Healthkathon 2025 4-Day Sprint
JKN-Flow: Hospital Management System
Integrated hospital management app with room allocation, electronic medical records, billing, rule-based BPJS claim fraud checks, and Gemini-powered chatbot.
Role Full-Stack Developer (Top Contributor)
Release date 2025-11
Category Web, AI, Healthcare
Technology
Laravel 11 PostgreSQL Tailwind CSS Alpine.js Gemini API
Project Overview
JKN-Flow was developed during an intensive 4-day sprint for Healthkathon 2025. It streamlines hospital administrative workflows for patients under Indonesia’s national health insurance (JKN/BPJS Kesehatan).
The system spans patient admission, ward allocation, electronic medical records, pharmacy inventory, billing, and automated fraud-scoring audits for BPJS claims.
Problems
- Long Physical Queues: JKN patients frequently spend hours waiting across service desks for referral checks and room availability.
- Risk of Claim Rejection: Manual claim checks are susceptible to diagnosis code mismatches (ICD-10/ICD-9-CM) or procedure duplication, delaying hospital reimbursements.
- Limited Patient Information: Patients and families struggle to access immediate guidance regarding clinic schedules and room vacancies.
Target Users
- Patients and Families: Queuing online and inquiring about referral procedures through a virtual assistant.
- Medical and Administrative Personnel: Documenting clinical notes, managing ward transfers, and verifying medication stock.
- Hospital Claim Verifiers: Auditing claim validity before submission to minimize claim disputes.
Technical Decisions
- Performant Monolith with Laravel 11: Chose Laravel 11 with PostgreSQL for rapid iteration during the hackathon, supported by strict relational schemas for medical histories.
- Lightweight Frontend (Tailwind CSS + Alpine.js): Fast UI response without heavy bundle overhead on clinic workstations or mobile devices.
- Rule-Based Claim Fraud Scoring Engine: Custom backend module auditing procedure codes, length of stay, and medication combinations against BPJS claim regulations.
- AI Patient Assistant via Gemini API: Interactive chatbot guiding patients through healthcare facility tiers with natural language understanding.
Challenges
- Multi-Table Relational Schema: Building an interconnected database architecture (rooms, patients, physicians, records, bills, claims) and functional interfaces in 4 days.
- Complex Claim Validation Rules: Translating statutory healthcare verification logic into automated checks without generating false positives during emergency admissions.
Measurable Results
- Comprehensive End-to-End Delivery: Completed 6 core hospital modules (Admission, EMR, Wards, Pharmacy, Billing, Claim Verification) within the 4-day hackathon sprint.
- Claim Anomaly Detection: Rule-based verification filters anomalies prior to official BPJS verifier review.
- Instant Response: Integrated AI chatbot provides immediate administrative procedural guidance.