Statistical Improvements in Weekly-Updated Cumulative Estimates of Flu Vaccination Coverage for Children in the United States

• Activity Number: 352 - Small Area Estimation, Analysis of Complex Sample Survey Data, and New Advances for Health Surveys
• Type: Contributed
• Date/Time: Thursday, August 12, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Survey Research Methods Section
• Author(s): Kennon Copeland and Nada Ganesh* and Lin Liu and Tammy Santibanez and James Singleton
• Companies: NORC at the University of Chicago and NORC at the University of Chicago and NORC at the University of Chicago and CDC and CDC
• Keywords: Cumulative vaccination coverage; Composite estimation; Variance reduction; Estimates revisions; Surveys and questionnaires

Abstract

The NIS-Flu is a national random-digit dialing cellular telephone survey of households, sponsored by the CDC, from which weekly estimates of cumulative influenza (flu) vaccination coverage for children 6 months through 17 years of age are generated by NORC at the University of Chicago. The survey has been conducted during each flu season (October-June) since the 2010-11 season.

The current NIS-Flu estimator uses a composite estimation technique, combining direct estimates from all prior weeks to generate weekly “enhanced” estimates with smaller variance than the direct estimate from a survey week. However, the current estimator is subject to weaknesses such as larger variance for initial estimates of vaccination coverage for the immediate prior month and instability of estimates for weeks within a month. Thus, research has been conducted on methods intended to improve accuracy and stability of the weekly cumulative estimates.

We will present descriptions of alternative estimators using conditional probabilities, Kaplan-Meier estimation, and smoothing with prior seasons’ data; performance evaluation results; and determination of the NIS-Flu estimator to be used for the 2021-22 season.