Authors: Mohammad Zaid Khan, Dr. Arvind Jaiswal
Abstract: The rapid growth of digital health platforms has created an urgent need for intelligent clinical decision-support tools that can interpret patient-reported symptoms and streamline care navigation. This work presents an enhanced architecture for MediTrack, a healthcare management platform, through the integration of a BioBERT-powered symptom triage and specialist recommendation engine. Leveraging domain-specific language representations, the system processes free-text symptom descriptions, identifies likely clinical categories, and recommends appropriate medical specialties with improved accuracy and contextual relevance. The proposed architecture combines natural-language preprocessing pipelines, BioBERT inference modules, probabilistic triage scoring, and a rule-augmented recommendation layer. Furthermore, the integration design emphasizes scalability, interoperability with existing MediTrack services, and compliance with healthcare data-protection standards. Experimental evaluation using benchmark clinical-symptom datasets demonstrates significant gains in classification performance and user-experience efficiency. This enhancement positions MediTrack as a more responsive, intelligent, and patient-centric digital health orchestration platform.
