BACKGROUND: Data from population-based surveys play a critical role in informing programs, interventions, and policies to improve public health and health equity. However, white non-Hispanic and higher-income individuals are often over represented in these surveys, leading to gaps in knowledge about the health and opportunities to improve health of other critical populations. The Best Starts for Kids Health Survey (BSKHS) is a population-based survey about the health and well-being of families and children 5th grade and younger in King County, WA. With the goal of achieving a truly representative sample, BSKHS took unique equity-based approaches throughout the survey. This presentation will describe the success of these approaches and will discuss implications and lessons learned for future population-based surveys.
METHODS: The equity-based approaches in BSKHS included oversampling by race/ethnicity and geography; offering the survey, announcement letters, and other materials in six languages (Chinese, English, Russian, Somali, Spanish, and Vietnamese); conducting community outreach prior to and during data collection; conducting text message-based recruitment in addition to phone call recruitment; preferentially calling predicted families of color or families speaking a language other than English; and including a community-based convenience sample to augment the sample size of small populations (e.g., families in specific racial/ethnic subgroups, who are LGBTQ, or who are unstably housed). For further discussion of BSKHS methods, see Wong et al., (CSTE abstract submitted).
RESULTS: Data collection was 80% complete at the time this abstract was submitted. To date, children of color represent 55% of BSKHS participants. In 2015, 52% of children 0 to 9 years in King County were children of color. To provide a comparison, 16% of the 2015 Behavioral Risk Factor Surveillance System respondents in King County were people of color, less than half of the expected 37% needed to be representative of King County overall. Eleven percent (11%) of surveys have been taken in a language other than English. In King County in 2015, 10% of the population 5 years and older spoke a language other than English and spoke English less than very well. Qualitative findings also suggest that non-English speaking families participated because interviewers were not only bilingual but also bicultural.
CONCLUSIONS: Preliminary data suggest that the equity-based approaches of BSKHS attained a more representative sample than did previous population-based surveys in King County.