A SERVICE-TIME PREDICTION MODEL IN SIMULATION OF QUEUING ANALYSIS FOR DECISION SUPPORT IN HEALTHCARE

Abstract

The Radiology Department of a hospital in Najran city in Saudi Arabia is seeking ways to improve patient experience and use current resources more efficiently as they face growing visits numbers of patients. This study's identified primary key performance indicators are patient’s waiting time and staff’s idle time. The impact on patient waiting time and radiographers' idle time were explored in this study by using data mining techniques to predict the service time. The same simulation technique is used to study the impact of assigning a type of patients to a fast track, or separate unit for low-acuity patients in the Radiology Department using an operational research queue-based Monte Carlo simulation in a spreadsheet-based decision support tool. The model combined the principles of queuing theory. Also, it expanded the discrete event simulation in order to account for patients' arrival time rate and service time. In addition, the Department queue system was designed and analyzed by using the simulation model. The prediction model has been deployed into the decision support tool. Developing this tool aims to analyze the effect of changing particular aspects of the system on the total waiting time. The simulation indicates that the main problem is not the shortage of resources, but it is ineffective queue system management. Simulation results exhibited that the ability to accurately predict the service time and assign patients to a particular type of scanning room like a fast track minimized overall average waiting times 48.6 minutes to 40.4 minutes in the department during operation hours. This modeling approach with a decision support tool could be efficiently distributed and inform healthcare decision-makers of implementing a fast track or comparable system on patients' waiting times.