APPENDIX A

BEHAVIORAL RISK FACTOR SURVEILLANCE SYSTEM

DESCRIPTION AND METHODOLOGY

The Behavioral Risk Factor Surveillance System (BRFSS) was established by the U.S. Centers for Disease Control and Prevention (CDC) in Atlanta in order to allow states to determine the prevalence of selected health risk factors among their adult populations. The West Virginia Bureau for Public Health was one of the 15 initial participants in the BRFSS in 1984. By 1993, there were 50 participating states, plus the District of Columbia. The BRFSS allows states to monitor trends in health behaviors, as well as compare data with other states. The survey technique used, i.e., interviewing a random sample of state residents by telephone, is a faster and more cost-effective way of obtaining this information than in-person interviews.

The BRFSS survey represents a collaborative effort between the Bureau and the CDC. The Bureau provides telephones, office space, interviewers, and supervision of the data collection; financial assistance, a standardized set of core questions, computer-assisted telephone interviewing software, computers, data processing services, and analytic consultation are provided by CDC. Additional questions of topical interest may be added by individual states. West Virginia's survey will include additional questions on osteoporosis risk factors beginning in 1998.

The survey sample is selected by random digit dialing (RDD) rather than through telephone directories since the latter do not include new or unlisted numbers. Sampling is conducted in a multistage cluster design based on the Waksberg Sampling Method for RDD, which eliminates many unassigned and business phone numbers from the selection process.

Based on a list of all state area prefixes provided to CDC, 10-digit numbers are randomly computer-generated and returned to the state. The numbers are dialed to identify the ones that belong to households. By replacing the last two digits with all of their possible combinations, a bank of 100 telephone numbers is generated from each number found to be residential. Calls are then made until three interviews (called a cluster of interviews) are completed from each bank. Completion of 67 clusters of three interviews per cluster results in the monthly sample size of 201 interviews, yielding an annual sample size of approximately 2,400 interviews. Within each household, the actual respondent is chosen randomly to avoid possible biases related to the time of day and household telephone answering preferences. The data must then be weighted to compensate for households with differing numbers of adults who, therefore, have different probabilities of being selected. Survey responses are weighted to the 1990 census age and sex distribution to more accurately estimate the actual prevalence of behavioral risk factors among the state's population.

For additional information on the BRFSS, contact Fred King, BRFSS Coordinator, Office of Epidemiology and Health Promotion, 350 Capital Street, Room 165, Charleston, WV 25301-3701, (304) 558-9100/FAX (304) 558-1553.

Back to Reference