ML model predicts Covid severity, helps in decision-making

A centralised repository of Covid-19 health records built by US researchers, last year, has been helpful in tracking the progression of the disease over time and could eventually be used as the basis for decision-making tools.

New York: A centralised repository of Covid-19 health records built by US researchers, last year, has been helpful in tracking the progression of the disease over time and could eventually be used as the basis for decision-making tools. The National Covid Cohort Collaborative (N3C) is a centralised, harmonised, high-granularity electronic health record repository that is the largest, most representative Covid-19 cohort to date.

This multicenter data set can support robust evidence-based development of predictive and diagnostic tools and inform clinical care and policy, said a team of researchers from those including at Universities of Colorado, Michigan, Rochester Medical Center, and Johns Hopkins. The cohort study, published in the JAMA Network, used data from 34 medical centres and included over 1 million adults -- 174,568 who tested positive for Covid-19 and 1,133,848 who tested negative between January 2020 and December 2020.

"This cohort study found that Covid-19 mortality decreased over time during 2020 and that patient demographic characteristics and comorbidities were associated with higher clinical severity," said Tellen D. Bennett, from the Department of Pediatrics at Colorado's School of Medicine. "The machine learning models accurately predicted ultimate clinical severity using commonly collected clinical data from the first 24 hours of a hospital admission," Bennett added.

Read: World hunger at a 15-year high due to Covid: United Nations

The machine learning models accurately predicted clinical severity -- mortality was 11.6 per cent overall and decreased from 16.4 per cent in March to April 2020 to 8.6 per cent in September to October 2020. The most powerful predictors in the ML models are patient age and widely available vital signs and laboratory values. These models, although intended as examples of how N3C can be used, could also be the basis for generalisable clinical decision support tools, the researchers said.

The N3C harmonises data from a very large number of clinical sites, which is important because significant site-level variation in critical metrics, such as invasive ventilatory support and mortality, has been reported. Expected trajectories can contribute to practitioner decision-making about what a patient will need, Bennett said.

(IANS)