OBJECTIVES: What are the factors that set the stage for health status and outcomes in the United States (U.S.)? This complex question is rarely considered in a comprehensive way. The current study employs an artificial intelligence analysis to assess the accuracy of adding a measure of social capital (i.e., public trust) to the Ecological Framework of Population Health in predicting U.S. county-level life expectancy and COVID-19 mortality rates.
STUDY DESIGN: Descriptive, cross-sectional, retrospective analysis.
METHODS: The current study utilized several U.S. county-level datasets representing the Ecological Framework of Population Health, including measures of culture, politics, policy, socioeconomics, lifestyle behaviors, and both chronic disease risk factors and diagnoses. A social media generated index of social capital, i.e., public trust, was added to the framework as a crosscutting variable to determine its efficacy in predicting county-level life expectancy and COVID-19 mortality rates using a non-linear artificial intelligence statistical method.
RESULTS: Analysis revealed significance in predicting both life expectancy (R(2) = 0.803) and COVID-19 deaths (R(2) = 0.548), with the optimal model employing 27 and 12 features, respectively. The public trust index was retained in both final models, ranking as the 5th and 6th most important predictors for life expectancy and COVID-19 mortality, respectively.
CONCLUSIONS: The present study expands the body of work exploring forcing factors of population health by demonstrating the potential utility of a measure of public trust, derived from social media posts and integrated into the Ecological Framework of Population Health, in predicting important U.S. county-level health outcomes.