TRUST-AWARE AUTONOMOUS PRIOR AUTHORIZATION ORCHESTRATION FRAMEWORK USING FHIR, AI, AND MULTI-AGENT HEALTHCARE SYSTEMS

Abstract

Prior authorization is essential to healthcare administration as a process to ensure that treatments are both clinically necessary and cost-effective prior to approval by the healthcare insurance company. In most cases, however, one or more of these data exchange and/or manual efforts coincide, increasing the burden of data management and the time needed to deliver patients' care. As HL7 Fast Healthcare Interoperability Resources (FHIR), an innovative healthcare framework that allows electronic healthcare data to be shared and exchanged much faster and in significantly richer detail, becomes increasingly embraced, recent regulatory developments like the CMS-0057-F Final Rule and the work of artificial intelligence (AI) provide exciting new avenues to modernize prior authorization for the benefit of patients. This survey would be keen to collect and discuss the latest published research related to interoperability in healthcare (IEE) systems, FHIR format, AI-powered clinical decision support systems, explainable AI, multi-agent systems in healthcare, and the potential and benefits of using these tools to enhance the PA work process. It also identifies key areas that should be tackled regarding issues of trust, transparency, auditability, regulatory control and human oversight, and gaps in the field. Following the evidence review, the paper presents the proposed Trust-Aware Autonomous Prior Authorization Framework (TAPAF), which integrates interoperability considerations, accountabilities, governance, and AI agents to enable efficient, explainable, and patient-centric prior authorization processes.