MedChain TN

AI + Blockchain Medical Inventory Management System for Tamil Nadu Healthcare Supply Chain

Problem Statement

Tamil Nadu's healthcare supply chain serving 77 million people suffers from five critical failures:

Problem	Impact
No demand forecasting	Drug shortages during malaria & dengue seasons
Mutable inventory records	Stock manipulation goes undetected
No drug verification	Counterfeit medicines reach patients
Poor expiry management	~18% of batches wasted
No decision support	Reactive, manual procurement decisions

Solution

MedChain TN combines blockchain immutability with AI intelligence to fix every layer of the supply chain.

Blockchain  →  tamper-proof batch records + ownership transfers
AI          →  demand forecasting + anomaly detection + expiry prediction
Mobile App  →  unified interface for manufacturers, distributors, patients

System Architecture

┌─────────────────────────────────────────────┐
│           React Native Expo App             │
│  ┌──────────┬──────────────┬─────────────┐  │
│  │ Patient  │ Distributor  │Manufacturer │  │
│  │ 4 Tabs   │   4 Tabs     │   4 Tabs    │  │
│  └────┬─────┴──────┬───────┴──────┬──────┘  │
└───────┼────────────┼──────────────┼──────────┘
        └────────────┼──────────────┘
                     │ HTTPS / REST + JWT
        ┌────────────▼─────────────┐
        │    Node.js API Gateway   │
        │    JWT + RBAC Middleware │
        └──┬───────────┬───────────┘
           │           │
    ┌──────▼───┐  ┌────▼─────────┐
    │PostgreSQL│  │ Flask ML API │
    │ + Redis  │  │  3 Modules   │
    └──────────┘  └──────────────┘
           │
    ┌──────▼──────────────────┐
    │      ethers.js          │
    │         ↕               │
    │  Ganache Local Chain    │
    │  4 Solidity Contracts   │
    └─────────────────────────┘

Tech Stack

Layer	Technology
Mobile App	React Native + Expo SDK 51
Backend	Node.js + Express
ML Service	Python + Flask
Database	PostgreSQL + Redis
Blockchain	Solidity + Hardhat + Ganache
ML Models	scikit-learn, XGBoost, Prophet, Keras
Maps	react-native-maps
Charts	victory-native
Auth	JWT + expo-secure-store

Repository Structure

/medchain-tn
  /mobile                   ← Expo React Native app
    /app
      /_layout.tsx
      /auth/
      /(patient)/
      /(distributor)/
      /(manufacturer)/
    /components
      /shared/
      /patient/
      /distributor/
      /manufacturer/
    /hooks/
    /services/
  /backend                  ← Node.js API Gateway
    /src
      /routes/
      /controllers/
      /middleware/
      /blockchain/
      /ml/
      /db/
    app.js
  /ml-service               ← Python Flask ML API
    /data
      /raw/
      /processed/
      /synthetic/
    /models/
    /notebooks/
    /src
      /features/
      /training/
      /evaluation/
      /api/
    app.py
  /contracts                ← Solidity Smart Contracts
    /contracts/
      MedAccessControl.sol
      MedBatch.sol
      MedTransfer.sol
      MedAudit.sol
    /scripts/
    /test/
    hardhat.config.js
  /data                     ← Dataset generation
    /synthetic/
      constants.py
      generate_usage.py
      generate_movements.py
      generate_expiry.py
      generate_regional.py
      run_all.py
  /docs/
  docker-compose.yml
  README.md

Getting Started

Prerequisites

node >= 18
python >= 3.10
docker + docker-compose
ganache (desktop or CLI)
expo-cli

1. Clone the Repository

git clone https://github.com/your-org/medchain-tn.git
cd medchain-tn

2. Set Up Environment Variables

cp .env.example .env

Fill in your .env:

# Ganache
GANACHE_RPC_URL=http://127.0.0.1:7545
GANACHE_CHAIN_ID=1337
GANACHE_MNEMONIC="your twelve word mnemonic from ganache ui"
GANACHE_PRIVATE_KEY_ADMIN=0x...
GANACHE_PRIVATE_KEY_MANUFACTURER=0x...
GANACHE_PRIVATE_KEY_DISTRIBUTOR=0x...

# PostgreSQL
POSTGRES_URL=postgresql://medchain:password@localhost:5432/medchain_tn

# Redis
REDIS_URL=redis://localhost:6379

# JWT
JWT_SECRET=your_jwt_secret_here
JWT_EXPIRES_IN=24h

# ML Service
ML_SERVICE_URL=http://localhost:5000

3. Start All Services

docker-compose up -d

This starts: Ganache · PostgreSQL · Redis · Node Backend · Flask ML Service

4. Deploy Smart Contracts

cd contracts
npm install
npx hardhat run scripts/deploy.js --network ganache
npx hardhat run scripts/seed.js --network ganache

5. Generate Datasets + Train ML Models

cd data/synthetic
pip install -r requirements.txt
python run_all.py

cd ../../ml-service
pip install -r requirements.txt
python src/training/train_all.py

6. Start the Mobile App

cd mobile
npm install
npx expo start

Scan the QR code with Expo Go on your device.

📊 Dataset Plan

Three datasets covering Tamil Nadu's pharmaceutical supply chain across 10 districts, 12 drugs, 3 facility types, 24 months.

Dataset	Rows	Purpose
Drug Usage History	~3,500	Demand Forecasting
Stock Movement Log	~5,000	Anomaly Detection
Batch Expiry & Wastage	~2,000	Expiry Prediction

Districts Covered

Chennai · Coimbatore · Madurai · Trichy · Salem
Tirunelveli · Vellore · Erode · Thoothukudi · Kanchipuram

Drugs Covered

Paracetamol · Amoxicillin · Metformin · Ibuprofen · Omeprazole
Amlodipine · Artemether · Ciprofloxacin · Insulin · Salbutamol
Ondansetron · Chloroquine

Disease Season Flags (Tamil Nadu)

Dengue / Vector-borne   Oct–Jan   → Paracetamol, Ciprofloxacin
Malaria                 Jun–Sep   → Artemether, Chloroquine
Respiratory             Dec–Feb   → Salbutamol, Amoxicillin
Gastroenteritis         May–Jun   → Ondansetron, Omeprazole
Diabetes                Oct–Nov   → Metformin, Insulin

⛓️ Smart Contracts

Contracts Overview

Contract	Responsibility
`MedAccessControl.sol`	Role management (Admin, Manufacturer, Distributor, Patient)
`MedBatch.sol`	Drug batch registration + QR hash storage
`MedTransfer.sol`	Ownership transfer between supply chain participants
`MedAudit.sol`	Expiry tracking + immutable audit trail

Network Configuration

Development:   Ganache local   (port 7545, chainId 1337)
Testing:       Polygon Amoy Testnet
Production:    Polygon PoS Mainnet

Key Contract Functions

// MedBatch.sol
registerBatch(drugName, region, quantity, expiryDate, qrHash)
getBatch(batchId)
verifyBatch(batchId) → isValid, status

// MedTransfer.sol
initiateTransfer(batchId, to, quantity, toRegion)
acceptTransfer(transferId)
getTransferHistory(batchId)

// MedAudit.sol
getAuditTrail(batchId)
getExpiringBatches(withinDays)

Running Tests

cd contracts
npx hardhat test --network ganache

ML Service

3 Intelligence Modules

Module 1 — Demand Forecasting

Models:    Random Forest · XGBoost · Facebook Prophet
Endpoint:  POST /predict/demand
Output:    predicted_qty_30d, predicted_qty_60d, predicted_qty_90d,
           confidence_interval, reorder_recommendation

Module 2 — Anomaly Detection

Models:    Isolation Forest · Autoencoder (Keras) · Z-Score
Endpoint:  POST /anomaly/detect
Output:    is_anomaly, anomaly_type, confidence_score,
           flagged_features, recommended_action

Module 3 — Expiry Waste Prediction

Models:    Gradient Boosting · Logistic Regression · Survival Analysis
Endpoint:  POST /predict/expiry
Output:    expiry_risk, probability_of_waste,
           redistribution_recommendation

API Reference

Auth

POST  /auth/login
POST  /auth/logout
GET   /auth/me

Drugs

GET   /drugs
GET   /drugs/:id
POST  /drugs

Blockchain

POST  /batch/register
GET   /batch/:batchId
GET   /batch/verify/:qrHash
POST  /transfer/initiate
POST  /transfer/accept
POST  /transfer/reject
GET   /audit/:batchId

ML Predictions

GET   /predict/demand
POST  /anomaly/detect
POST  /predict/expiry

Alerts

GET   /alerts
PATCH /alerts/:id/resolve

Mobile App

User Roles & Tab Structure

Patient (4 Tabs)

Home          ← drug search + availability + stock status
Scan QR       ← verify drug authenticity via blockchain
Prescriptions ← track active + past prescriptions
Notifications ← low stock alerts for prescribed medicines

Distributor (4 Tabs)

Inventory     ← stock levels, batch breakdown, incoming shipments
Transfers     ← initiate + accept + track batch transfers
Forecast      ← AI demand predictions 30/60/90 days
Chain History ← full blockchain transaction log

Manufacturer (4 Tabs)

Add Batch     ← register drug batch on blockchain + generate QR
Supply Chain  ← map view of all batches + regional distribution
Distributors  ← activity, stock levels, demand vs supply gaps
Alerts        ← anomaly feed + expiry warnings + resolve actions

Demo Flow

End-to-end demo in 5 steps:

Step 1 — Manufacturer registers a Artemether batch
         → batch written to Ganache chain
         → QR code generated

Step 2 — Manufacturer transfers batch to Madurai distributor
         → transfer pending on-chain

Step 3 — Distributor accepts transfer
         → ownership updated on-chain
         → inventory dashboard reflects new stock

Step 4 — Patient scans QR code
         → app queries blockchain
         → authenticity confirmed: manufacturer, expiry, chain of custody shown

Step 5 — Anomaly alert fires
         → ML service flags unusual stock drop on Insulin batch
         → manufacturer sees alert with confidence score 0.91
         → alert escalated for investigation

Security

✓ JWT required on all non-public endpoints
✓ RBAC middleware checks role before every write operation
✓ Private keys never exposed to mobile app
✓ All blockchain writes go through Node backend server wallet
✓ Patient role is read-only — zero write access to chain
✓ Anomaly alerts auto-escalate at confidence score > 0.85
✓ Ganache --deterministic flag keeps wallet addresses stable across restarts

Built for Tamil Nadu. Designed for every healthcare supply chain.

Name		Name	Last commit message	Last commit date
backend		backend
contracts		contracts
docs		docs
hf-26-025		hf-26-025
ml-services		ml-services
.gitignore		.gitignore
MEDCHAIN.jpg		MEDCHAIN.jpg
README.md		README.md
docker-compose.yml		docker-compose.yml

Latest commit

History

Folders and files

Repository files navigation