National Estimates of Risks/Behaviors: Behavioral Risk Factor Surveillance System and the National Health Interview Survey

The National Health Interview Survey (NHIS) has been the pre-eminent survey for obtaining national health data on the U.S. population since the 1950s. Data collection is performed by the U.S. Census Bureau and is centralized. In-person interviews within households are conducted using a nationally representative sample of respondents. Data are collected from adults and children, and both self-reports and proxy reports are included. In contrast, the Behavioral Risk Factor Surveillance System (BRFSS) began in the early 1980s and expanded to all states in the early 1990s. The BRFSS was designed to collect statelevel risk factor data on adults 18 years of age or older. Data collection is decentralized, with each state collecting data through a contractor or though the state health department. Data are obtained through telephone surveys of representative samples within each state. Only selfreported data are obtained from respondents. The two surveys use similar questions for several measures. Because the BRFSS is now nationwide, it is potentially a source for national estimates, but the comparability of the NHIS and national estimates from the BRFSS is unclear. This presentation examines the comparability of BRFSS national estimates with NHIS estimates for health care coverage, current smoking status, body mass index, hypertension, diabetes, influenza vaccination and pneumonia vaccination among several populations. The 1997 data files were used since this was the most recent year available for the NHIS at the time the study was initiated. Prior research suggests that it is appropriate to combine state data (Iachan, et al, 1999). Our examination of selected sample design issues and quality indicators also lent support to combining state data to produce national estimates. The BRFSS sample size is not proportional to state populations, but range in size from 1,505 to 4,923, with a total of 135,582 respondents in 1997. Each state’s data are weighted to the state population (CDC, 2001). The national estimates for BRFSS were based on summing sample records for all states using the final post-stratified weight.