
The Baymax system architecture demonstrates a practical approach to healthcare navigation through NPI database integration, city-managed hospital preferences, and user-centered design principles.
4.1 AI Healthcare Navigation Workflow
The core system processes user health inquiries through various AI models to provide personalized doctor recommendations and healthcare navigation assistance.
User Input
"My brother has a high fever and headache"
Chain-of-Thought Analysis
Parallel Data Sources
NPI Registry
Web Search
AI Processing
Unified Response
AI Model Integration
- • Natural language processing for symptom analysis
- • Medical specialty matching algorithms
- • NPI database query and filtering
- • Geospatial analysis for location-based recommendations
Healthcare Data Integration
- • National Provider Identifier (NPI) database access
- • City-managed hospital URL repository
- • Provider specialty and credential data from NPI
- • Location and geographic data mapping
- • Direct contact information extraction
5. Proof of Concept: Weekend Demo
Watch the working prototype in action. This demo showcases the Baymax system built over a single weekend, demonstrating NPI-based doctor discovery, city-managed hospital integration, and AI-powered healthcare navigation.
Demo Highlights
- • Real-time doctor search using NPI database integration
- • User-friendly interface for healthcare navigation
- • Location-based provider recommendations
- • Direct access to doctor contact information
- • Proof that complex healthcare problems can be solved rapidly
6. Ethical Decision: Respecting Organizational Boundaries
Following the successful demonstration of the working prototype, my supervisor informed me that our organization is planning to build something similar and innovate with AI in the healthcare space. While I have no knowledge of the specific details or timeline of those initiatives, I made the immediate decision to halt further development of Baymax.
This decision stems from my core values and professional principles. Some features were removed and hardcoded in the live demo out of respect for my workplace and to ensure there is no potential competition with organizational initiatives.
Looking Forward: While this particular project has been paused, I remain deeply committed to healthcare accessibility innovation. Perhaps in the future, I could contribute to whoever builds something like this, or if my organization does pursue AI innovation in healthcare. For now, this project stands as a proof of concept and a testament to what can be achieved when technology is focused on solving real community healthcare challenges.
Removed Functionality
Some features were removed and hardcoded to respect organizational boundaries and avoid any potential competition with company initiatives:
Advanced data source configuration and AI integration settings were part of the original prototype
What Remains in Demo
- • Basic NPI-based doctor search functionality
- • Location-based provider recommendations
- • User interface and navigation
- • Core healthcare navigation concept
- • Proof of weekend development capability
Removed Features
- • Advanced data source configuration
- • Multiple AI integration options
- • Custom search provider settings
- • Priority-based source management
- • Extended healthcare database connections
7. Future Vision: From Weekend Prototype to Healthcare Revolution
While the Baymax project was halted due to organizational considerations, the weekend prototype validated a powerful concept: AI can rapidly transform healthcare accessibility when properly focused on community needs. The working demo serves as proof that complex healthcare navigation challenges can be solved with the right approach and adequate resources.
The project's success in such a compressed timeframe demonstrates the exponential potential when dedicated engineering teams, healthcare partnerships, and community integration support are properly aligned toward solving real healthcare accessibility challenges.
Proven Weekend Achievements
- • Functional NPI-based doctor discovery system
- • City-managed hospital URL integration
- • Location-based healthcare provider search
- • Symptom-to-specialty matching algorithm
- • Direct contact information access
- • Working demo validation
Full Team Potential
- • Production-scale healthcare system integration
- • Advanced AI models for health prediction and triage
- • Comprehensive healthcare provider networks
- • Multi-city deployment and scaling capabilities
The Exponential Engineering Effect
If one developer could build a working healthcare navigation prototype in a weekend, a dedicated team of engineers could revolutionize healthcare accessibility across entire communities in a matter of months.
- • Weekend prototype → Full team could deliver production system in 4-6 weeks
- • Solo development → Team collaboration enables 10x feature development speed
- • Basic demo → Professional healthcare partnerships unlock enterprise integration
- • Single community focus → Multi-city deployment becomes achievable at scale
8. Conclusion: The Power of Rapid Healthcare Innovation
The Baymax weekend prototype demonstrates that practical healthcare solutions can emerge from focused, community-centered development approaches. By leveraging the National Provider Identifier (NPI) system and enabling cities to manage their preferred hospital URLs, we created a working system that helps people find the right doctor and avoid unnecessary emergency room visits.
Despite being halted due to organizational considerations, the project successfully validated that AI can rapidly address healthcare navigation challenges when properly applied. The working demo proved that with just a weekend of focused development, we could build a system that pulls real doctor profiles, provides direct contact information, and helps people navigate the confusing healthcare landscape.
This research validates that the future of healthcare accessibility lies not in solving every complex problem at once (like insurance terminology), but in tackling specific, critical pain points—like finding the right doctor before ending up in the wrong place.
Key Research Contributions
- • Proof that NPI-based healthcare navigation systems can be rapidly prototyped
- • Validation of city-managed hospital URL preferences for local healthcare access
- • Demonstration of AI-powered symptom-to-specialty matching in a weekend build
- • Framework for avoiding ER redirects through intelligent doctor discovery
- • Evidence of the exponential potential with dedicated engineering resources
The Weekend That Could Have Changed Healthcare
Explore how a single weekend prototype demonstrated the transformative potential of AI-powered healthcare navigation and imagine what dedicated engineering teams could accomplish for community health accessibility.