Blockchain-based Electronic Health Records (EHR) Management with OCR Assisted Summarization

A decentralized application built with ReactJS, Flask, Solidity, and IPFS, that allows users to store and retrieve their medical records securely. The application has built-in OCR capabilities that enable users and their trusted doctors to analyze the uploaded reports by extracting critical medical terms, including disease names and prescribed drugs.

Code     Project Report

The advantages of having a secure, immutable and decentralized Electronic Health Record (EHR) database:

  • Single version of the truth verified by the consensus of the participating hospitals
  • Easy to share selective or all EHRs as consented by the patient
  • Full medical history of a patient at one single point
  • Easy verification of medical prescription
  • Redacted EHRs for research purposes
  • Increased transparency
  • No insurance fraud

This application app has 2 main users:

  • Patients, who can:
    1. Upload a document to the blockchain. The document is added as a node in IPFS which returns a hash. The hash is then stored on the blockchain.
    2. View the uploaded documents.
    3. Analyse the uploaded documents. The text from the document is extracted and NER (Named Entity Recognition) is performed on the text using BERN (Biomedical Named Entity recognition and multi-type Normalization).
    4. Analyse their reports to find keywords related to Drugs or Diseases.
    5. Add a trusted doctor to view their medical documents.
  • Doctors, who can:
    1. Upload a medical document about a certain patient to the blockchain.
    2. View a certain patient's uploaded document.