Patient and Epidemiological Factors Associated With Influenza Testing in Hospitalized Adults With Acute Respiratory Illnesses, 2016-2017 to 2019-2020

Alexandra F Dalton,Alexia Couture,Malini B DeSilva,Stephanie A Irving,Shruti Gohil,Suchitra Rao,Rebecca V Fink,Allison L Naleway,Zijing Guo,Devi Sundaresan,Rebecca J Birch,Sarah Ball,Kai Zheng,Toan C Ong,Carrie Reed,Catherine H Bozio

DOI: https://doi.org/10.1093/ofid/ofad162

2023-03-25

Abstract:Background: Data are limited on influenza testing among adults with acute respiratory illness (ARI)-associated hospitalizations. We identified factors associated with influenza testing in adult ARI-associated hospitalizations across the 2016-2017 through 2019-2020 influenza seasons. Methods: Using data from 4 health systems in the United States, we identified hospitalizations that had an ARI discharge diagnosis or respiratory virus test. A hospitalization with influenza testing was based on testing performed within 14 days before through 72 hours after admission. We used random forest analysis to identify patient characteristics and influenza activity indicators that were most important in terms of their relationship to influenza testing. Results: Across 4 seasons, testing rates ranged from 14.8%-19.4% at 3 pooled sites and 60.1%-78.5% at a fourth site with different testing practices. Discharge diagnoses of pneumonia or infectious disease of noninfluenza etiology, presence of ARI signs/symptoms, hospital admission month, and influenza-like illness activity level were consistently among the variables with the greatest relative importance. Conclusions: Select ARI diagnoses and indicators of influenza activity were the most important factors associated with influenza testing among ARI-associated hospitalizations. Improved understanding of which patients are tested may enhance influenza burden estimates and allow for more timely clinical management of influenza-associated hospitalizations.