An AI-Based Web Application for Prescription Digitization and Medication Reminders in Elderly Care

Elderly patients managing multiple prescriptions face a higher risk of missing doses and
experiencing harmful drug-drug interactions (DDIs). This project creates a web-based
prescription and medication reminder system powered by AI, specifically designed for older
users. The system digitizes and stores prescriptions using Optical Character Recognition (OCR)
and medical Named Entity Recognition (NER). It maps drug data to standardized vocabularies
like RxNorm.
A hybrid DDI detection engine combines rule-based checks with machine learning to identify
potential interactions and generates clear alerts for caregivers and clinicians. The reminder
module provides various notifications, including visual alerts, push notifications, voice
prompts, and SMS/phone calls, ensuring accessibility for users with different levels of comfort
with technology and any disabilities. Missed-dose detection triggers notifications for caregivers
and follows a customizable escalation policy.
References:
To prioritize safety, privacy, and compliance, symptom-based over-the-counter (OTC)
suggestions are clearly labeled as non-prescriptive and require clinician validation when
activated. The implementation uses modern OCR models, such as TrOCR, and includes
explainable machine learning for DDI detection. It also features a senior-friendly user interface
(UI) with large buttons and voice-first workflows.
Deliverables include a prototype web application, evaluation of OCR/NER accuracy on
anonymized prescriptions, assessment of DDI detection, and user-acceptability testing with
elderly participants and caregivers. By integrating AI-driven prescription analysis with reliable
reminder systems, the project improves adherence, reduces medication errors, and supports
independent living for elderly patients.