Conducting a BRFS Among Minority Populations in Michigan

BACKGROUND:  A priority of the Michigan Department of Community Health (MDCH) Health Disparities Reduction and Minority Health Section (HDRMHS) is to improve the availability of health related data for racial and ethnic minorities, a priority shared at the national level. Within the Michigan adult population, 3.0% are Asian/Pacific Islander, 4.7% are Hispanic/Latino, and 4.9% (after adjusting for under-reporting) are Arab/Chaldean. Accurate and timely data among these groups at the population level are limited. The statewide Michigan Behavioral Risk Factor Survey (MiBRFS) obtains adequate health information from White and Black adults, but too few respondents of other racial and ethnic minorities are captured for robust analyses. Therefore, HDRMHS developed additional data collection methods to improve equity in health data availability. 

METHODS: A stand-alone BRFS was conducted among Asian and Hispanic adults in 2012 and among Arab/Chaldean adults in 2013. For the Hispanic and Asian stand-alone BRFS, survey researchers at Michigan State University (MSU) utilized surname databases developed by Survey Sampling International (SSI). The probability samples were based on Michigan landline phone directories whose surnames matched the SSI databases. MDCH, MSU, Wayne State University (WSU), and a community center collaborated to help develop and implement a BRFS among Arab/Chaldeans. MSU survey researchers drew a probability sample from Michigan landline phone directories using an Arab/Chaldean surname list developed for WSU cancer epidemiology studies. Telephone interviews were completed in English, Spanish, and Arabic. Data from participants in the MiBRFS who identified as Asian, Hispanic, or Arab/Chaldean were combined with data from the respective stand-alone surveys and weighted to produce three final data files. 

RESULTS:  There were a total of 510 Asian, 659 Hispanic, and 536 Arab/Chaldean participants in the 2012 and 2013 combined BRFSs. The Arab/Chaldean surname approach was very efficient for the landline samples, with yield rates for three geographic strata (High density: 27.9%, Medium: 6.2%, Low: 1.8%) closely matching the expected yields based on Census profiles tracts (High: ≥ 10%, Medium: ≥4-<10%, Low: <4%). Some Arab/Chaldean participants requested information with locating a medical provider and other services, requests not previously experienced with other BRFS projects.

CONCLUSIONS:  Conducting BRFSs among minority populations provided critical health related data not previously available by the MiBRFS alone. The Arab/Chaldean surname approach appeared effective and a similar approach could be used in other communities. The Arab/Chaldean BRFS identified a population with high need for healthcare and services. Data will be useful for identifying barriers and developing culturally appropriate programs.