MediLink: A Secure Role-Based Healthcare Management Platform with Integrated Rule-Based Clinical Decision Support and Multilingual Voice Interaction
Digital healthcare adoption has expanded rapidly in recent years, yet many platforms continue to address isolated aspects of care delivery rather than offering an integrated patient experience. Existing telemedicine, electronic health record, and emergency-response systems typically operate as independent services, contributing to fragmented care, communication barriers, and delayed emergency response in linguistically diverse and connectivity-constrained regions. This paper presents MediLink, a hybrid online-offline healthcare support platform that consolidates rule-based symptom triage, multilingual interaction, electronic medical record management, doctor-patient communication, appointment scheduling, and GPS-based emergency ambulance coordination within a single Node.js and MongoDB-based system. The platform follows a layered architecture comprising presentation, application, business-logic, and database layers, with a cross-cutting authentication and role-based access control module. Its symptom-assessment component uses a rule-based, pattern-matching decision-support engine rather than a trained predictive model, deliberately designed as a modular layer so that a machine-learning or naturallanguage-processing-based model can be integrated in future iterations without architectural redesign. The system has undergone functional validation covering module-level workflows, database CRUD operations, authentication, appointment scheduling, and notification dispatch; formal quantitative evaluation involving benchmark datasets, load testing under varying traffic, and large-scale usability studies has not yet been conducted and is identified as future work. This paper discusses the system's design rationale, architecture, module composition, and validation approach, and positions MediLink against prior single-purpose telemedicine, electronic health record, and emergency-dispatch systems identified in the literature. The discussion highlights that architectural integration, rather than a single novel predictive algorithm, constitutes the primary contribution of this work, while acknowledging the need for further clinical and performance validation before real-world deployment.