Hospital-Patient Connected Smart Medication Management for Older Adults: Item-Level Confirmation, Selective Re-Reminders, and Traceable Alerts

This proposal presents the design of a hospital-patient connected smart medication management system for older adults. The system is designed around item-level medication tracking rather than conventional time-slot-only reminders. Each scheduled medication item maintains an independent state, allowing users to confirm items individually or through a low-interaction batch action while preserving separate event records for each medication item. Items that remain unconfirmed can be selectively re-reminded after a configurable grace period. The proposed architecture consists of an Android-oriented patient application, a mock clinician dashboard, and a shared relational backend. The patient application uses on-device local notifications to maintain reminder functionality after schedule synchronization even when continuous network connectivity is unavailable. Prescription changes, medication events, simulated symptom reports, and clinician-facing alert states are synchronized through the shared backend. The prototype is planned using Expo/React Native and TypeScript for the patient application, React for the clinician dashboard, and Supabase/PostgreSQL for authentication, relational data management, row-level access control, and realtime synchronization. The proposal also defines an OS-aware notification strategy, offline event synchronization, older-adult-oriented interaction principles, traceable exception handling, and quantitative engineering evaluation criteria. This is a pre-implementation system proposal. It does not use real patient records or real hospital EMR data, does not objectively verify medication ingestion, and does not provide autonomous diagnosis, prescribing, dosage modification, or medication-specific clinical decision-making. Demonstration and evaluation are planned using synthetic data.