Improving Hospital-wide Patient Scheduling Decisions by Clinical Pathway Mining

Daniel Gartner,I. Arnolds,S. Nickel
DOI: https://doi.org/10.3233/978-1-61499-564-7-1066
Abstract:Recent research has highlighted the need for solving hospital-wide patient scheduling problems. Inpatient scheduling, patient activities have to be scheduled on scarce hospital resources such that temporal relations between activities (e.g. for recovery times) are ensured. Common objectives are, among others, the minimization of the length of stay (LOS). In this paper, we consider a hospital-wide patient scheduling problem with LOS minimization based on uncertain clinical pathways. We approach the problem in three stages: First, we learn most likely clinical pathways using a sequential pattern mining approach. Second, we provide a mathematical model for patient scheduling and finally, we combine the two approaches. In an experimental study carried out using real-world data, we show that our approach outperforms baseline approaches on two metrics.
What problem does this paper attempt to address?