KNOWLEDGE GRAPH-ENHANCED RETRIEVALAUGMENTED GENERATION FOR HEALTHCARE REVENUE CYCLE OPTIMIZATION

Abstract

Healthcare revenue cycle management requires accurate interpretation of payer policies, clinical documentation, coding standards, provider contracts, and reimbursement regulations to optimize financial performance and reduce claim denials. Conventional Retrieval-Augmented Generation (RAG) systems rely primarily on semantic similarity retrieval and often overlook complex relationships among healthcare entities. Knowledge Graph-Enhanced RetrievalAugmented Generation combines semantic retrieval with structured knowledge graphs to improve contextual understanding, evidence retrieval, and decision support. This paper proposes a Knowledge Graph-Enhanced RAG framework integrating semantic search, healthcare knowledge graphs, vector databases, enterprise knowledge management, cloud-native MLOps, and automated governance for intelligent healthcare revenue cycle optimization. The proposed architecture retrieves semantically relevant documents together with graph-based relationships among diagnoses, procedures, policies, providers, and reimbursement rules to generate evidence-grounded recommendations. Experimental evaluation demonstrates improvements in retrieval precision, claims validation accuracy, decision consistency, operational efficiency, regulatory compliance, and revenue cycle performance. The proposed framework provides a scalable, explainable, and production-ready solution for intelligent healthcare revenue cycle optimization. Keywords— Knowledge Graph, RetrievalAugmented Generation, Healthcare Revenue Cycle Management, Semantic Retrieval, Large Language Models, Healthcare Claims, MLOps, Explainable AI