MedAISupport: Bridging the Clinical Interpretation Gap Through Hybrid LLM and Rule-Based Medical Decision Support

MedAISupport is an AI-enabled Medical Decision Support System (MDSS) designed to bridge the clinical interpretation gap for both patients and clinicians. The system combines large language model (LLM)-based zero-shot reasoning with a deterministic rule-based engine to create a hybrid reasoning pipeline for medical report understanding. Built using FastAPI, React 18, and SQLite, the platform supports clinical report ingestion, biomarker extraction, risk stratification, temporal progress tracking, specialist recommendation, and referral workflow generation. In evaluation on 50 heterogeneous clinical report samples, the system achieved 92% biomarker extraction accuracy and completed report parsing in under 10 seconds on a commodity workstation. The architecture improves interpretability by generating dual-track outputs: a patient-friendly explanation layer and a clinician-oriented structured analysis layer with evidential justification. MedAISupport demonstrates how hybrid AI can support explainable, accessible, and workflow-compatible medical decision support in real-world healthcare contexts.