Gathering deep insights into communities from advanced data science methods and geospatial analytics
Public health officials at the federal and state levels have called for a better way to measure, predict, and adjust for social factors in health care and population health. The RTI Rarity project takes an “artificially intelligent” approach to inform decisions concerning community-level social, behavioral, environmental, and economic factors for quality health care. By curating a national database of more than 200 area-level social determinants of health (SDoH) measures within ten domains at the Census tract, ZIP code, and county levels across the U.S., the RTI Rarity tool provides high-resolution insights into factors that strongly influence health outcomes.
The RTI Rarity tool uses supervised machine learning, including random forests and other state-of-the-art predictive methods, to create local social inequity (LSI) scores drawing on the SDoH measures. The health equity analysis tool and its underlying data allow for the development of both within-state and cross-state summary scores and ten domain-specific sub scores informed by our conceptual framework. The scores yield meaningful insights into the neighborhood-level factors driving local health outcomes.