Simulation and optimization of hospital registration window based on FlexSim
This was a course project during my undergraduate studies and was published in a Chinese journal. The course paper can be downloaded here. The following is a translation of the paper.
Abstract:
The queuing problem in hospital is becoming more and more serious. This paper investigates the queuing system of registration window of a hospital in Nanjing. The data of arrival time interval and service time of each window are collected, and its distributions are fitted by Matalab. The model of the system is set up and simulated with FlexSim. According to the simulation results, the optimization scheme is put forward. By increasing the number of windows opened to patients from 3 to 4, the average waiting time is reduced by 430 seconds, and the queuing time is reduced by 69%.
Key words: hospital；registration window；queueing system；simulation；FlexSim
Introduction
With the improvement of people's health needs, the number of medical patients continues to increase. The queuing problem in hospitals has also become increasingly prominent, causing a lot of inconvenience to patients seeking medical treatment. In particular, the service efficiency of the registration and payment system is of great significance in reducing patients' waiting time. Modeling and simulation software can be used to study the queuing phenomenon and propose optimization plans based on the simulation results, thereby improving the service efficiency of the medical system and improving the patient medical experience.
The author investigated the queuing situation of the registration window in a hospital in Nanjing, used Matalab to process and fit the collected queuing and service time data, and combined with FlexSim simulation technology, simulated the situations when different numbers of registration windows were opened. Improvement plans are proposed through comparison of simulation results.
Investigation
After long-term field investigation, it was found that the hospital has a total of 6 registration windows, of which 3 windows (No. 3, 4 and 5) are open for a long time, and 3 windows (No. 1, 2 and 6) are rarely open.
The registration window of the hospital is responsible for not only registering patients when they enter the hospital, but also accepting payment services before patients purchase medicine or see a doctor. During the investigation, it was found that queuing at the hospital's registration window was common, especially during peak hours.
Data Fitting
Probability distribution of patient arrival time intervals
The collected raw data were calculated and extracted to obtain multiple sets of data on patient arrival time intervals. Using Matalab to perform distribution fitting on these data, it can be found that the time intervals for customers to arrive at the registration window all obey exponential distribution.
Probability distribution of window service times
Extract the registration/payment time of each patient from the original data, and use Matalab to perform distribution fitting on these data. It can be found that the service time of windows 2, 4, and 5 obeys the normal distribution, and that of windows 1, 3, and The service time of No. 6 follows an exponential distribution.
FlexSim Modeling
Model entity composition
①Number of Generator: 1 (simulates patient arrival).
②Number of Absorber: 1 (simulates patient leaving).
③Number of Processor: 6 (simulated service window).
④Number of Temporary storage area: 6 (simulated patient waiting queue).
⑤Number of Operator: 6 (simulated staff).
Simulation
As shown in Figure 4, the temporary entity (patient) arrives at the system through the generator, and by judging the number of people queuing at each open window, it chooses to join the queue with the smallest number of people queuing. If the selected processor is in service, the patient will first wait in the waiting area. Then it goes through the processor (registration window) to handle business, and finally enters the absorber and leaves the system.
Result Analysis
The situations when opening 2, 3, 4, 5 and 6 windows were simulated in the model respectively, and the corresponding indicators under opening different numbers of windows were obtained. As shown in Table 1:
| Number of windows opened | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|
| Average window utilization | 0.95 | 0.94 | 0.89 | 0.74 | 0.62 |
| Average waiting time/s | 682 | 622 | 192 | 82 | 68 |
When three windows are opened, the average window queuing time is about 682 seconds, and when four windows are opened, the average window queuing time is about 192 seconds. Analysis shows that from 3 windows to 4 windows, the average queuing time is greatly reduced, while the average window utilization is only reduced by about 5%. If five more windows are added, the average queuing time will only be reduced by about 110 seconds based on the opening of four windows, but the average window utilization will be reduced by 15%. If the number of windows is further increased to 6, the average utilization of the windows further decreases but the average waiting time does not improve significantly.
Suggestions for Improvement
Based on the simulation results of the FlexSim model, the following improvement suggestions are put forward:
(1) Increase the number of registration and payment windows currently opened every day from 3 to 4.
① Before improvement: When 3 windows are opened, the average utilization rate is 94% and the average waiting time is about 622 seconds.
②After improvement: When 4 windows are opened, the average utilization rate is 89%, and the average waiting time is about 192 seconds.
After increasing the number of registration and payment windows opened every day from 3 to 4, the average window utilization rate only decreased by 5%, but the average waiting time was reduced by about 430 seconds, and the queuing time was reduced by 69% compared with before. The improvement effect is obvious. With. If the number of windows opened is further increased, the marginal returns brought about by the improvement will be diminishing relative to the increased costs.
(2) During special periods when the number of patients coming for medical treatment is small or large, the number of existing windows should be flexibly adjusted by comprehensively considering costs and benefits.
Conclusion
This article uses FlexSim modeling and simulation technology to study the number of hospital registration windows, and proposes an improvement plan based on the simulation results. The application of industrial engineering in medical systems has broad prospects. By using industrial engineering methods, it can optimize department staffing, reduce waiting or walking waste, improve medical treatment processes, and many other aspects, thereby effectively improving the hospital's service efficiency and alleviating the strain on medical resources. Provide patients with high-quality medical services.