Methods: Administrative records from the 2010–13 influenza seasons were reviewed for ARI hospitalizations using pre-selected diagnostic codes. PCR- and antigen-based panel assays were available. Testing (within 3 days of admission) and antiviral prescriptions were compared by patient age, high-risk conditions, and intensive care unit (ICU) admission. Antivirals prescribed without the aid of diagnostic tests were considered empiric. Season- and age-adjusted rates were estimated from Poisson regression.
Results: Among 27,137 ARI hospitalizations, 33% were tested for influenza, 9% were positive, and 2% were prescribed influenza antivirals. Overall, 25% of tests were PCR-based. Treatment was strongly associated with testing (rate ratio=15.0, p-value<0.01). While testing was greater among young children (64%) than adults ≥65 years (32%), antiviral treatment was greater among older adults (2% v. 0.8% in those <2 years). Testing, though not treatment, was significantly higher among inpatients with high-risk conditions (p-value<0.01). Empiric therapy was rare, occurring for 0.6% of ARI hospitalizations and 0.9% of tested admissions, and not associated with age, high-risk condition, or ICU admission. Among patients with laboratory-confirmed influenza, 65% were treated.
Conclusion: In our study, although not all hospitalizations were influenza-associated, testing for influenza was uncommon and empiric antiviral treatment was rare among inpatients with ARI during influenza seasons. While laboratory-confirmed influenza cases were often treated, reliance on antigen-based assays can miss true influenza cases due to false negative results. Exploration of the cost and benefits of test-and-treat strategies that include use of newer molecular assays, is warranted.
M. A. Rolfes,
J. I. Meek, None
A. M. Fry, None
S. S. Chaves, None