Beyond Discharge

An Android application was developed that integrates Google ML Kit OCR and FHIR standards to retrieve, process, and analyze discharge summaries for predicting ICU readmissions. The app uses a Flask-based backend with a trained Linear SVC model, achieving 85% prediction accuracy. A dual-input pipeline was implemented to enable real-time data retrieval from FHIR servers and OCR-based image scanning, significantly improving accessibility and accuracy in medical record analysis for predictive healthcare solutions.