A team at Massachusetts General Hospital (MGH) have developed a new artificial intelligence-based method which they claim can predict the prognosis of individual patients with Covid-19 admitted to hospital.
The tool, nicknamed CoVA, can – according to its developers – be used to rapidly and automatically determine which patients are most likely to develop complications and need to be hospitalised.
Dr Gregory Robbins, an infectious diseases physician on the MGH Biothreats team, said he recognised the need for a more sophisticated method to identify outpatients at greatest risk for experiencing negative outcomes.
A wide-ranging team of experts in neurology, infectious disease, critical care, radiology, pathology, emergency medicine and machine learning designed the Covid-19 Acuity Score (CoVA) based on input from information on 9,381 adult outpatients seen in MGH’s respiratory illness clinics and emergency department between 7 March and 2 May, 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 Dr Brandon Westover, an investigator in the Department of Neurology and director of Data Science at the MGH McCance Center for Brain Health.
CoVA was then tested in another 2,205 patients seen between 3 May and 14 May.
Among the 30 predictive factors are demographics like age and gender, Covid-19 testing status, vital signs, medical history and chest X-ray results (when available).
The top five predictors are age, diastolic blood pressure, blood oxygen saturation, Covid-19 testing status and respiratory rate.
The team behind CoVa say this AI risk tool is unique in being based on such a large patient sample, being prospectively validated, and in being specifically designed for use in the outpatient setting, rather than for patients who are already hospitalised.
The full background to, and testing of, CoVa can be found in the Journal of Infectious Diseases here: https://academic.oup.com/jid/advance-article/doi/10.1093/infdis/jiaa663/5938525