UNIVERSITY PARK, Pa. — Guodong (Gordon) Pang has been awarded $150,000 by the National Science Foundation (NSF) for research that aims to improve discharge predictions within intensive care units (ICUs) and a patient’s subsequent flow through the hospital system.
The collaborative research project, titled “Physiologically Based Optimization of ICU Management,” will be funded for three years.
Collaborators Andrew Schaefer of Rice University and Gills Clermont of the University of Pittsburgh Medical Center have received $216,120 and $33,883, respectively, from the NSF for the work they are contributing to the project.
“This research will integrate stochastic and dynamic models of physiology with patient flow management in order to improve the outcomes of patients admitted to the ICU,” said Pang, associate professor of industrial engineering. “The main objective is to show that effective leveraging of clinical indicators can significantly improve the efficiency of ICU operations.”
Approximately 20 percent of hospital operating costs can be attributed to ICUs and that percentage has been increasing, explained Pang. ICUs are also tightly connected to other hospital units, such as operating rooms, emergency departments, etc.
“Therefore, improved management of ICUs at the strategic level will considerably impact patient flow while improving patient outcomes and reducing health care expenses,” said Pang. “The efficient operation and management of ICUs are critical to providing high quality of care while controlling operating costs.”
Pang, Schaefer and Clermont’s research will introduce a physiologically based stochastic and dynamic Transfer Readiness Score, which can better predict a patient’s outcome. An anticipative bed-request scheme will be developed using the created readiness score to better predict when a patient will likely be ready for transfer out of the ICU.
The research will create a new class of score-based queuing and stochastic network models, in which the service distribution captures the high variability of length of stay in the ICUs and patient physiology. Score-based routing control policies and decentralized network optimization algorithms for the new models will be developed to improve the flow of patients throughout the hospital system.
According to the Society of Critical Care Medicine, more than 5.7 million patients are admitted annually to ICUs in the United States. Approximately 20 percent of acute care admissions are to an ICU and up to 58 percent of emergency department admissions result in an ICU admission. All of these patients share the need for frequent assessment and greater need for technological support compared to patients admitted to non-ICU beds.
Length of stay in ICUs in hospitals within the United States has been estimated at 3.8 days; however, it varies depending on patient and ICU attributes.
“Prolonged stays in ICUs and hospitals, in general, are emotionally and financially difficult on patients and their families,” said Pang. “This new research primarily focuses on improving quality of health care delivery throughout the patient’s hospital stay to try to ease some of the anxieties patients and families face.”