From the beginning of the COVID-19 pandemic, one of the ongoing challenges for providers has been to predict the severity of the cases they are encountering as patients are admitted to the hospital.

To that end, a recent study by researchers at the National Institutes of Health has tapped the largest data repository, roughly two million medical records, in an attempt “to develop a clinically useful model that accurately predicts severity using data from the first day of hospital admission."  

The cohort included 174,568 adults who tested positive for SARS-CoV-2, of which 18.6% were hospitalized. Of those, 6,565, or about a fifth, followed what researchers called a "severe clinical course": invasive ventilatory support, extracorporeal membrane oxygenation, a discharge to hospice or death.

In the study, which was published in JAMA Network Open,  “the team found that inpatient mortality decreased over time from March and April 2020 to September and October. Treatment patterns also changed, with use of antimicrobial and immunomodulatory medications shifting over the course of the pandemic.”

Moreover, the researchers determined “that patient demographic characteristics and comorbidities were associated with higher clinical severity. The machine learning models accurately predicted ultimate clinical severity using commonly collected clinical data from the first 24 hours of a hospital admission.”

For the study, the researchers relied on the Data Enclave of the National COVID Cohort Collaborative, or N3C, which was formed to accelerate understanding of SARS-CoV-2 and develop a novel approach for collaborative data sharing and analytical data during the pandemic. The N3C is composed of members from the National Institutes of Health Clinical and Translational Science Awards Program and its Center for Data to Health, the IDeA Centers for Translational Research, the National Patient-Centered Clinical Research Network, the Observational Health Data Sciences and Informatics network, TriNetX, and the Accrual to Clinical Trials network.

As a result, the researchers noted, “this report provides a detailed clinical description of the largest cohort of US COVID-19 cases and representative controls to date. This cohort is racially and ethnically diverse and geographically distributed.”

One of the study’s primary outcomes, the team explained, was the establishment of “expected trajectories for many vital signs and laboratory values among patients with different clinical severities. Expected trajectories can contribute to practitioner decision-making about what a patient will need.”

The team noted the new models could act as a basis for generalizable clinical decision support tools, but they cautioned that development of such tools would require additional work at deploying healthcare systems.

Photo by sanjeri/Getty Images