Expert-based probabilistic modeling of workflows in context of surgical interventions

Abstract

To provide assistance functions in context of surgical interventions, the use of medical workflows plays an important role. Workflow models can be used to assess the progress of an on-going surgery, enabling tailored (i.e., context sensitive) support for the medical practitioner. Subsequently, this provides opportunities to prevent malpractices, to enhance the patient's outcome and to preserve a high level of satisfaction. In this work, we propose a framework which enables a formalization of medical workflows. It is driven by a dialog of medical as well as technical experts and is based on the Unified Modeling Language (UML). An easy comprehensible UML activity serves as a starting point for the automatic generation of more complex models that can be used for the actual estimation of the progress of a surgical intervention. In this work, we present translation rules, which allow to transfer a given UML activity into a Dynamic Bayesian Network (DBN). The methods are presented for the application example of a cholecystectomy (surgical removal of the gallbladder)

Interactive Decision Support: A Framework to Improve Diagnostic Processes of Cancerous Diseases Using Bayesian Networks

“…In this work, we introduce translation rules, automatically transferring a given UML activity into a BN. Modified from [8].…”

“…Thereby, the vertex set V represents the set of random variables. While a directed edge V k → V m within the set of edges E represents a direct dependency between two variables, the independence of two variables is symbolized by a missing edge [8].…”