Behavioral Risk Factor Surveillance System (BRFSS) Dataset Details

    To contact the board with suggestions or comments, you may send an email to the following address: cohid@state.co.us

Overview

The Colorado Behavioral Risk Factor Surveillance System (BRFSS) is an ongoing statewide telephone survey designed to monitor the prevalence of health behaviors and preventive health practices associated with the leading causes of disease, disability, and premature death. The Colorado BRFSS was initiated in 1990 as a joint project of the Colorado Department of Public Health and Environment and the Centers for Disease Control and Prevention (CDC). The Survey Research Unit at the Colorado Department of Public Health and Environment conducts the survey and selects respondents using a random digit dialing sampling technique. The sample size of the BRFSS has grown over time. While scientific survey procedures are followed in all phases of the survey, the data should be regarded as estimates. Colorado residents without telephones and those who cannot be interviewed in either English or Spanish are excluded from the study. BRFSS measures are self-reported and are subject to several sources of error. For example, prevalence estimates only capture those with diagnosed conditions. In addition, recall bias -- trouble remembering information such as when blood pressure was last checked -- may distort estimates.

For more information, please go to the Colorado BRFSS website (http://www.cdphe.state.co.us/hs/brfss/)or the CDC website (www.cdc.gov/brfss).

Health Statistics Regions

These 21 Health Statistics Regions (HSRs) are aggregations of counties developed by the Health Statistics Section of the Colorado Department of Public Health and Environment (CDPHE) in partnership with state and local public health professionals. The HSRs were developed using statistical and demographic criteria. More information on the 21 HSRs is available by contacting Alyson Shupe, Health Statistics Section, CDPHE at health.statistics@state.co.us.

Top of Page

Small numbers

Due to the relatively low number of surveys completed each year, two years of data are combined to produce more stable estimates. Since not all questions are asked each year, some estimates are based on only one year of data. Estimates are not provided for counties or PMRs with fewer than 50 respondents.

Weighting

Population-based estimates are produced by weighting BRFSS data for the probability of selection and poststratifying to reflect the age and sex distribution of the selected area. Probability of selection can vary by telephone number, geographic region, the number of adults in the household, and the number of telephones in the household. For example, weights for the number of adults in the household adjust for the lower probability of selection for respondents from larger households compared to respondents from smaller households. That is, a person in a one-adult household is twice as likely as a person in a two-adult household to be selected for interview.

Confidence Intervals

The prevalence estimates reported on CoHID estimate the percent of the Colorado adult population who exhibit certain risk behaviors (e.g., 23 percent of Colorado adults are current smokers). Since the estimates are based on a sample and not the entire population, there is some error. Confidence intervals take this error into account and present a range in which the “true value” is likely to fall (e.g. between 21.1 and 24.3 percent of Colorado adults are current smokers).

Interval width is influenced both by the degree of certainty sought (e.g., 95 percent versus 99 percent certainty) and the standard error. A high degree of certainty (e.g. 99 percent) increases interval width, as does a small sample size. Wide confidence intervals are associated with less reliable estimates and should be interpreted with caution.

Top of Page

Back to CoHID home page