2794. Testing and Treatment in Patients Hospitalized with Suspected Influenza Pneumonia

• Published in: Open forum infectious diseases Vol. 6; no. Supplement_2; p. S987
• Main Authors: Deshpande, Abhishek, Klompas, Michael, Bartley, Patricia, Yu, Pei-Chun, Haessler, Sarah, Zilberberg, Marya, Imrey, Peter, Rothberg, Michael
• Format: Journal Article
• Language: English
• Published: US Oxford University Press 23.10.2019
• Subjects: Abstracts
• ISSN: 2328-8957; 2328-8957
• DOI: 10.1093/ofid/ofz360.2471

Abstract Background Influenza is a leading cause of community-acquired pneumonia (CAP). Little is known about the effect of influenza testing on antimicrobial treatment among adult patients hospitalized with CAP. We quantified prevalence of testing and impact of positivity on treatment with antibacterials, antivirals, and outcomes. Methods We included adults admitted with pneumonia in 2010–2015 to 179 US hospitals contributing to the Premier database. Patients had CAP if radiographic evidence of pneumonia and antimicrobial treatment were present on day 1. We assessed influenza testing and compared antimicrobial utilization and outcomes of patients who tested positive vs negative vs not tested. Using mixed logistic regression and gamma generalized linear mixed models, we assessed the impact of influenza testing on inpatient mortality, length of stay (LOS) and cost. Results Among 166,273 patients with CAP, 38,665 (23.2%) were tested for influenza; 11.5% of these tested positive. The influenza testing rate increased from 15.4% in 2010/7–2011/6 to 35.6% in 2014/7–2015/6, ranging from 28.8% during flu season (October–May) to 8.2% in other months. Positive tests were more common during flu season (12.2% vs. 2.8%, P < 0.001). Patients tested for influenza were younger (66.6 vs. 70.3 years), less likely admitted from SNF (5.4% vs. 7.9%), with fewer comorbidities (2.9 vs. 3.3). Of patients tested for influenza, positive patients were younger (66.3 vs. 68.8 years), less likely admitted from SNF (5.2% vs. 6.8%), with more comorbidities (2.9 vs. 2.7) (all comparisons P < 0.001). Patients testing positive more likely received antivirals, were slightly less likely to receive antibacterials (Figure 1), but received shorter antibacterial courses than negative patients (5.3 vs 6.4 days, P < 0.001). Influenza tests were associated with reduced odds of in-hospital mortality compared with no testing (adjusted OR 0.71, 95% CI 0.63–0.81) and positive vs. negative tests with reduced costs (0.95, 0.92–0.99) and LOS (0.97, 0.94–0.99) (Figure 2). Conclusion In a large US inpatient sample hospitalized for pneumonia, only 23.2% of the patients were tested for influenza, but testing varied widely by hospital. A positive influenza test was associated with antiviral treatment but had minimal impact on antibiotic prescribing. Disclosures All authors: No reported disclosures.