Knowledge-augmented Agentic AI for Mental Health Medication Information Seeking

2026-06-24

AI Summary

This study addresses the longstanding fragmentation of psychiatric medication safety information between authoritative yet abstract regulatory reports and authentic but unverified patient narratives, which lacks an integrative approach that preserves traceability and distinguishes levels of evidence. The authors propose the first multi-agent knowledge graph framework that synergistically integrates data from Reddit, WebMD, and the FDA Adverse Event Reporting System. Leveraging large language models, the system achieves high-precision entity recognition (F1 = 0.969 for drugs, F1 = 0.973 for conditions) and employs standardized ontologies—ATC-N, ICD-10, and MedDRA—for semantic alignment. This approach enables source-aware fusion of patient-generated content with regulatory data while maintaining distinct evidence tiers, revealing that community-reported signals can precede FDA alerts by hundreds of days. The resulting framework establishes an auditable, traceable knowledge infrastructure for psychiatric drug safety.

Abstract

Patients increasingly seek medication information online, yet safety knowledge for psychiatric drugs is split between regulatory adverse-event records, which are authoritative but abstract, and patient narratives, which are experience-near but unvalidated. Integrating them without conflating evidence and anecdote is especially consequential in psychiatry, where poorly contextualised information can amplify fear, nocebo responses, and non-adherence. Here we develop a provenance-aware, knowledge-graph-based multi-agent framework unifying 466,525 Reddit posts, 60,782 WebMD reviews, and twenty years of U.S. FDA Adverse Event Reporting System records for nine antidepressants. A large-language-model entity-recognition pipeline benchmarked against physician annotations reached highest F1 scores of 0.969 for medications and 0.973 for conditions. The two community platforms were far more concordant with each other (overlap up to a Jaccard similarity of 0.905) than with regulatory reports, indicating that patient-generated data form a partly independent safety signal. For sertraline, many adverse events appeared in community sources hundreds of days before the corresponding FDA date. A Neo4j knowledge graph grounded in ATC-N, ICD-10, and MedDRA vocabularies preserves provenance, keeping every claim traceable and regulatory facts distinct from patient experience. These results establish source-aware integration as a route to more auditable psychiatric medication information, with usefulness and patient benefit to be tested prospectively.

Problem

• knowledge integration
• mental health medication
• adverse event reporting
• patient narratives
• regulatory data

Innovation

• knowledge-augmented AI
• provenance-aware integration
• multi-agent framework
• mental health medication safety
• patient-generated data