|
Activity Number:
|
315
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #302144 |
|
Title:
|
Confidence Intervals for Estimates of Excess Mortality
|
|
Author(s):
|
Al Ozonoff*+ and Xiaopeng Miao and Po-Yung Cheng and William W. Thompson
|
|
Companies:
|
Boston University School of Public Health and Boston University School of Public Health and Centers for Disease Control and Prevention and Centers for Disease Control and Prevention
|
|
Address:
|
715 Albany Street, Crosstown Center, Boston, MA, 02118,
|
|
Keywords:
|
Influenza ; Surveillance ; Excess mortality ; Confidence intervals
|
|
Abstract:
|
The 122 Cities Mortality Reporting System (CMRS) is one component of the national influenza surveillance program operated by the Centers for Disease Control and Prevention (CDC). Participating cities submit weekly counts of pneumonia and influenza (P&I) deaths and all-cause (AC) deaths. CDC models the square root of the arcsine of the ratio of P&I deaths to AC deaths using an Andrews robust regression method. The model estimates both a seasonal baseline and threshold value for determining influenza-associated mortality. Prior work has demonstrated the benefit of robust regression methods to reduce estimation bias. In this work we investigate variability, and consider the statistical issues involved in constructing confidence intervals (CIs) for estimates of excess mortality. Using historical data from the 122 CMRS, we compare model-based CIs to empirical CIs from a resampling procedure.
|
