Advancing health equity through implementation science: Identifying and examining measures of the outer setting

Background: Implementation science (IS) could accelerate progress toward achieving health equity goals. However, the lack of attention to the outer setting where interventions are implemented limits applicability and generalizability of findings to different populations, settings, and time periods. We developed a data resource to assess outer setting across seven centers funded by the National Cancer Institute's IS Centers in Cancer Control (ISC³) Network Program. Objective: To describe the development of the Outer Setting Data Resource and characterize the county-level outer context across Centers. Methods: Our Data Resource captures seven key environments, including: (1) food; (2) physical; (3) economic; (4) social; (5) health care; (6) cancer behavioral and screening; and (7) cancer-related policy. Data were obtained from public sources including the US Census and American Community Survey. We present medians and interquartile ranges based on the distribution of all counties in the US, all ISC³ centers, and within each Center for twelve selected measures. Distributions of each factor are compared with the national estimate using single sample sign tests. Results: ISC³ centers’ catchment areas include 458 counties and over 126 million people across 28 states. The median percentage of population living within ½ mile of a park is higher in ISC³ counties (38.0%, interquartile range (IQR): 16.0%–59.0%) compared to nationally (18.0%, IQR: 7.0%–38.0%; p < 0.0001). The median percentage of households with no broadband access is significantly lower in ISC³ counties (28.4%, IQR: 21.4%–35.6%) compared the nation overall (32.8%, IQR: 25.8%–41.2%; p < 0.0001). The median unemployment rate was significantly higher in ISC³ counties (5.2%, IQR: 4.1%–6.4%) compared to nationally (4.9%, 3.6%–6.3%, p = 0.0006). Conclusions: Our results indicate that the outer setting varies across Centers and often differs from the national level. These findings demonstrate the importance of assessing the contextual environment in which interventions are implemented and suggest potential implications for intervention generalizability and scalability.