Analysis and Modelling of An Acute Stroke Treatment Process

Abstract

Stroke is a leading cause of severe adult disability and death. Ischemic stroke is treatable with thrombolysis and endovascular thrombectomy (EVT). However, fast treatment is critical for improving patient outcomes. The literature reports many quality initiatives to reduce the treatment times by implementing improvement strategies to the treatment processes. In addition to these quality initiatives, industrial engineering methodologies have been used to improve the acute stroke treatment processes. This thesis comprises of three publications related to improving acute stroke treatment process in a single comprehensive stroke centre (CSC). The thesis objectives are 1) to identify the current state of research and gaps in the application of simulation modelling of the acute stroke treatment processes using a systematic literature review; 2) to analyze and find the time distribution of each task in the treatment process of a single centre using a time and motion study; 3) to model acute stroke treatment pathways, considering EVT and sub-tasks, in the single centre, and assess the impact of improvement strategies to decrease treatment times from emergency department (ED) arrival. The literature was systematically reviewed, and the results revealed that there is a need to develop a simulation model for EVT, a newer treatment, and include a finer resolution of tasks within the treatment processes. To identify system-related delays, all relevant tasks and their respective time distributions, a time and motion study was conducted in a Nova Scotian CSC. The results from the time and motion study were used to develop a discrete event simulation (DES) model to create a simulation model of the acute stroke treatment process in the centre. The simulation model was tested using nine improvement strategies. Door-to-CT time (DTCT), door-to-needle time (DNT) and door-to-groin puncture time (DGPT) are primary outcome measures. The combination of all improvement strategies resulted the highest time reduction for all outcome measures: DTCT (14.2 vs 10.1 min, p<0.001), DNT (39.4 vs 23.0 min, p<0.001), and DGPT (67.9 vs 38.5 min, p<0.001). Sub-task identification and modelling can help improve the accuracy of a discrete event simulation modelling. Future studies can use the developed model for their own centres.