Evaluation of Small-Area Estimation Methods for Use by the Behavioral Risk Factor Surveillance System

• Activity Number: 296
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): Haci Akcin*+ and Carol Gotway Crawford and Derek Ford and Martin Frankel and Michael Battaglia and Xingyou Zhang and Theodore J Thompson and Betsy Cadwell-Gunnels
• Companies: Centers for Disease Control and Prevention and Centers for Disease Control & Prevention and Centers for Disease Control and Northrop Grumman and Baruch College, CUNY and Battaglia Consulting Group and CDC and Centers for Disease Control and Prevention and Center for Disease Control & Prevention
• Keywords: small area ; BRFSS ; county estimates

Abstract

There is an ongoing need to assess health status at both the state and county levels. Estimates of health status help determine demand for health care and inform health policy decision making. The Behavioral Risk Factor Surveillance System (BRFSS) is one of the world's largest ongoing random digit-dialing telephone surveys designed to provide state-level estimates of health status. In 2002, BRFSS introduced The Selected Metropolitan/Micropolitan Area Risk Trends to provide direct estimates of certain health indicators for selected areas with 500 or more respondents. Roughly 300 counties have estimates of disease prevalence. However, the cost of obtaining reliable estimates for all counties is prohibitory large. Several approaches for modeling county level estimates from BRFSS have been proposed. We provide an overview of approaches and contrast the methods with respect to model assumptions, specification and validity, computational time, and benchmarking. An analytic comparison of a subset of estimates and standard errors is presented to evaluate accuracy against direct estimates. We summarize differences in estimates between approaches where direct estimates are not reliable.