A HYBRID KNOWLEDGE-BASED SYSTEM FOR THERAPY ADJUSTMENT IN GESTATIONALDIABETES

This poster describes a system to analyze self-monitoring data of gest ational diabetic patients, for obtaining an assessment of their metabo lic control with the final goal of supporting decision-making in thera py adjustment. The system is able to manage incomplete data and to mak e temporal reasoning under uncertainty, the two most important constra ints when analyzing ambulatory monitoring data. Two different formalis m have been used to represent and manage the knowledge: a dynamic Baye sian network and a production system based on rules. The outcomes prov ided by the whole system are: information on possible patient transgre ssions of the prescribed treatment and recommendations of treatment ad justments.