Researchers at Wake Forest University School of Medicine have modified an early-warning tool that is commonly used to determine if hospitalized patients are getting sicker. With these changes, researchers may have developed a way for busy emergency departments (ED) to assess the risk of incoming patients and guide the critical decision of whether patients requiring hospitalization should be placed in an intensive care unit (ICU) or a standard room.
Determining patient risk in an emergency department can be difficult, said senior author Chadwick Miller, M.D., M.S., an assistant professor of surgical sciences and emergency medicine. Clinicians often have less than a couple of hours to observe a patient, he added – and things can change minute-by-minute in the ED.
“In the fast-paced emergency department setting, it would help to have a tool that can quickly determine whether patients should be admitted to a normal bed or to a higher level of care,” Miller said.
According to the Centers for Disease Control and Prevention, more than 119 million patients were treated in U.S. emergency departments in 2009. Being able to more accurately predict the level of care they’ll require could have a significant effect on quality of care, Miller said. For example, correctly assigning a patient to an intensive care unit means access to early, more aggressive care that can prevent deterioration of the patient’s medical condition.
Miller and colleagues studied an existing hospital-based tool called the modified early warning score (MEWS) that includes a series of questions about the patient’s health state. It has typically been used to determine if hospitalized patients are getting sicker.
The researchers reviewed records from 280 patients being admitted to the hospital from the ED. Using MEWS, 82 percent of the patients fell into the category of “intermediate risk,” a gray area where it can be unclear whether a higher or a lower level of care is needed. MEWS predicted that 18 percent of patients required intensive care, but a look back at patients’ records told the team that, in fact, 27 percent of the patients ultimately needed a higher level of care.
In addition to under-predicting the percentage of patients needing intensive care, MEWS also under-performed in predicting the number of patients at “low risk,” assigning none of the nearly 300 patients to the category and placing them instead in that “gray zone” that could lead to expensive, unnecessary and potentially harmful over-treatment.
The analysis showed that to be used in the ED setting – where accuracy and appropriate judgment calls are crucial – MEWS would need to be re-worked. The research team added key bits of information to the tool – questions about whether the patient arrived at the ED via ambulance, received intravenous antibiotics in the ED, the patient’s length of stay in the ED, and gender. They found that the revised model, dubbed MEWS Plus, performed better than the original MEWS.
“Its ability to accurately classify people as low-risk is much improved,” Miller said. The MEWS Plus model classified about 22 percent more patients in an appropriate risk category than the original MEWS model did, Miller cautioned, however, that future investigations are needed to evaluate whether this tool should be widely implemented.
The MEWS Plus research team, funded in part by the Division of Healthcare Research and Quality Outcomes of Wake Forest University Baptist Medical Center and the Wake Forest University Translational Science Institute, included John Galliard, M.D., Howard Blumstein, M.D., Doug Case, Ph.D., and Catherine Messick, M.D., all of the School of Medicine, and Corey Heitz, M.D., now a faculty member of Wright State University.