Modeling of clinical practice guidelines for interactive assistance in diagnostic processes

Clinical Practice Guidelines (CPGs) include recommendations for actions and therefore provide a frame of reference for the medical practitioner during diagnostic processes. To facilitate the implementation of this recommendations we propose an interactive assistance. For this reason the CPGs of Chronic Myeloid Leukemia (CML) and Myelodysplastic Syndromes (MDS) are modeled using UML activities and Bayesian networks. To lower the barriers of CPG implementation we establish an interface between experts and models. It is based on translation rules transferring UML activities to Bayesian networks. The resulting models are used to provide an innovative assistance function allowing intuitive deviations from given CPG recommendations.