Modeling the treatment scheme of sustained ventricular tachycardia with a Bayesian belief network

Patients with sustained ventricular tachycardia (SVT) need intensive electrophysiological testing for a treatment decision. A computer-based support for this decision has to take into account the clinical experience with SVT, the facets of the required electrophysiological tests, and the general characteristics of the potential therapy. A Bayesian belief network (BBN) is developed that allows the test-therapy relationships to be modeled. In the arrhythmological context, the events (the nodes in the network) are cardiological findings, pathophysiological states, test results, and therapeutic decisions. The basics of BBN as a tool for the design of medical decision support are presented and the essential knowledge issues of the arrhythmological problem are treated