|
|
|
A COVID-19 risk prediction tool developed at UMass Medical School and deployed at the DCU Field Hospital helped UMass Memorial Health Care hospitals manage the COVID-19 surge in Worcester. The Decompensation Electronic COVID Observational Monitoring Platform Triage (DE-COMP-Triage) provided a score to determine which patients at the field hospital were at highest risk of rapid deterioration and, thus, should be transferred to a regular hospital with an intensive care unit.
Worcester’s DCU Field Hospital was set up to care for patients with mild to moderate symptoms in order to free up hospital beds for sicker patients. But a challenge in caring for patients with COVID-19 is not knowing when those with mild to moderate symptoms might worsen to the point that they require intubation or are so sick that they are likely close to death.
“This collaboration was driven out of the need for guidance for clinicians and case managers, and we needed to move fast,” said David McManus, MD, professor of medicine, who in his capacity as vice chair of medicine for clinical affairs in the Department of Medicine, was a member of the field hospital leadership team. “As we worked on a guiding document that pulls in information about how to care for COVID patients adapted for a field hospital environment, it became clear we needed inclusion and exclusion criteria to triage patients in that environment to UMass Memorial Medical Center for a higher level of care.
McManus worked with MD/PhD candidate Apurv Soni to analyze patient data. Soni found that presentation vitals, age and medical history of type 2 diabetes could be used to calculate DE-COMP-Triage score. The score classifies patients into the categories of low, low-intermediate, high-intermediate and high risk of requiring mechanical ventilation or death.
“We needed to focus on the patient information available at their time of entry and in the first few hours after that, like demographics, vitals, physical exam findings and rapidly obtained lab values,” said Soni. “While they’re not the only ones that are associated with bad outcomes with COVID, these are the ones that are most efficient in stratifying patients with low versus high risk of decompensation. Our objective was to have an evidence-based tool that supports clinician’s decision-making in identifying patients least likely to decompensate and thus eligible for care in low acuity setting.”
As of May 17, half of all COVID-19 patients that presented to the emergency departments at UMass Memorial Health Care hospitals were classified as low-risk and only 1 percent decompensated. By contrast, about 15 percent of patients were classified as high-risk by DE-COMP-Triage score and nearly two-thirds of them decompensated.
The project ramped up quickly with support from the Population & Quantitative Health Sciences and the Office of Research at UMMS; clinical leadership for the field hospital, UMass Memorial and Baystate Health; the EPIC electronic health record team at UMass Memorial; and more than 50 UMMS volunteers. The field hospital operated from March through May.
Soni had experience conducting global health research in India and he completed his PhD dissertation research in the lab of Jeroan Allison, MD, chair and professor of population and quantitative health sciences.
“This was a transformative experience. The generosity of my mentors in bringing me into this work and trusting me with it speaks to the unique environment of UMass Medical School where faculty empower trainees to perform at this level,” he said. “This has been a very tangible experience that has allowed me to merge my research degree in biostatistics and epidemiology with my clinical training in real time.”
Since its initial rollout, DE-COMP-Triage has been put into use in the full-service hospitals. A more sophisticated version of the original model that can predict imminent decompensation in the next 12 hours is being readied for clinical testing through a partnership between UMass Memorial Health Care and the UMass Medical School departments of medicine and population & quantitative health sciences.
Related stories on UMassMedNow:
NIH invests in rapid innovation and development for COVID-19 testing
Startup incubator aims to spur life-saving innovations
LISTEN: How digital devices play a role in screening patients for disease
Soni and Fahey honored by Consortium of Universities for Global Health for best manuscript
