Johns Hopkins develops electronic triage tool to better prioritize ED patients
Algorithm predicts patient outcomes based on a systems engineering approach and advanced machine learning methods.
Johns Hopkins researchers have developed an electronic triage tool to help emergency departments quickly and more accurately determine those patients who are critically ill and to assign priority treatment levels.
The e-triage tool identifies relationships between predictive data and patient outcomes by leveraging an algorithm based on a systems engineering approach and advanced machine learning more commonly used in industries outside of healthcare, such as defense, finance and transportation.
“Machine-based learning takes full advantage of electronic health records and allows a precision of outcomes not previously realizable,” says Gabor Kelen, MD, director of the Department of Emergency Medicine and professor of emergency medicine at the Johns Hopkins University School of Medicine. “Decision aids that take advantage of machine learning are also highly customizable to meet the needs of an emergency department’s patient population and local healthcare delivery systems.”
Also See: ED access to HIE data can boost efficiency and quality
During triage, EDs typically use a subjective assessment called the emergency severity index (ESI) to assign a score from Level 1 for patients who are the most critically sick to Level 5 for patients who are the least sick.
However, in a multi-site retrospective study of almost 173,000 ED visits—published in the Annals of Emergency Medicine—the EHR-embedded tool demonstrated overall equal or improved identification of patient outcomes compared to ESI and more accurately classified ESI Level 3 patients.
“E-triage demonstrates an opportunity to apply advanced predictive analytics to large-scale electronic health record data to support triage decision-making and improve patient risk management in the ED,” the article concludes.
According to Scott Levin, associate professor of emergency medicine at the Johns Hopkins University School of Medicine, out of the more than 65 percent of visits triaged to ESI Level 3, the tool identified about 10 percent—or more than 14,000, ESI Level 3 patients—who may have benefitted from being place in a more critical priority level. These patients were at least five times more likely to experience a critical outcome, such as death, admission to the ICU or emergency surgery, and two times more likely to be admitted to the hospital.
“When a patient comes in and we collect the patient’s information, the e-triage tool is comparing that patient to hundreds of other like patients to make a prediction on the patient’s outcome,” says Levin.
In addition, he notes that the tool was able to increase the number of patients place in a lower priority level—such as Level 4 or 5—to help minimize the number of low-acuity patients waiting for and overusing scarce resources. He contends that the technology has the potential to help with the crisis of ED crowding across the country.
“We have deployed to tool and have been using it for nine months now,” adds Levin, who says e-triage links to Johns Hopkins’ Epic EHR system. “We extract patients’ vital signs, chief complaint information and all their prior medical history, and that is leveraged by the machine-learning algorithm that predicts a patient’s likelihood of dying, being admitted to an intensive care unit, having some sort of emergent procedure and the likelihood of being admitted to the hospital.”
Currently, the internally developed e-triage tool is being used at The Johns Hopkins Hospital and Howard County General Hospital— both member hospitals of Johns Hopkins Medicine—and is expanding to the rest of the health system. But Levin sees the technology being commercialized for implementation by other healthcare organizations.
“We’re in the process of refactoring the tool somewhat so that it can easily plug into other hospital EHRs,” concludes Levin, who is a co-founder and holds equity in StoCastic LLC, a Johns Hopkins start-up that is offering the tool.
Under a license agreement between StoCastic and The Johns Hopkins University, Kelen and Levin are entitled to royalty distributions on the technology described in the study, which was funded by the Agency for Healthcare Research and Quality and the National Science Foundation.
The e-triage tool identifies relationships between predictive data and patient outcomes by leveraging an algorithm based on a systems engineering approach and advanced machine learning more commonly used in industries outside of healthcare, such as defense, finance and transportation.
“Machine-based learning takes full advantage of electronic health records and allows a precision of outcomes not previously realizable,” says Gabor Kelen, MD, director of the Department of Emergency Medicine and professor of emergency medicine at the Johns Hopkins University School of Medicine. “Decision aids that take advantage of machine learning are also highly customizable to meet the needs of an emergency department’s patient population and local healthcare delivery systems.”
Also See: ED access to HIE data can boost efficiency and qualityDuring triage, EDs typically use a subjective assessment called the emergency severity index (ESI) to assign a score from Level 1 for patients who are the most critically sick to Level 5 for patients who are the least sick.
However, in a multi-site retrospective study of almost 173,000 ED visits—published in the Annals of Emergency Medicine—the EHR-embedded tool demonstrated overall equal or improved identification of patient outcomes compared to ESI and more accurately classified ESI Level 3 patients.
“E-triage demonstrates an opportunity to apply advanced predictive analytics to large-scale electronic health record data to support triage decision-making and improve patient risk management in the ED,” the article concludes.
According to Scott Levin, associate professor of emergency medicine at the Johns Hopkins University School of Medicine, out of the more than 65 percent of visits triaged to ESI Level 3, the tool identified about 10 percent—or more than 14,000, ESI Level 3 patients—who may have benefitted from being place in a more critical priority level. These patients were at least five times more likely to experience a critical outcome, such as death, admission to the ICU or emergency surgery, and two times more likely to be admitted to the hospital.
“When a patient comes in and we collect the patient’s information, the e-triage tool is comparing that patient to hundreds of other like patients to make a prediction on the patient’s outcome,” says Levin.
In addition, he notes that the tool was able to increase the number of patients place in a lower priority level—such as Level 4 or 5—to help minimize the number of low-acuity patients waiting for and overusing scarce resources. He contends that the technology has the potential to help with the crisis of ED crowding across the country.
“We have deployed to tool and have been using it for nine months now,” adds Levin, who says e-triage links to Johns Hopkins’ Epic EHR system. “We extract patients’ vital signs, chief complaint information and all their prior medical history, and that is leveraged by the machine-learning algorithm that predicts a patient’s likelihood of dying, being admitted to an intensive care unit, having some sort of emergent procedure and the likelihood of being admitted to the hospital.”
Currently, the internally developed e-triage tool is being used at The Johns Hopkins Hospital and Howard County General Hospital— both member hospitals of Johns Hopkins Medicine—and is expanding to the rest of the health system. But Levin sees the technology being commercialized for implementation by other healthcare organizations.
“We’re in the process of refactoring the tool somewhat so that it can easily plug into other hospital EHRs,” concludes Levin, who is a co-founder and holds equity in StoCastic LLC, a Johns Hopkins start-up that is offering the tool.
Under a license agreement between StoCastic and The Johns Hopkins University, Kelen and Levin are entitled to royalty distributions on the technology described in the study, which was funded by the Agency for Healthcare Research and Quality and the National Science Foundation.
More for you
Loading data for hdm_tax_topic #better-outcomes...