BACKGROUND: Best Starts for Kids (BSK) is a new initiative to improve the health and wellbeing of King County, WA by investing in prevention and early intervention for children, youth, families, and communities. However, very little population-based local area data exist to assess needs or evaluate impacts on young children. Therefore, we designed and deployed the Best Starts for Kids Health Survey (BSKHS) to assess the health and wellbeing of families and children 5th grade and younger in King County.
METHODS: We developed novel sampling frame methodologies for this survey, utilizing an equity and social justice-based approach while constrained in budget. Participants were randomly sampled using contact information from two frames. For children ages 0-5 years, we used King County resident births to form the sampling frame; for children in kindergarten through 5th grade, we worked with 18 school districts. Oversampling was conducted by race/ethnicity and across the four large geographic regions of King County. In addition to the population-based sample, we included a convenience sample component to supplement information about small populations such as children with child welfare involvement. For further discussion of the outreach and equity-based approaches, see Moore et al. (CSTE abstract submitted). Surveys were administered September 2016-January 2017. Online/telephone/mail modes were used and the survey was available in 6 languages in the online/phone modes.
RESULTS: This methodology required considerable internal staff time for frame development, including contacting individual school districts and development of district-level data sharing agreements. For the 0-5 year old frame, we are able to include only children who were County residents at the time of birth, but we estimate that this excluded only 7% of children. For the elementary school frame, key limitations are inclusion of only children enrolled in public schools (81% of children in the age range) and exclusion of children due to confidentiality reasons (varied by district, range 0%-45% excluded). Although the field period is only 80% complete at the time of abstract submission, 3% of cases are ineligible (mainly due to moving out of the County) and 10% were unable to be contacted due to no or incorrect phone number.
CONCLUSIONS: This novel sampling frame for a population-based survey of young children efficiently identified households with children and parent/guardian contact information for survey conduct, as compared with random digit dial methods or other nonspecific population-based sampling methodologies.