Improving accuracy of influenza-associated hospitalization rate estimates

Abstract

Diagnostic test sensitivity affects rate estimates for laboratory-confirmed influenza–associated hospitalizations. We used data from FluSurv-NET, a national population-based surveillance system for laboratory-confirmed influenza hospitalizations, to capture diagnostic test type by patient age and influenza season. We calculated observed rates by age group and adjusted rates by test sensitivity. Test sensitivity was lowest in adults ≥65 years of age. For all ages, reverse transcription PCR was the most sensitive test, and use increased from<10% during 2003–2008 to ≈70% during 2009–2013. Observed hospitalization rates per 100,000 persons varied by season: 7.3–50.5 for children<18 years of age, 3.0–30.3 for adults 18–64 years, and 13.6–181.8 for adults ≥65 years. After 2009, hospitalization rates adjusted by test sensitivity were ≈15% higher for children<18 years, ≈20% higher for adults 18–64 years, and ≈55% for adults ≥65 years of age. Test sensitivity adjustments improve the accuracy of hospitalization rate estimates.