Abstract T P206: An Executable Graphical Model for In Silico Prototyping of Processes for Acute Stroke Care

Abstract

Background: Strong evidence shows neurologic outcomes in acute ischemic stroke (AIS) worsen with delay from symptom onset to thrombolytic therapy. Yet this onset-to-treatment (OTT) time has not decreased in most systems of care over the past decade. Even the in-hospital, “door-to-needle” (DTN) component of this delay is unimproved, notwithstanding exceptions in institutions where innovative quality improvement efforts have borne fruit. Objective: Provide a basis for visualizing, communicating, and simulating stroke care system configuration and performance to facilitate the quality improvement efforts necessary for reducing DTN and OTT times in AIS. Methods: We developed an executable, graphical model of acute stroke care, employing the hierarchical colored Petri net (CPN) formalism. The top level of the hierarchy sets the epidemiologic context, including demographics and background processes like stroke prevention and onset. At deeper levels, we elaborate time-critical processes that contribute to OTT: stroke recognition, EMS activation and transport, and many emergency department (ED) processes. Key ED innovations described in the literature were modeled: EMS prenotification, a direct-to-imaging transport strategy, process parallelism, and telestroke capability. Results: Our CPN model has provided a platform for detailed, realistic prototyping and simulation of acute stroke care processes. The performance characteristics of process configurations with multiple, interacting innovations were evaluated and compared. Conclusions: In silico care process prototyping permits evaluation of proposed innovations in simulated settings. Using an intensively graphical simulation modeling methodology adds value by promoting “visual consensus” regarding care process structure and function, among stakeholders in a quality improvement initiative.