Integrated framework of process mining and simulation-optimization for pod structured clinical layout design

This paper proposes a three-phase framework to leverage hospital tracking data of patient visits while designing healthcare layouts with pod structures. The first phase proposes a process mining algorithm that modifies the Probabilistic Determining Finite Automata (PDFA) with Particle Swarm Optimization (PDFA-PSO) algorithm to predict the significant patient workflows from hospital historical data. The second phase employs simulation modeling to solve a right-sizing problem to determine the optimal size of the layout pods and the frequency of flows between the different clinical locations. The final phase uses an Unequal Area Facility Layout Problem (UAFLP) to determine the layout typology. The proposed process mining and simulation model are vital steps to measure the frequency between spaces and pod areas, which are needed to solve the UAFLP for outpatient settings. The proposed framework is validated using a case study for a renovation project of a large heart and vascular clinic in the US. The research shows that process mining is an efficient tool to extract a subset of significant patient pathways among 90 pathway variants and build a more realistic simulation that reflects behavioral and operational aspects. The research shows that the PSO algorithm is efficient in estimating the PDFA parameters and improving the prediction accuracy of the extracted patient pathways. In addition, the research shows that Genetic Algorithm with Placement Staretegy is an efficient algorithm for layout automation.

Automated summary

This paper introduces a three-phase framework for designing healthcare layouts using hospital tracking data, with the use of process mining and simulation modeling to extract important patient pathways and improve prediction accuracy.