Behavioral Risk Factor Surveillance System (BRFSS) Sampling Method
The BRFSS collects data from adults aged 18 years or older. The Centers for Disease Control and Prevention (CDC) coordinate the survey, but it is conducted separately by each state. Each state conducts thousands of interviews every year. In Washington State, over 1,000 interviews are completed monthly.
Most BRFSS data are gathered through telephone interviews (see below). However, changes in culture and technology make gathering data this way increasingly difficult, so we are gathering data through paper surveys and over the Internet to see how they compare to data gathered by telephone. Most potential participants will be contacted by phone for an interview, but others may receive a letter directing them to a web site, or a package including a paper version of the survey and a stamped, self-addressed return envelope.
Most BRFSS data in Washington State are collected through telephone interviews. Households are selected from blocks of potential phone numbers in an area, including unlisted numbers. Separate lists of landline and cell phone numbers are used. Unlisted telephone numbers are included in these lists. Calls are made seven days a week during both the daytime and the evening. Once a household is contacted, one adult is randomly selected to be interviewed. This is usually the adult with the next birthday. Washington State contracts with a company called ICF International to conduct the interviews.
Like most states, Washington uses a computer-assisted telephone interviewing (CATI) software program. When a CATI program is used, a questionnaire is displayed on a computer screen during each interview, and the interviewer enters the responses directly into a computer. A CATI program offers several advantages:
- Data entry errors are minimized.
- Questions not applicable to the current respondent (for example, age- or sex-specific questions) are automatically skipped.
- Responses can be immediately checked, and those that are found unacceptable (such as impossible body weight) are immediately brought to the interviewer's attention so that they can be corrected.
BRFSS interviews are “scripted” to obtain consistent results. The interviewers are trained to read the questions in specific ways. However, this can sometimes make the flow of the interview sound unnatural to the person being interviewed.
The BRFSS relies on information reported directly by the respondent, so it may be subject to a number of sources of possible error. How questions are worded may elicit responses in a certain way and can result in what is called "measurement error." Similarly, the ability to accurately recall details varies by person and how much time has passed since the event they are trying to recall, which leads to "response error." It is also possible that the people who choose to participate are different than those who do not. Interviews are conducted only in English and Spanish in Washington State, so adults who are not able to be interviewed in English or Spanish are not included. Households without telephones are not contacted. Thus, BRFSS findings can only be generalized to English and Spanish speaking adults living in households with telephones. This is called “selection bias.”
There is no reason to believe that these sources of bias change significantly from year to year, so even if the results are not completely accurate they can be compared over time. This allows us to determine if the prevalence of a given condition is increasing or decreasing, which is usually what we want to know.
Results from the BRFSS
The BRFSS data has been used to track changes in behavior and to measure progress toward achieving national, state, and local public health objectives. For example, Healthy People 2020 is a set of goals for improving the health of the nation by reducing risk factors for diseases and injury and improving the use of health services. Many of these risk factors are measured by the BRFSS, and progress toward meeting many of the Healthy People 2020 objectives is measured using national BRFSS datasets that include data from Washington State.
State and local health assessment reports rely extensively on BRFSS data to examine the behaviors that underline health risks and access to care for the population of Washington. For example, the Health of Washington State is the Department of Health's periodic review of Washington's well-being. It asks three general questions: How healthy is Washington State as a place to live? How healthy are we as a community of people sharing that place? And, how does health vary between distinct populations living in Washington? It relies on a substantial number of BRFSS measures to answer these questions. This information is located in the Health of Washington State report.