Influenza Aberration Detection Models for Eight Large Cities in the United States

• Activity Number: 392
• Type: Invited
• Date/Time: Thursday, August 12, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Section on Government Statistics
• Author(s): William W. Thompson*+ and Eric Weintraub and David Shay and Lynnette Brammer and Nancy Cox and Keiji Fukuda
• Companies: Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention
• Keywords: aberration detection ; Poisson regression

Abstract

We applied Poisson regression models to data obtained from several cities in the U.S. to identify daily aberrations in deaths that were associated with influenza. From the 1990 through 2000, we obtained mortality data from the National Center for Health Statistics, influenza data from the World Health Organization, and temperature data from the National Oceanic and Atmosphere Association. Poisson regression models were used to identify influenza-associated aberrations in deaths by city. After controlling for temperature, there was still significant variation in influenza parameter estimates between cities. Influenza was associated with 4-14% of pneumonia and influenza deaths, 3-5% of respiratory and circulatory deaths, and 1-3% of all-cause deaths. In addition to temperature, other city-level confounders were examined to assess the sensitivity and specificity of influenza aberration detection models by city. Future models will also assess interactions between climate factors and influenza.