Team-based multi-agent system for early detection of adverse drug reactions in postmarketing surveillance

Discovering unknown adverse drug reactions (ADRs) in postmarketing surveillance as early as possible is highly desirable. In the U.S., the Food and Drug Administration (FDA) has provided Web-based forms for spontaneous reporting of possible ADRs. Nevertheless, the process of analyzing and interpreting the reports, collecting additional relevant information, and drawing reliable conclusions requires collaboration between experts with different and complimentary skills (e.g., epidemiologists, biostatisticians, pharmacists and physicians). Multi-agent systems have been shown to be a promising approach for tackling distributed problem solving, especially when data sources and knowledge are distributed, and coordination and collaboration are required. Hence, we propose a team-based multi-agent framework for early detection of ADRs. In this framework, intelligent agents assist a team of experts based on a human decision making model called Recognition-Primed Decision (RPD). Fuzzy logic is used to determine the degree of similarity for retrieving experience in the RPD model. We describe our preliminary system design and illustrate its potential benefits for assisting FDA expert teams in early detection of previously unknown ADRs.