Race, Ethnicity, and Linguistic Isolation as Determinants of Participation in Public Health Surveillance Surveys

INTRODUCTION: To plan, implement, and evaluate programs designed to improve health conditions among racial and ethnic minority populations in the United States, public health officials and researchers require valid and reliable health surveillance data. Monitoring chronic disease and behavioral risk factors among such populations, however, is challenging. This study assesses the effects of race, ethnicity, and linguistic isolation on rates of participation in the Behavioral Risk Factor Surveillance System (BRFSS). METHODS: County-level data from the 2003 BRFSS survey and 2000 U.S. census were used to examine the effects of race, ethnicity, and linguistic isolation on six measures of survey participation (i.e., rates of resolution, screening, cooperation, response, language barriers, and refusal). RESULTS: Participation rates were significantly lower in counties with higher percentages of black people and people who did not speak English. Response rates decreased by 4.6% in counties with the highest concentration of black residents compared with counties with few black residents. Likewise, response rates decreased by approximately 7% in counties in which a larger percentage of the population spoke only Spanish or another Indo-European language compared with counties in which all residents spoke English. CONCLUSION: The negative relationship between the percentage of Spanish-only–speaking households and participation rates is troubling given that the BRFSS is conducted in both Spanish and English. The findings also indicate that more needs to be done to improve participation among other minorities. Researchers are investigating several ways of addressing disparities in participation rates, such as using postsurvey adjustments, developing more culturally appropriate data-collection procedures, and offering surveys in multiple languages.