Simulation Optimization-Based model for Decision-Making in the stroke clinical pathway

Healthcare is highly complex, sensitive and needs constant improvements. Several works have already been developed to support those processes. However, finding the optimum solution takes much work and time. Multi-Objective Genetic Algorithms (MOGA) improve the results by finding the optimal trade-off between multiple conflicting objectives and exploring the problem space more thoroughly. This study presents an enhanced framework that integrates Process Mining (PM), Discrete Event Simulation (DES), and Multi-Objective Genetic Algorithms (MOGAs) into an optimized, end-to-end pipeline. This framework builds upon an existing non-optimized approach to enable decision-makers to explore and implement Key Performance Indicator (KPI)-oriented solutions directly from raw log data. The framework navigates vast and complex solution spaces by embedding MOGAs into the KPI-oriented simulation process, delivering optimized scenarios with improved performance, boosting decision-making efficiency. The clinical stroke pathway, covering symptoms’ onset to hospital discharge, was utilized as a case. This research demonstrates how optimization techniques with classical techniques into one unified framework can accelerate healthcare improvements, offering scalable applications to other domains beyond stroke care. The results demonstrate that using MOGA leads to improved solutions compared to the non-optimized framework, and this approach can be evaluated in short periods due to its performance. The findings underscore the solutions’ sensitivity to changes in simulation parameters, emphasizing the importance of considering multiple objectives when dealing with complex decision-making problems in the healthcare industry. Future studies are suggested to extend the model, compare the effectiveness of different optimization methods within the framework, and test the framework’s applicability to other domains.