Fuzzy sets and decision making system for thyroid hormone profile

Automation of the analytical specimen processing tasks in clinical laboratories is provided by the laboratory information and automation systems. Integration of these systems increase throughput, lower the cost, improve and ensure quality standards, and store and keep data for knowledge maintenance and mining. Among several important tasks, the key one concerns with data acquisition from analyzers and their processing and interpretation as well. This paper introduces a real-time knowledge based system whose purpose is to select the analytical test results that must be audited by an expert from those results which are considered as coherent. The system, developed at Laboratorio Fleury, comprises a rule based fuzzy classifier and a learning module. The emphasis here is on the fuzzy classification and selection system as a means to increase decision making capabilities in advanced laboratory information and automation systems. Actual experiments with thyroid hormone profile tests reveals that the approach succeeded in screening a significant amount of coherent results.