Examining County-Wide Impacts of Early Childhood Obesity-Reduction Efforts in LA County

• In Event: 2-144 - Reach, Implementation, and Impacts of Efforts to Reduce Early Childhood Obesity In LA County

Integrative Statement

Given the negative consequences early childhood obesity can have for healthy child development (Reilly & Kelly, 2010), there has been an increase in national and regional efforts to combat child obesity. While research has examined the effectiveness of programs to reduce childhood obesity, less is known about impacts of collective efforts to reduce early childhood obesity at county levels. In LA County, a leading public grant-making organization funded a broad set of activities (e.g., park development, food demonstrations, breastfeeding promotion) to combat obesity under the Reducing Early Childhood Obesity (RECO) initiative. This study seeks to answer the question: What is the collective impact of RECO on child obesity rates in LA County?

Determining the impact of a broad portfolio of interventions on a single large area like LA County presents unique challenges. Although it is possible to assess county-wide changes in child outcomes, it is difficult to ascertain whether any changes resulted from RECO or from something else in LA County. In addition, LA County is the most populous and one of the most racially/ethnically diverse counties in the US, and it is geographically larger than 4,000 square miles, making it difficult to identify an appropriate comparison county.

To address these challenges, this study employs a synthetic comparison group design (Abadie 2010). The synthetic comparison group approach provides a transparent analytic method for constructing a counterfactual (i.e., comparison group) against which to estimate effects on large aggregate units such as counties by comparing the treatment county to “donor counties” or other counties that could be used as counterfactuals. The method then weights donor counties in order to create a synthetic county with similar observable characteristics to the treatment county.

For this study, we used two outcome data sources – overweight/obesity data for children aged 2-5 from Southern California Kaiser Permanente and data on other child health behaviors from the California Health Information Survey (CHIS). Additional county-level demographic data from the American Community Survey was also used to support synthetic comparison group construction. Kaiser data were available bi-annually from 2009 through 2016 for seven counties; CHIS data were available annually or semi-annually from 2007 through 2016 for 34 counties.

Examining child obesity trends in LA County over time reveals declines since the implementation of RECO activities began (Figure 1). To determine whether this favorable decline could be attributed to RECO, we then compared time trends in synthetic LA County to actual LA County in the period after RECO activities began. Figure 2 shows that while child obesity rates declined in both LA County and the synthetic county, obesity in LA County decreased to a greater degree, suggesting a favorable county-level impact of RECO. Additional analyses will include falsification tests to assess the likelihood that the divergence in trends is attributable to random chance. We will also examine county-level impacts on intermediate child outcomes such as healthy eating, physical activity, and screen time.

Discussion will focus on the utility of collective efforts to improve child health behavior and outcomes at the county-level.

Authors

• Lauren Olsho, Abt Associates Inc.
  Presenting Author

• Molly Brune, Abt Associates
  Non-Presenting Author

• Michelle Blocklin, Abt Associates
  Non-Presenting Author