University of Massachusetts Amherst

Search Google Appliance


Making ERs Work Better

Hari Balasubramanian

Hari Balasubramanian, assistant professor in our Mechanical and Industrial Engineering Department, is working with emergency department directors and faculty at UMass Memorial Medical Center to identify bottlenecks, improve efficiency, and reduce patient waiting time in the emergency room. They plan to assess the usefulness of two engineering methods — computerized discrete event simulation (DES) and wireless identification technology, also known as radio-frequency identification (RFID) — to study how variables interact, such as doctor and nurse staffing levels, intake flow and procedures, and injury or disease severity, to affect efficient use of resources.

They are using mathematical models and wireless tagging techniques to study hospital emergency room operations under a two-year $150,000 Life Sciences Moment grant from the President’s Office and the Medical School.

Catching a plane on time, we seldom think about the systems planning required to achieve that success, as Balasubramanian says. “When things work, we seldom ask why. But many operations management researchers are asking exactly this question, as well as why things go wrong. They are also designing and testing more efficient operations models.”

A new faculty member, Balasubramanian is teaming up with Eric Dickson, Katherine Harrison, and others at UMass Memorial Medical Center.

As Balasubramanian explains, DES modeling software uses actual time-motion data on patient and clinician activities, which can be used to simulate a variety of scenarios that can confront ER staff. Once ranges have been established for the variables, the modeler can ask the computer to run 200 or 250 randomly created mock patients through the system, basically simulating a busy day at the ER.

“It’s understood that a model will never capture reality perfectly what happens in a real emergency department,” Balasubramanian points out. "But a successful one should match it with a minor error rate."

With the DES model established, the research team and emergency department directors can run various scenarios and experiment with targeted interventions to test the effects of, for example, moving intake and registration tasks from the patient’s first ER interaction or event to later, at the bedside, without having to experiment in real time and put real patients at risk.

In a second phase of this study, Balasubramanian and colleagues plan to tag patients with RFID bracelets to track their progress through the system using multiple time stamps. These data should help to further test and validate patient outcomes, including time to discharge or assign a patient to an inpatient bed, for example, and to identify bottlenecks and treatment delays, which may put patients at higher risk. This study phase should help Dickson, Balasubramanian, and colleagues understand patient flow in an emergency department at a level of detail not yet described in the medical literature, they say. They hope to begin initial RFID tagging this spring and refine deployment including tests of interventions next summer.

In the future Balasubramanian, Dickson, and colleagues hope to include collaborators in other emergency departments, such as the one at Baystate Medical Center in Springfield. (February 2010)