QueueCure '26 – Patient-First Hospital Queue Platform

Wooble Full Stack Hackathon
A distributed, real-time hospital queue management system built with a microservices architecture to solve the OPD coordination crisis in India.

---

🏥 The Problem

India's outpatient department (OPD) struggle isn't a lack of doctors—it's a lack of coordination. 76% of clinics still use paper tokens, leading to 2-3 hour wait times and zero visibility for patients. QueueCure replaces chaos with a unified, real-time pipeline for both online bookings and physical walk-ins.

---

🏗 System Architecture

QueueCure is built using a Decoupled Microservices Architecture designed for high availability and real-time synchronization.

graph TD
    A[Patient App - React] --> G[API Gateway / Ingress]
    B[Receptionist Dashboard - React] --> G
    
    G --> C[Auth Service :3002]
    G --> D[Hospital Service :3001]
    G --> E[Queue Service :3003]
    G --> F[Triage Service :3004]
    
    C & D & E --> DB[(Supabase Cloud)]
    E <--> R[(AWS Elasticache - Redis)]
    F --> AI[Groq/Claude AI API]
    
    R -- SSE Stream --> A
    R -- SSE Stream --> B

Loading

Tech Stack

• Frontend: React (Vite, TypeScript, Tailwind CSS)
• Backend: Node.js, Express (Microservices)
• Database: Supabase (PostgreSQL)
• Real-time/PubSub: AWS Elasticache (Redis)
• AI Engine: Groq (Llama 3.3) for Triage
• Containerization: Docker (Single-stage)
• Orchestration: Kubernetes (HPA, ConfigMaps, Secrets)

---

📡 Real-time Sync Flow (SSE + Redis)

Unlike traditional polling, QueueCure uses Server-Sent Events (SSE) backed by Redis Pub/Sub to ensure instant updates across all devices.

Receptionist [Call Next] 
       │
       ▼
POST /queue/call-next (Queue Service)
       │
       ├─► Update DB (Supabase)
       ├─► Recalculate Doctor Lag
       └─► Redis PUBLISH "queue:{doctor_id}" { type: 'CALL_NEXT', ... }
                               │
       ┌───────────────────────┴───────────────────────┐
       ▼                                               ▼
SSE Endpoint (:3003/stream)                     SSE Endpoint (:3003/stream)
[Connected Patient Pod A]                       [Connected Patient Pod B]
       │                                               │
       ▼                                               ▼
UI Update: "You're Next!"                       UI Update: "Position: 1"

---

🧠 Core Algorithm: ETS (Estimated Time of Service)

The ETS isn't just a guess; it's a dynamic calculation based on real-time doctor performance.

Formula: ETS = (Tokens Ahead × Avg Consult Time) + (Tokens Ahead × 2m Buffer) + Doctor Lag

Doctor Lag Calculation: Lag = Rolling Average of (Actual Session Time - Estimated Session Time) As the doctor runs late or ahead, the entire queue's ETS shifts automatically and notifies patients via SSE.

---

🚀 API Endpoints

🔑 Auth Service (:3002)

• POST /auth/login - Handles Patient OAuth and Receptionist Credentials.
• POST /auth/verify - Validates JWT for cross-service authorization.

🏥 Hospital Service (:3001)

• GET /hospitals - Discovery with filters for city, specialty, and rating.

📋 Queue Service (:3003)

• POST /tokens - Create online or walk-in tokens.
• GET /queue/stream?doctorId=X - SSE endpoint for live queue updates.
• POST /queue/call-next - Advance the queue & trigger Redis broadcast.
• POST /tokens/switch-suggest - AI-powered hospital switching.

🤖 Triage Service (:3004)

• POST /triage - AI symptom analysis (Routine/Urgent/Emergency).

---

🧪 Real-World Validation (40/40 Score)

The system has been validated against 20 critical Indian hospital scenarios:

ID	Scenario	Result
TC-001	Morning Rush - 50 simultaneous bookings handled without collisions.	✅ PASS
TC-003	Emergency - Chest pain cases auto-jump to position 1.	✅ PASS
TC-004	Doctor Lag - ETS adjusts dynamically as consultation time drifts.	✅ PASS
TC-011	Capacity - Queue limits prevent overcrowding (Max 50/doctor).	✅ PASS
TC-013	Privacy - PII (Names/Phones) restricted in public queue views.	✅ PASS
TC-014	Load Test - 500 concurrent users handled via AWS Elasticache.	✅ PASS

---

🛠 Deployment

Local (Docker Compose)

Important: Since we are using AWS Elasticache and no local Redis container, you MUST update the REDIS_URL in your .env file to your actual Elasticache endpoint. Note: Using localhost or 127.0.0.1 will fail inside Docker.

docker-compose up --build

Kubernetes (Production)

# Apply Base Namespace & Configs
kubectl apply -f micro-k8s/base.yaml

# Apply Microservices
kubectl apply -f micro-k8s/services.yaml

---

🔒 Security & Privacy

• HMAC Tokens: QR codes are signed with HS256 to prevent tampering.
• Data Sanitization: Public APIs never leak Patient Phone/Gov IDs.
• TLS Redis: Standardized connection logic for encrypted AWS Elasticache clusters.

---

Built by Team QueueCure for Wooble Hackathon '26