Date of Defense


Date of Graduation



Industrial and Entrepreneurial Engineering and Engineering Management

First Advisor

Tycho Fredericks

Second Advisor

Steven Butt


This senior design project was completed at Ascension Borgess Hospital located in Kalamazoo, Michigan. The mission of Ascension Borgess Hospital is to provide patient centered care. In 2018, 562 patients left the Emergency Department (ED) without being seen, which is a large safety risk for the patients and hospital. Based on this, as well as other obstacles, this project focused on meeting the mission of Ascension Borgess by improving key ED metrics that impact patient centered care. In order to meet that mission, this study aims to improve patient experience, safety and quality of the care provided, profitability for the hospital, and staff satisfaction. This will be accomplished by reducing the number of patients that leave without being seen, reducing the total patient length of stay, and reducing the ineffective use of staff.

Discrete-event simulation software, ProModel, was used to test changes in the ED without disrupting the current system. The current state model was created in ProModel through observations, data collection, and provided hospital databases. The locations, entities (patients), arrival cycles of the entities, processing code, resources (staff), and path networks were created in the current state model. Patients that enter the ED are assigned an acuity level during the triage process which is based upon the severity of the patient's symptoms as well as the number of resources they could potentially occupy from the hospital. The acuity scale ranges from 1 to 5, 1 being high severity and 5 being nonurgent symptoms. The current state model was verified using visual testing and stress testing, and then it was validated using paired sample t-tests on key outcomes.

Four outputs from each simulation model were analyzed in order to measure the three metrics from the project objectives which were the number of patients that leave without being seen (LWBS), the average patient length of stay (LOS), and the inefficient use of staff. The four metrics that were analyzed for each model were: the number of patients that leave without being seen, the average patient length of stay, the percent of time a patient was blocked in the system, and the resource underutilization.

After the current state model was created, four different models were tested against the current state model. The first model created was vertical care, which focuses on patients with acuity level four and five meaning that their symptoms are nonurgent and can be treated in a vertical chair. After they receive treatment, they wait in a results pending area. This helps reduce cleaning cost and utilizes less resources from the hospital. The results from this model showed no changes from the current state metrics.

The second model created focused on check-in and triage changes. This model utilized two triage locations to reduce registered nurses (RN) from doing non-registered nurse related activities. Model two was able to reduce the number of patients that left without being seen by 13% and the percent of time a patient was blocked in the system by 16%.

The third model tested was a proposed lobby layout. This layout was created by essential ED staff that implemented new components into the lobby. Those components were four triage bays, minimal waiting area, two kiosks for digital registration, a separate location for heart attack testing, and a results pending area. The results of this study showed a reduction in LWBS by 22% and a percent of time a patient was blocked by 20%. Based on the simulation outputs and systematic layout planning (SLP), a revised lobby layout was created. This layout reduced the number of triage bays to three and the number of EKG curtain locations to one. The results from this output were promising with a reduction in the number LWBS and percent blocked by 37% and 22%, respectively.

Finally, the fourth model which added an admit holding unit was created and tested. This model added a waiting area for patients being admitted into the hospital for further treatment which created more ED bed availability. The outputs from this model were the most significant as the reduction in LWBS was 69% and the percent blocked was reduced by 40%.

The models were compared through two methods: statistical analysis and a cost analysis. The statistical analysis used ANOVA testing, specifically a Tukey-Kramer test, to determine if the outputs produced were statically different from the current state model. Model 3 showed the most statistically different results for almost every acuity. Additionally, this analysis showed statistically different metrics in model 4 for patients with acuity 1, 2, and 3. This is expected as only patients with more severe symptoms are typically admitted into the hospital. Next a cost analysis was performed which looked at resource usage cost as well as return on investment (ROI) and payback period. Model 4 was the best according to the cost and investment analysis. Model 4 was significantly lower than the other three models for yearly patient resource usage cost by about one million dollars. The return on investment for model 4 was the highest at 63% and had the lowest payback period at 1.7 years.

Based on the results of the study, two short-term recommendations were made. The first was to implement model four, the admit holding unit, as it showed the greatest improvements to the system and cost metrics. Additionally, model two was recommended to be implemented due to its small initial investment of under $74,000. The results of this model were positive for such a small change and investment. Finally, the long-term recommendation was to implement model 3, the proposed lobby layout. This is due to the systems metric benefits and ROI of 30%. The results of model four were very significant to the system and drastically different than the analysis on the other models. This showed that the underlying issue was due to admission blockages throughout the hospital. Therefore, it is recommended to complete a root-case analysis on the admissions issues throughout the entire hospital as implementing model 4 is a surface solution. To complete this, the entire hospital would need to be modeled as patients leaving the ED can move to any department throughout Ascension Borgess.

Finally, the proposed lobby layout and the admit holding unit were combined and a simulation was created. This simulation representing 0ppo These overall recommendations accomplished two of the three metrics: the number of patients that leave without being seen and the patient length of stay. Overall, this study was able to improve patient experience, safety and quality of the care provided, and profitability of the hospital.

Access Setting

Honors Thesis-Open Access