Estimates of US Influenza-Associated Deaths Calculated from Three Poisson Regression Models for the 1976/77 Through the 2004/05 Seasons

• Activity Number: 34
• Type: Contributed
• Date/Time: Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Hong Zhou*+ and William W. Thompson and Po-Yung Cheng and Cecile Viboud and Corinne Ringholz and Eric Weintraub and Joseph Bresee and David K. Shay
• Companies: CDC and CDC and CDC and Fogarty International Center and Fogarty International Center and CDC and CDC and CDC
• Keywords: influenza ; respiratory and circulatory ; mortality ; modeling

Abstract

We compared estimates of influenza-associated deaths calculated from three Poisson regression models that incorporated influenza virus surveillance data. Weekly underlying respiratory and circulatory deaths were modeled for the 1976/77 through 2004/05 respiratory seasons by using viral surveillance data directly, using deviations in viral surveillance data above a seasonal spline trend, or using deviations in viral data above a Serfling baseline. Annual average deaths for these three models were 27,469, 16,220 and 21,828, respectively. Estimates from each model were highly correlated (R>=0.94). Estimates from the models using deviations in viral surveillance data above a spline trend and using deviations in viral data above a Serfling baseline were more conservative and more consistent with expected patterns for more severe (i.e., influenza A (H3N2) predominant) influenza seasons.