N3C Study Shows Health Factors Before COVID-19 Infection May Predict Long COVID Risk

September 10, 2024

Scientists studied electronic health records from 55,257 people in the N3C database. (Shutterstock)

A person’s use of health care services before a COVID-19 infection may be the strongest predictor of whether they will get long COVID after an infection. These findings are according to a study of electronic health records (EHRs) from 55,257 people in the National Clinical Cohort Collaborative (N3C) database. The study results could help clinicians better focus preventive efforts on people at greater risk for getting long COVID.

Researchers at the University of California, Berkeley, and their colleagues conducted their study as part of the NIH Long COVID Computational Challenge. Long COVID includes a range of long-term symptoms that people have after acute COVID-19. The N3C database has EHRs from more than 8 million people who have had COVID-19.

The scientists used a machine-learning tool called SuperLearner to assess N3C EHRs. The tool correctly predicted whether people would go on to have long COVID. The individual risk factor with the strongest long-COVID predictive power was having a greater number of health care visits overall, more so than visits for any one specific medical issue. The time period with the strongest long-COVID predictive power began more than 37 days before the person got COVID-19. That period was a better predictor than the time periods right before, during or after the person had COVID-19. Other key risk factors included the following:

• Having a greater number of health care visits during acute COVID-19
• Having a viral lower respiratory infection during acute COVID-19
• Being an older age
• Having a greater number of health care visits after COVID-19
• Using systemic corticosteroids before getting COVID-19

The study appears in JMIR Public Health and Surveillance.