Image: 4X-image/iStock/Getty Images
A new artificial intelligence-based score considers multiple factors to predict the prognosis of individual patients with COVID-19 seen at urgent care clinics or emergency departments.
The tool, created by Harvard Medical School investigators based at Massachusetts General Hospital, can be used to rapidly and automatically determine which patients are most likely to develop complications and need hospitalization.
Reporting in The Journal of Infectious Diseases, a team of experts designed the COVID-19 Acuity Score (CoVA), a machine-learning model based on data from 9,381 adult outpatients seen in Mass General’s respiratory illness clinics and emergency department between March 7 and May 2, 2020.
“The large sample size helped ensure that the machine learning model was able to learn which of the many different pieces of data available allow reliable predictions about the course of COVID-19 infection,” said M. Brandon Westover, HMS associate professor of neurology and director of data sciences at the Mass General McCance Center for Brain Health.
Westover is one of three co-senior authors of the study, with Gregory Robbins, HMS associate professor of medicine at Mass General, and Shibani Mukerji, HMS assistant professor of neurology and associate director of Mass General’s Neuro-Infectious Diseases Unit.
The researchers tested CoVA on data from another 2,205 patients seen between May 3 and May 14. In this prospective validation group, 26.1 percent, 6.3 percent and 0.5 percent of patients experienced hospitalization, critical illness or death, respectively, within seven days.
CoVA demonstrated excellent performance in predicting which patients would fall into these categories.
“Testing the model prospectively helped us to verify that the CoVA score actually works when it sees ‘new’ patients in the real world,” said study first author Haoqi Sun, HMS instructor in neurology and a research faculty member in the Mass General Clinical Data Animation Center.
Among 30 predictors, which included COVID-19 testing status, vital signs, medical history, chest x-ray results and demographic categories such as age and gender, the top five predictors were age, diastolic blood pressure, blood oxygen saturation, COVID-19 testing status and respiratory rate.
“While several other groups have developed risk scores for complications of COVID-19, ours is unique in being based on such a large patient sample, in being prospectively validated and in being specifically designed for use in the outpatient setting, rather than for patients who are already hospitalized,” Mukerji said.
“CoVA is designed so that automated scoring could be incorporated into electronic medical record systems. We hope that it will be useful in case of future COVID-19 surges, when rapid clinical assessments may be critical,” Mukerji added.
The authors of the study include experts in neurology, infectious disease, critical care, radiology, pathology, emergency medicine and machine learning. The impetus for the study began early in the U.S. epidemic when Massachusetts experienced frequent urgent care visits and hospital admissions. As an infectious diseases physician and as part of the Mass General Biothreats team, Robbins recognized the need for a more sophisticated method to identify outpatients at greatest risk for experiencing negative outcomes.
Adapted from a Mass General news release.
