Fuzziness in healthcare-associated infection monitoring and surveillance

Automated identification, monitoring, and reporting of health care-associated infections in intensive care units by connecting intensive-care medical information, laboratory information, and Arden-Syntax-based clinical decision support systems was proven feasible and operable. Raw clinical and laboratory data of patients are transferred to the system´s data warehouse. Ontologies listing all of the applied terms and a structured knowledge base with procedural and rule-based formalizations of the involved concepts and their relationships between them are part of the system. The inherent linguistic uncertainty of clinical terms is modeled by fuzzy sets. Degrees of compatibility (fuzzy degrees of membership) between data and the respective clinical terms under consideration are propagated by fuzzy logic. Uncertainty when it is a part of clinical propositions is modeled by truth values, and propagated by fuzzy logic as well. This large-scale system is known as Moni-ICU and runs as a routine clinical application at Vienna General Hospital, Austria.