Classifying and Differentiating Clinical Coding of Respiratory Syncytial Virus, Influenza, and COVID-19 in OpenSAFELY Between 2016 and 2024

Broad computable phenotypes for RSV, influenza, and COVID-19 in English EHRs (Sep 2016–Aug 2024) increased misclassification risk for mild cases versus narrow phenotypes.

• Why it matters: Testing data are often missing in coded records, creating a need for reliable phenotypes that can still distinguish respiratory virus infections using computable definitions.
• What they did: Using OpenSAFELY with NHS England approval, the team designed and assessed narrow and broad phenotypes for RSV, influenza, and COVID-19, comparing their seasonal patterns to surveillance data and evaluating misclassification by severity and age.
• The result: Seasonal timing matched surveillance for both phenotype types, but broad definitions misclassified mild cases more often; for severe cases, misclassification risk was higher in infants than older adults, enabling virus classification without testing information.

AI summary of the abstract · may contain errors · not medical advice · read the paper.