The Pavilion at the Hospital of the University of Pennsylvania is rising as a towering example of the value of behavioral research in health-care building design.

By Christina Hernandez Sherwood

Photos by Addison Geary

On a February day in 2016, architects, designers, contractors and Penn Medicine employees gathered in a Market Street conference room to assess their progress on the Pavilion, the inpatient hub planned for the Hospital of the University of Pennsylvania (HUP).

The team had spent months on a meticulously thought-through design of the massive, 1.5-million-square-foot facility. They had worked collaboratively, bringing together Penn Medicine’s own clinical experts with architecture, design, and construction professionals who specialize in health care, and they had relied on the latest research about how to organize clinical care spaces to be effective for both clinicians’ and patients’ needs—for example, incorporating so-called “onstage” and “offstage” zones, which put staff in front of or out of view of patients, respectively. They had even spent months of work and hundreds of thousands of dollars on the construction of a 30,000-square-foot mockup of half an inpatient floor and offered tours for hundreds of employees—and in their first set of surveys, respondents frequently endorsed many of the building’s design details as “ideal.” Now the design team was ready to hear the results from another, more involved tour of that foam-and-cardboard space to decide if the initial design of the Pavilion’s inpatient floors was truly viable.

The verdict: It wasn’t. By and large, Penn Medicine clinical and support staff who spent hours in the mockup running through simulations of their day-to-day workflows gave certain key aspects of the plan a thumbs down. Among the criticisms: Certain hallways weren’t wide enough. Just 42 percent of simulation participants felt patient bathrooms were designed to accommodate staff assistance. And only about half of visitors to the mockup thought the dedicated patient elevators were optimally located.

After the presentation, the design team had some positive attributes from their original plan they knew to keep—but to address the criticisms, they began again. “We always knew we were going to be making some changes, but not to the extent we ended up doing,” said Stephen Greulich, senior project manager. “We didn’t think we’d be redesigning the building.”

Yet while this rejection might sound like a failure, the Pavilion design process was working exactly the way it was intended. The multilayered process of creating this new hospital building uses behavioral research to guide design choices at each step. That meant asking the people who would work in the new hospital how they thought it should be designed, listening to their feedback, and testing their assumptions. As a result, when it opens in 2021 the 17-story Pavilion will not only be Penn’s biggest and, at $1.5 billion, most expensive building project to date. It will also be a robust architectural expression of the spirit of innovation, one that could stand as a towering example of the evolution of health-care facility design from art toward science. The Penn team has taken the process of building a cutting-edge hospital as an opportunity to discover along the way how one should be designed.

A New Approach and a Nagging Feeling

The last time an inpatient facility was erected from scratch as part of HUP’s network of interconnected buildings at 34th and Spruce Streets in Philadelphia was in the 1990s. At that time, behavioral research was seldom used to inform design. Twenty-five years later, techniques such as mockups and real-life simulations are much more common—though the scale of their use in health care building design is still limited. But when the team charged with building the Pavilion was invited to take risks and think outside the confines of the traditional hospital, a whole new world opened up. (Unlike HUP’s existing inpatient buildings, the Pavilion will be a freestanding structure across the street, adjacent to HUP’s outpatient facility, the Perelman Center for Advanced Medicine, on Civic Center Boulevard; bridges and tunnels will connect the entire complex.)



Kevin Mahoney, senior vice president and chief administrative officer for the University of Pennsylvania Health System at the time of this photograph in 2018 (CEO of UPHS since July 2019), has been involved in planning for new Penn Medicine buildings for two decades. He had key roles in deciding to create the integrated PennFIRST team and in suggesting the full-scale mockup of half of an inpatient Pavilion floor.

The first innovation was the mid-2015 creation of the Pavilion’s specially assembled project team, PennFIRST. As an “integrated project delivery” team, PennFIRST brings together staff from the various groups involved in the project into a new, unique entity—Penn Medicine employees working together with outside architects and designers from the global health care design firm HDR and international architect Foster+Partners, as well as engineering design and construction management experts from BR+A, L.F. Driscoll, and Balfour Beatty. Now three years into the Pavilion design and construction process, the PennFIRST team of about 100 still occupies a full-floor co-location space in University City. There, moveable walls and tape on the floor are shifted as designs evolve, a screen in the kitchen live streams a view of the Pavilion construction site, and Penn nurses, who are part of the project team, share space with designers and architects. “I wanted everybody to sit around the table like a large family and talk issues through,” said Kevin Mahoney, executive vice president and chief administrative officer for the University of Pennsylvania Health System. “Everybody had to be together.”

Mahoney, who for two decades has been involved with planning new buildings on nearly every piece of Penn Medicine real estate across the region, said he hoped using an integrated project delivery contract would avoid a common pitfall of these projects: bickering among the architect, contractor and owner. With this contract, everyone is at equal risk—or reaps equal reward. If the project comes in under budget, the savings are shared by the architect, contractor, and owner, Mahoney said. If it goes over budget, the penalty is shared, too.

Despite assembling a veritable dream team of health care building designers, Mahoney wasn’t happy with the first Pavilion plan. “I was uncomfortable that we were about to spend this much money,” Mahoney said, “and I had this nagging feeling in the pit of my stomach.”

At Mahoney’s suggestion, in the fall of 2015, the PennFIRST team rented a warehouse in Philadelphia’s Northern Liberties neighborhood and hired union contractors to construct a full-scale styrofoam mockup—the largest in Penn Medicine history—of half an inpatient floor based on that first Pavilion design. “This was about proving out how the building worked,” Greulich said. “We can’t do this halfway and find out in five years there’s a major flaw.”

More than 600 people from all levels of Penn Medicine were bussed to the warehouse for scheduled facilitated tours of the mockup and to give feedback on the design—and the feedback was overwhelmingly positive. “I kept sitting there saying, ‘That doesn’t make any sense to me because I still don’t think it’s right,’” Mahoney said. That’s when he called in a Penn Medicine specialty that had never before been consulted on building design: simulation.

A Simulation Innovation

Historically, simulation at Penn Medicine was used for educating staff and assessing competence. The simulation team joined Penn Medicine Academy, the change management, learning and performance improvement focused branch of the human resources department, about four years ago, at the same time a cluster of Ebola virus cases in the United States was prompting hospitals to prepare for the worst. The Penn Medicine Academy team—shifting from staff training to emergency management—tested Ebola response at three Penn Medicine facilities, evaluating the spaces, teams, and processes, and providing feedback.

Coming off the Ebola simulations just as preliminary work began on the Pavilion, Cindy Morgan, Penn Medicine’s vice president for learning and organizational development, had a breakthrough. Simulation had played a role in how staff learned to use new health system spaces before—most recently, in training for the transition of Penn’s trauma center from HUP to a new facility at Penn Presbyterian Medical Center. Why not use simulation, she thought, in the process of actually developing the facility design for the Pavilion? If future Pavilion employees could interact with the design as realistically as possible, perhaps they could tell the PennFIRST team how the building could work better. “The project has a feeling of being the first of its kind,” Morgan said. “As long as we’re working within the spirit of innovation and trying things that haven’t been done before, it allowed us to be courageous, to put our toe in the water and say, ‘We have an idea that we’d like to try.’” In a meeting with PennFIRST, Morgan volunteered the Penn Medicine Academy team for the Pavilion project. Mahoney, for one, was smitten.

Rather than using simulation to train for infrequent, high-stakes scenarios, like an operating-room fire or patient allergic reaction, simulation at the Pavilion was meant to let clinicians interact in the space as if it were an ordinary day. It took weeks for the simulation team to design the dozens of interweaving—and increasingly complex—workflows to simulate in the mockup over a four-hour period. At the same time, a cohort of about 100 Penn Medicine staff from a variety of clinical service lines, perioperative services, radiology, lab, pharmacy, and all support services were assembled to simulate their jobs in the preview Pavilion.

part bArch

A full-scale mockup of a patient floor (below, photo by Debra B. Foster) offered Penn Medicine staff a realistic sense of how they would move and interact in space at the new Pavilion, something not possible with images such as floor plans and renderings (above). 

The first bus of simulation participants pulled up to the warehouse in late December 2015. Bundled in coats, hats and gloves (the mockup had no centralized heating), they took in the maze of white foam with real and fake medical equipment positioned throughout. After an introduction, each participant began acting out a basic workflow specific to his or her typical routine.

The simulations were designed to range from simple workflows to highly complex ones.

A nurse’s simplest simulation might begin with their arrival at work—exiting the elevator, putting a lunch in the refrigerator, clocking in—through to collecting assignments for the day, greeting patients, refilling medications, and picking up fresh linens, supplies, or lab testing results. Another simulation workflow for the same nurse might be more complex—consulting with a physician and discussing X-rays with a radiologist, assisting a patient from the bed to the bathroom. By the nurse’s final—and most complex—round of simulations, he or she is working with staff from just about every department. The nurse rounds with an interdisciplinary team in a patient room, while environmental services staff move garbage through the hallways, patient lunches are delivered, and physical therapists work nearby. It’s a chance to see how work would be impacted with everything happening simultaneously in the unit.

Lynn Schuchter, MD, chief of the Division of Hematology and Oncology, was among the simulation participants. Though she’s been through half a dozen new office spaces and clinics during her time at Penn Medicine, Schuchter said reading floor plans or walking through a single-room mockup were inferior to simulating lifelike workflows. “You experience the space in a very real and different way,” she said. “It was real life. People took the exercise seriously.” Schuchter said she was staggered by the level of commitment to the simulation process.

As participants pretended their way through the mockup, PennFIRST and Penn Medicine Academy team members were the exercise’s eyes and ears. Tasked with collecting observational data from the simulations and producing actionable insights for the PennFIRST team, a performance improvement consultant employed a combination of monitoring methods. Participants were encouraged to think out loud during their simulations, while note-takers recorded what they saw and heard. Facilitators lobbed questions about different aspects of the space. (For instance, did the Pavilion design make it difficult for nurses to keep tabs on their patients?) Stationary GoPro cameras were recording video throughout the mockup (one captured footage of an environmental services employee struggling to move a garbage cart to its designated area), while videographers made the rounds. More data was collected after the simulation sessions through online surveys, focus groups on specific tasks and topics, such as elevators, and debriefing sessions where participants verbalized their thoughts and wrote them on sticky notes.


Getting Back to the Drawing Board

Within about a month, the Phase 1 simulation sessions ended—and the number crunching began. As the Penn Medicine Academy team sifted through some 5,500 comments, they grouped them into common themes. It didn’t take long to make perhaps the most striking discovery: Though tour and simulation participants took the same survey on the mockup design, their responses were statistically different. Simulation participants voiced many more negative comments than those who only took the tour, said Gretchen Kolb, director of simulation. “In some spaces, no one had anything nice to say,” she said. “The fact that it was so different was really eye opening.”

The PennFIRST team pored over the feedback, and set to redesign the building, during an intense six weeks in early 2016, said project manager Lauren Valentino, MHA(c). They emerged with a significantly redesigned inpatient unit, one that fundamentally changed the entire Pavilion. “It was essentially designed from the inside out,” Valentino said.

The first inpatient design already had large patient rooms—all sized for intensive care—and thus are “conversion ready,” meaning any room could be adapted for different acuity patients without the expense and hassle of extensive renovations. That fit the PennFIRST team’s overarching mantra for the Pavilion to be “future-proof” or adaptable to changing health care needs and technologies for decades to come. But the arrangement of those rooms was in flux even before the mockup. It became clear during the simulations that having two 32-bed units per floor, separated by a common elevator lobby, would not be adaptable enough; the 32-bed units were too large if all were used for critical care, where demands on staff are higher and the patient-to-staff ratio is lower, said Chris Bormann, architect and director of health for the design firm HDR. The new, post-simulation design included three 24-bed units with two elevator cores, a reconfiguration that took a full floor off the Pavilion and created an entirely new building footprint—an entirely new exterior shape.


(left) Kathy Gallagher, MS, BSN, RN, NE-BC, a clinical liaison for PennFIRST, pictured at left with Senior Project Manager Stephen Greulich, works with Penn Medicine staff such as teams from security and the Emergency Department (ED) to plan their future operations in the new Pavilion emergency room. Tape on the floor delineates the security check zone at the entrance. 

(Above)Using pushpins, strings, and sticky notes, ED staff worked with Gallagher to map out the flow of a patient’s journey through the department to plan for changes in the new space.















One of the biggest interior changes was the location of patient bathrooms, which were originally located on each room’s headwall, closer to the patient’s pillow. During the simulations, as clinicians mimicked assisting a patient from bed to bathroom, they had difficulty navigating around the equipment next to the bed. As a result, the bathrooms were moved to the footwall.

Those changes were just the beginning. Elevators were centralized. Corridors became easier to navigate and more natural light shone into the building’s clinical areas. Solid walls around offstage support areas were replaced with panes of vision glass to make staff less isolated.


Finding Flaws—In Time to Fix Them


After more than 600 Penn Medicine employees toured a life-size mockup of half of an inpatient floor of the Pavilion, feedback was overwhelmingly positive.

But when it came to identifying practical flaws in the design, taking a tour was no match for simulating real-world tasks.

After a tour, most staff rated the patient elevators’ location and route to rooms as “ideal.”

But after simulating their real work activities in the mockup, barely half thought the patient elevators were ideally situated.

This is just one example of how, in a 30,000-foot life-size mockup of the space, simulation revealed design flaws that a facilitated tour did not. The interactive experience of simulating regular work activities in the mockup yielded more qualitative and quantitative data and identified more opportunities for design modification than a tour alone. Other key themes that required changes included difficulties with the entrance to the unit, location of the patient bathrooms within the patient rooms, clinical support spaces, teaming spaces, unit-specific spaces, isolation rooms and functional needs not previously considered by the design team.



When that feedback was returned to the design team, the entire project team went back to the drawing board.


The original layout had two 32-bed inpatient units per floor, separated by a common elevator lobby core. After Penn Medicine staff in a variety of roles ran through simulated work activities in the full-scale styrofoam mockup of half a patient floor, they noted problems with the elevator locations. Simulation participants also pointed out that the units were too large for critical care, where demands on staff are higher and the patient-to-staff ratio is lower.



Taking all the feedback into account, the design team substantially reconfigured the space. Now, in addition to changes within patient rooms and within the units, each patient floor of the Pavilion has three 24-bed units with two elevator cores. The reconfiguration took a full floor off the building and created an entirely new footprint and exterior shape.


Setting the New Stage

By April 2016, the original warehouse mockup was struck down and rebuilt to reflect the post-simulation revised design. “This is what we all feel is the most remarkable thing: they took our feedback,” Schuchter said. “They valued everybody’s view and they redesigned it.” The revised mockup was ready for a second simulation round with a cohort of 57 people. This time, the reaction was completely flipped. Participants praised the new design with comments like, “Wow, that was my idea. You changed that because I said there was a problem.”

These new attitudes were also reflected in the data. Though they were asked mostly the same questions as in Phase 1, participants’ feelings about the Pavilion design changed drastically. All told, the second simulation cohort overwhelmingly believed the modified design would provide a positive experience for employees (90 percent), visitors (96 percent) and patients (98 percent). The PennFIRST team was happy, too. Participants still had suggestions on how to further improve the design, but the suggestions required only smaller-scale modifications, not the larger-scale redesign instigated by the first round of simulations. Instead of relocating the elevators, for instance, the team was asked to make them larger.


A life-size mockup of an operating room occupies a corner of the full-floor co-location space where the cross-disciplinary PennFIRST team works together to plan for the Pavilion. Pictured are Project Manager Lauren Valentino, Clinical Liaison Kate Newcomb-DeSanto, MSN, RN, MSW, and Senior Project Manager Stephen Greulich.

The simulations and mockup were repeated with a proposed Pavilion perioperative unit before they were demolished and materials recycled. In the end, the entire mockup and simulation process cost $785,000. That’s about half a percent of the Pavilion’s $1.5 billion budget. “The cost we put in paid for itself,” Mahoney said, “because we designed the building a different way.”

With the overall inpatient design finalized, the PennFIRST team launched into more detail-oriented decisions. At each subsequent step, they are using evidence-based design techniques to lead the way. For instance, they decided on same-handed patient rooms (that is, rooms that are identically oriented, rather than mirror images of each other) throughout the Pavilion based on research that such rooms were more efficient for staff and help reduce errors in emergency situations. In the Pavilion, the patient is always on the left, the bathroom is always on the right and the supplies are always in the same place. Similarly, evidence-based design was used in the decision to forego a centralized caregiver station in exchange for multispecialty collaboration spaces and nurse workstations located closer to patients just outside their rooms with windows for observation.


A Design to Shape Experience—for Everyone

As the simulations focused largely on how Penn Medicine staff would interact with the Pavilion, there was also a parallel effort underway, using human experience mapping, to make the building more accessible and hospitable for a broad range of patients and visitors. Using a demographic database of Penn Medicine patients, a PennFIRST team created several unique “personas,” fictional patients who would visit the Pavilion. One was Lillian, an older woman who cared for her grandchildren in a multi-generational household. Arriving at the Pavilion’s emergency department with chest pains is the beginning of Lillian’s, and her family’s, interaction with the new hospital.

In large workshops at the PennFIRST space, nurses, physicians, radiology technicians, administrators, dietary staff, and former patients imagined the most comfortable ways of moving Lillian and other personas through the Pavilion. What would Lillian’s journey look like? What would her family need while she was in the hospital? How could they get help with parking, or making follow-up appointments or getting prescriptions filled? What aspects of the building’s design would help—or hinder—their experience?

Another group is considering how the Pavilion will affect the feelings of those who work there. Lisa Bellini, MD, GME’94, vice dean for academic affairs, is part of a team endeavoring to include well-being spaces on every floor of the Pavilion—private rooms for programs and services to help clinicians handle stressful moments and avoid burnout. The spaces are expected to include mothers’ rooms for lactation, places for night-shift workers to sleep and multi-functional rooms for meditation groups and other gatherings. Also part of the plan are general wellness rooms where clinicians can take a timeout after a difficult family meeting or adverse event. “There’s no private place to go to take a step back and decompress,” Bellini said. “These places would really try to address this issue.”



The nature of work itself in the Pavilion is on the planning agenda, too. Through “future state workshops,” the PennFIRST team encourages staff to think about the ways they currently work with the goal of creating best practices. A bonus: some groups identified practices they could implement immediately, Valentino said.

With three more years until the opening of the Pavilion, Penn Medicine magazine has only just begun to tell its story. Look for more coverage in future issues on topics that include the building’s economic impact, patient experience, and technologies. Receive regular updates on these and other news stories that matter to you by signing up for our email subscriptions at

The Pavilion will also be a training ground for the next generation of Perelman School of Medicine students, said J. LarryJameson, MD, PhD, the school’s dean and executive vice president of the University for the health system. While gaining knowledge, medical students are also learning attributes of clinical care, he said, and in the new Pavilion will watch teams deliver high-quality care in a building designed for that purpose. “We’re transferring more than just knowledge and technical skills to the students,” he said. “The learners will be practicing the future of medicine from the very beginning based on this design.”

With about three years of Pavilion work completed—and another three to go—there’s still much work to be done before the hospital is fully constructed and occupied. The simulation team is expecting to play a role in the Pavilion transition, helping onboard employees to the space and orienting providers to their new workflows. In the meantime, the team is taking their success at the Pavilion to other outposts of Penn Medicine, including emergency preparedness and rapid response simulations at Penn Medicine Cherry Hill, and helping to inform the design of the new spine center at Pennsylvania Hospital. “The proof of concept was a positive one,” Morgan said. “I think there will be a willingness to use simulation moving forward.”

Meanwhile, PennFIRST team members are sharing their learning from the Pavilion project beyond Penn Medicine by speaking at conferences across the country about their innovative work, Mahoney said. “As a university, part of our role is to disseminate knowledge,” he said.

The team also continues to work as an integrated unit, innovating along the way. Because of the integrated project delivery contract, Bormann said, HDR still has about 30 employees involved in the Pavilion helping to save money and assisting with minor redesigns. And the PennFIRST team continues to run focus groups with Penn Medicine employees to nail down more specifics of the Pavilion design, such as room adjacencies and in-room outlets and utilities. There are full-scale mockups of a patient room, operating room, and emergency department bay in their Market Street co-location space; these continue to serve as a venue for additional prototyping and testing. The team will later add real furniture and design finishes under consideration to gather feedback on selections. As Valentino put it, “I don’t think users will ever not be part of the process.” The opportunity to discover, from real people’s behavior, how to design for the way they work best, is too important to pass up.

“We see this new Pavilion as both a reflection of all the things we’ve been doing well over the last 10 to 15 years and a statement of where we’re going to be in the next 100 years,” said Ralph Muller, CEO of the University of Pennsylvania Health System. “It’s a capstone to the building of new outpatient care space, new translational medicine space, new space for our medical students, and now a new hospital that will be state of the art, worldwide.”

Sign up here to subscribe to the email edition of Penn Medicine magazine. 

Share This Page: