But recognition rarely provides a resolution; figuring out exactly how to address a problem is often the most difficult part of finding a solution. Trial and error is generally not an acceptable approach to complex business dilemmas.
How, then, can businesses find solutions to pressing issues that call for change? Professor Edward Wasil uses mathematical models that draw on both quantitative and qualitative data to solve practical business problems. In short, his research specializes in high-tech problem solving.
Wasil, who serves as chair of the Information Technology department, illustrates the nature of his research using the “traveling salesman” problem.
A traveling salesman starts at an initial location, visits a number of cities, and then returns to the location in the shortest amount of time and distance. Taking all variables into account, what is the most efficient way for the salesman to complete his tour?
“It’s a classic problem that’s been around since the 1700s,” Wasil said. “Everybody has tried their hand at it, and it turns out to be a very difficult problem to solve.”
A modern-day equivalent of the traveling salesman is behemoth UPS — whose tagline, “We ♥ logistics™,” could easily apply to Wasil. The shipping service handles roughly 16 million deliveries each day, many of which need to reach their destination within 24 hours.
A Turning Point
Recently, Wasil’s research has turned to the health-care realm: Two of his latest projects tackle intricate hospital administrative hurdles, such as surgical scheduling and lowering the transmission rates of infectious diseases.
Wasil drew a comparison between hospital efficiency and that of Disney theme parks, where rides are constructed with the goal of moving guests through quickly. Hospitals need to employ the same tactics to get patients in and out efficiently.
“At the operational level, hospitals are dealing with these issues on a daily basis,” Wasil said. “They are just beginning to realize that they can rely on sophisticated operations-research techniques to help them find good solutions to their important problems.”
Wasil’s work in operations research focuses on network optimization. His mathematical models help businesses solve the problem of optimizing the delivery of goods and services within a particular set of constraints.
While optimization is the ultimate goal, the models Wasil and his co-researchers formulate often produce very good, near-optimal solutions in a short amount of computing time.
Human behavior is a major factor, and adds to the challenge of creating a model that mimics reality. Working through all possibilities is critical if you hope to arrive at an applicable solution, he added. “You’re aiming for a very good solution,” Wasil said. “You want to create something that is provably good and better than what you started with.”
Reducing a Bottleneck
Working with the University of Maryland’s Medical Center in Baltimore (UMMC), Wasil and his co-authors created a model to streamline the flow of cardiac surgery patients.
Immediately after surgery in an operating room, patients are taken to a Post-Anesthesia Care Unit (PACU) to recover. Patients usually stay in the PACU for about an hour to let the anesthesia wear off. From there, they are moved to an intensive care unit (ICU) and then to a step-down unit after a few days.
When ICU beds are not available, patients remain in the PACU, effectively “boarding” until beds downstream open up. The situation is undesirable to say the least: it requires unplanned overnight shifts for staff, can lengthen patients’ stays, and imposes extra costs. Wasil found that as many as 1,000 patients were boarded in the PACU at UMMC annually in 2007 and 2008.
Using hospital data, Wasil concluded that the arrival of patients for surgery was not matched with an equal number of discharges—a result of surgeries being scheduled in service blocks. Each block gives a particular service line (orthopedics, vascular, oncology, etc.) a specific operating room on a specific day. Individual surgeons are then assigned operating time for their cases within the relevant block.
“Block schedules tend to encourage thinking at the service-line level, instead of at the hospital level, and they do not incentivize desirable schedules in terms of hospital efficiency,” Wasil wrote.
To alleviate the boarding problem, the team initially proposed a three-phase scheduling approach that used integer programming and simulation to improve the flow of patients through the system by attempting to balance arrivals and departures. However, conflicting stakeholder preferences were a barrier to successful implementation, revealing more constraints on the solution. “There are lots of preferences physicians have—having clinicals on a specific day, for example—that are constraints you have to deal with,” Wasil said.
In response, rules of thumb were developed to build in flexibility and create a final, workable schedule. The team’s revised version highlighted the need for a system-wide approach to hospital performance, rather than individual departments (for example, orthopedics) operating in silos. It also revealed that a small number of alterations can have a significant impact in reducing the number of boarders.
UMMC has considered pieces of the model, and is looking into how to use the rules-of-thumb approach to adjust block schedules. “People get used to doing things a certain way and change becomes very difficult,” Wasil said. “Models can only propose solutions; people and businesses then have to adopt them.”
Professor Wasil, chair of the Information Technology Department, specializes in network optimization and applications of decision-aiding methods. He has published more than 100 technical articles in a wide variety of academic outlets including Operations Research, European Journal of Operational Research, and Manufacturing and Service Operations Management.
Wasil won the 2011 American University Award for Outstanding Scholarship for his dedication to and excellence in research.
In another project, Wasil and his co-authors built a model intended to examine the spread in hospitals of a bacterial infection known as methicillin-resistant Staphylococcus aureus (MRSA).
Hospital-acquired infections result in nearly 100,000 deaths per year, according to the Committee to Reduce Infection Deaths, an educational campaign founded by Betsy McCaughey, a former lieutenant governor of New York. Approximately 20,000 of those deaths involve MRSA. The infection’s spread is exacerbated by changes in hand-washing compliance among health-care workers (HCW); however, one of the team’s findings indicates that even total hand-washing compliance would not solve the MRSA problem on its own—although it is vital.
Unlike the PACU boarding model, the MRSA model used agent-based modeling and simulation to study the dynamics of MRSA transmission within a hospital. With this type of modeling, all players or agents, including the setting, are created in the simulation.
The model investigates the effectiveness of patient isolation—placing an infected patient in a single room; patient decolonization, which means removing the presence of bacteria on a patient’s skin; and improvement of HCW-to-patient ratios, decreasing the rate of transmission by lowering the number of patients with which a HCW engages.
The simulation results show that patient isolation is the best overall means of reducing transmission other than a one-to-one HCW-to-patient ratio — an idyllic state that is fiscally impossible. These findings were consistent with results published in relevant literature.
“Often a whole system can function better if the people behind the system are willing to change,” Wasil said.
After establishing the best-performing measure, Wasil and his team were next able to look at questions of significance to hospitals, including whether nurses or physicians are responsible for more MRSA transmissions and how isolation can best be applied to different hospital settings.
Nurses typically see patients more frequently than physicians, but physicians see a higher total number of patients, albeit less frequently. Contrasting service patterns make it difficult to predict which group is the primary source of transmission.
“The question of who is responsible for more transmissions is important so that hospitals can focus educational programs where they would have the most significant impact,” Wasil wrote. “Both populations [nurses and physicians] have different cultures and varying degrees of interaction with patients and therefore would require a different approach to reduce transmission.”
The model showed that the type of unit and type of HCW taken together can most accurately predict transmission dynamics.
Based on the results, the team honed in on three key recommendations: minimize the size of patient cohorts; screen for and isolate colonized and infected patients; and decrease the likelihood of transmission between HCWs and patients by decreasing contact and enforcing the use of gloves and gowns around colonized infected patients.
The team also found that both nurses and physicians are the roots of the problem, depending on the setting.
In a general ward, nurses account for 80 percent to 90 percent of contact with patients, making them more likely to spread the disease.
But in the ICU, nurses have fewer patients—sometimes a “cohort” of only two—whereas physicians may be responsible for up to 10 patients, who could have five different nurses assigned to them. As a result, physicians pose a greater threat in the ICU, since they can spread MRSA from one patient cohort to another. “The results of [our] experiment show that physicians account for almost all transmission in the ICU, except at extremely high values of hand-hygiene compliance,” Wasil and team wrote.
The Next Challenge
Wasil and his co-authors have already begun studying the transmission of MRSA between hospitals and long-term-care facilities. “We know MRSA is being spread by patients going back and forth,” Wasil said. “We’re looking at how to stop that from happening.”
He also plans to research the possible connection between surgeons’ discharge practices and patient readmissions to see if discharging patients too early is responsible for patient “bounce-backs.”
Beyond health care, Wasil is always on the lookout for a new problem. Like this one: assigning routes to snowplows in Colorado with as few uphill stretches as possible, thus making the job of clearing snow easier and more efficient.