Healthcare Simulation Transforms Patient Flow at Emory’s Teaching Clinic

Introduction

Emory Healthcare stands as the only academic medical center in the state of Georgia, with an impressive network of ten hospitals, over 580 locations, and 230 primary care facilities across the state. Ranked as the number one healthcare system in Georgia, Emory has built a reputation for excellence in patient care, medical education, and innovative healthcare solutions.

Among Emory’s extensive network of facilities is the Dunwoody Family Medicine clinic, a comprehensive teaching facility that opened in October 2024. This newly established clinic represents a significant upgrade from its previous location, expanding from 25 to 33 exam rooms to accommodate growing patient demand. The facility provides a wide range of services including primary care, family medicine, orthopedics, spine and cardiology, as well as imaging, laboratory services, ambulatory surgery, and physical therapy.

What makes the Dunwoody clinic particularly unique is its role as a teaching facility. As Victoria Jordan, Vice President of Process Optimization and Innovation for Emory Healthcare explains, “The Dunwoody Family Medicine clinic is a resident-driven clinic. In fact, over 70% of the providers within the clinic are residents themselves.” This teaching environment creates specific operational challenges that impact both patient experience and educational requirements.

With a capacity to see over 350,000 patients annually and a projection to serve more than 20,000 patients in 2025 alone, optimizing operations at this facility became a critical priority. To address this challenge, Emory Healthcare partnered with Georgia Tech Industrial and Systems Engineering senior design students and Simio to develop a simulation model that would help identify opportunities for improvement.

“We specifically wanted to demonstrate how we could use simulation at Emory Healthcare,” explains Dr. Jordan. “We don’t have a lot of people that have used it. And as we were working with our primary care group, they were anxious to see. So this was really more of a demonstration to start with.”

The Challenge

The Dunwoody Family Medicine clinic faced a complex operational challenge stemming from its dual mission of providing excellent patient care while serving as a teaching facility for medical residents. This created unique workflow requirements that significantly impacted patient wait times and overall clinic efficiency.

As a resident-driven clinic, the facility operates under specific educational protocols that affect patient flow. Residents at different stages of their training have varying levels of autonomy and supervision requirements:

• First-year residents in their initial six months must meet with a preceptor (supervising physician) in the middle of each patient appointment, with the preceptor returning to the exam room with the resident
• First-year residents in their second six months still meet with the preceptor mid-appointment but the preceptor no longer needs to return to the exam room
• Second-year residents can “stack” up to 2-3 patients before consulting with a preceptor
• Third-year residents can stack up to 3-4 patients before consulting with a preceptor

These supervision requirements created significant bottlenecks, particularly at the preceptor’s office. With only 2-3 preceptors available each day supervising 10 providers, 3-4 nurses, and 5-7 medical students, wait times accumulated throughout the day.

Data analysis revealed concerning patterns in patient wait times:

• 40% of patients waited longer than 10 minutes just to be roomed at the beginning of their appointment
• 50% of patients waited longer than 10 minutes after the nurse left for the physician to arrive
• Patients waited an average of 34 minutes during their appointments
• 61% of patients arrived less than 15 minutes before their scheduled appointment time (despite being asked to arrive 15 minutes early)
• 19% of patients arrived after their scheduled appointment time
• 77% of rooming was not completed by the appointment start time
• 93% of providers entered exam rooms after the scheduled appointment start time
• 90% of appointments ended later than projected

“We noticed system backlogs,” explained one of the Georgia Tech team members. “At the beginning of the day, there are very little wait times. As the session progresses, towards the middle of the day, the end of the morning session, and at the end of the day, appointments start ending later and later.”

The complexity of the clinic’s operations, with multiple interdependent processes and the unique preceptor-resident interaction model, made it difficult for staff to identify the root causes of delays and develop effective solutions. This environment presented an ideal opportunity for the application of simulation in healthcare to visualize, analyze, and optimize patient flow.

Solution Approach: The Power of Simulation in Healthcare

To address these complex challenges, Emory Healthcare partnered with Georgia Tech Industrial and Systems Engineering students to develop a comprehensive digital twin in healthcare using Simio simulation software. This approach allowed them to model the intricate operations of the Dunwoody Family Medicine clinic and test potential improvements without disrupting actual patient care.

Data Collection and Analysis

The project began with extensive data collection from multiple sources:

• Electronic Health Records: The team leveraged Emory’s robust database to extract information on patient flow, including check-in times, rooming start times, nurse service durations, and visit end times.
• Time Studies: Since the electronic records didn’t capture non-patient-facing activities (particularly preceptor interactions), the team conducted on-site time studies to collect data on how long providers spent with preceptors and in queue waiting for preceptor availability.
• Process Mapping: The team documented the detailed workflow for different types of providers (faculty, experienced residents, and first-year residents) to capture the unique aspects of this teaching clinic.

With this data in hand, the team performed statistical analysis to identify the most significant factors affecting each step in the patient journey. They discovered that:

• Provider experience level significantly impacted service time, with less experienced providers taking longer to treat patients
• Patient age group and appointment hour affected arrival patterns
• Patient age group, sex, and appointment type influenced nurse service time
• Appointment types showed distinct patterns that could be clustered for more accurate modeling

To avoid overfitting and simplify the model, the team conducted correlation analysis to cluster similar attributes. For example, they found that patient arrival patterns could be grouped into just three categories (8 AM, 1 PM, and all other hours) rather than modeling each hour separately.

Building the Digital Twin

Using Simio simulation software, the team created a detailed digital twin of the Dunwoody clinic that visually represented the physical layout, patient flow, and resource allocation. The simulation included:

• Visual representation of the clinic’s four pods, each containing 7-10 exam rooms
• Patient arrival and check-in processes
• Nurse rooming and initial assessment activities
• Provider-patient interactions
• Preceptor consultation processes
• Additional care procedures (lab tests, vaccinations, etc.)

“We tried to model the clinic visually as best as possible just to make this as useful as possible,” explained one of the Georgia Tech team members. The simulation allowed clinic staff to visualize patient movement, identify bottlenecks, and understand how delays propagated throughout the system.

A key innovation was the development of a data preprocessor tool that allowed the clinic to import actual patient schedules into the simulation. This enabled them to test specific days or scenarios by simply selecting a date, running the script, and importing the resulting CSV files into Simio.

Implementing Healthcare Simulation Standards

To ensure the model’s accuracy and reliability, the team followed healthcare simulation standards for validation:

• Face Validity: They verified that the simulation logic matched the actual clinic operations, particularly the complex preceptor-resident interaction patterns.
• Input Model Validation: They compared the simulation’s input distributions (like average nurse service time) against actual data from days not used in model development.
• System Interaction Validation: They evaluated how the various components worked together by comparing metrics like average time in system between the simulation and real-world observations.

This validation process revealed that while many aspects of the model accurately reflected reality, some refinements were needed. For example, the simulation strictly enforced a 1-to-1 assignment between patients and nurses, whereas in reality, nurses would sometimes help each other when backlogs developed.

Results and Business Impact

The healthcare simulation project delivered valuable insights that led to several practical recommendations for improving clinic operations. Through what-if analysis, the team identified four key opportunities for patient flow optimization:

1. Preceptor Assignment Changes

The simulation revealed that changing the preceptor assignment system from a 1-to-1 mapping (where each resident is assigned to a specific preceptor) to a first-come, first-served model could reduce wait times at the preceptor office by 31%. This simple operational change required no additional resources but could significantly improve patient flow.

2. Flexible Stacking Constraints

The team discovered that allowing residents to visit available preceptors opportunistically, rather than strictly adhering to their maximum stacking limits, could further reduce preceptor waiting time. For example, if a second-year resident had seen two patients (their maximum stack) but a preceptor was available, allowing them to consult immediately rather than waiting for their assigned preceptor would improve efficiency.

3. Strategic Room Assignments

One of the most straightforward yet impactful findings involved the physical layout of the clinic. The time study revealed that first-year residents (who need to consult preceptors most frequently) could reduce their travel time by 60% if they were assigned to Pod 4, which was closest to the preceptor room.

As Dr. Jordan noted, “It was interesting that one of the recommendations the team made was to move the first-year residents who have to check in with a preceptor after each visit to the pod that was closest to the preceptor office, which sounds completely obvious in hindsight, but it’s something that the leaders of the clinic were really happy to see because they’re like, ‘We see it every day and we just never thought about that.’”

4. Appointment Length Optimization

The analysis showed that the clinic’s standard 20-minute and 40-minute appointment slots didn’t always align with actual service times. By better matching appointment lengths to typical service durations for different visit types, the clinic could reduce both provider idle time and patient waiting time.

Business Value and Stakeholder Response

The impact of simulation training on patient care at Emory Healthcare extended beyond specific operational recommendations. The project delivered several broader benefits:

• Improved Patient Satisfaction: By identifying ways to reduce wait times, the simulation helped address a key factor in patient satisfaction.
• Enhanced Quality of Care: Reducing provider stress and time pressure allows for more focused patient interactions.
• Educational Value: The simulation provided a tool for demonstrating process improvement concepts to residents and staff.
• Stakeholder Buy-in: The visual nature of the simulation helped clinical staff understand and accept the recommended changes.

“The primary care team was really excited about the possibility of implementing some of the recommendations from the team,” Dr. Jordan explained. “They gave us very positive feedback. They said it really helped to have some fresh eyes look at the process and identify things that in hindsight seemed very obvious.”

Future Applications and Lessons Learned

The success of this initial healthcare simulation project has opened the door for expanded applications across Emory Healthcare. The organization is already planning next steps to build on this foundation:

Phase Two: Resident Scheduling Optimization

The next phase will focus on optimizing scheduling for residents and patients to ensure all residents complete the procedures required for their training. “Phase two will be to optimize scheduling for the residents and patients so that we can make sure all our residents complete all the procedures on their checklist,” explained Dr. Jordan.

Emory has already requested another Georgia Tech team to work on this optimization model for the fall semester, demonstrating their commitment to continuing this data-driven approach to healthcare process improvement.

Broader Implementation Across Emory Healthcare

Beyond the Dunwoody clinic, Emory sees potential for applying similar simulation models across their extensive network. “We have over 300 clinics across the Atlanta area,” noted Dr. Jordan. “We’re looking for how we can use similar modeling to optimize patient flow and resource usage in those as well.”

This expansion represents a significant opportunity to standardize best practices and improve operations throughout the Emory Healthcare system.

Key Lessons from the Project

Several valuable lessons emerged from this healthcare simulation initiative:

• Stakeholder Involvement is Critical: The project’s success depended on close collaboration between the Georgia Tech team, Simio consultants, and Emory Healthcare staff. As one team member noted, “Emory’s been a massive help. I don’t think we would be able to complete this project without their assistance.”
• Visual Simulation Enhances Understanding: The visual nature of the Simio model helped clinical staff grasp complex operational dynamics that weren’t apparent in their daily work.
• Data-Driven Decisions Build Confidence: The simulation’s ability to quantify the impact of proposed changes helped build confidence in the recommendations.
• Academic-Industry Partnerships Add Value: The collaboration between Emory Healthcare, Georgia Tech, and Simio demonstrated how academic institutions and industry partners can work together to solve real-world healthcare challenges.

Conclusion

The healthcare simulation project at Emory’s Dunwoody Family Medicine clinic demonstrates the powerful impact that digital modeling can have on healthcare operations, particularly in complex teaching environments. By creating a detailed digital twin of the clinic, the team was able to identify specific, actionable improvements that could significantly reduce patient wait times and enhance both the educational experience for residents and the care experience for patients.

As Dr. Jordan summarized, “Through this project, the Georgia Tech team, along with Simio’s facilitation support, did a great job of getting an initial digital twin in Simio. It gave us a solid preliminary model upon which we will build.”

The success of this project highlights the growing importance of simulation in healthcare as organizations seek to optimize resources, improve patient experiences, and maintain educational excellence in teaching facilities. By following healthcare simulation standards and leveraging advanced modeling capabilities, Emory Healthcare has established a foundation for continuous improvement that can be expanded across their entire system.

“We would like to specifically thank Simio for the work that they did to help us get this going,” Dr. Jordan concluded. “Greer and her team helped us make sure that we had a consistent system between the software that we were using internally and the software that the students were using. They also helped us with education and consulting support that was invaluable in our efforts.”

This case study illustrates how healthcare simulation can transform operations in ways that benefit all stakeholders—patients, providers, residents, and the healthcare system as a whole—while providing a roadmap for other organizations facing similar challenges.