Abstract
Neighborhood conditions significantly impact health and well-being, leading public health agencies to focus interventions on disadvantaged areas. Yet, a growing number of composite indices have been developed to measure neighborhood opportunity, deprivation, or vulnerability, and little is known about how they compare in relation to health. Our study compares 13 composite indices of neighborhood conditions commonly used in public health research. Using data for over 57,000 US census tracts, we evaluate correlations among indices and assess their associations with four health outcomes-poor mental health, diabetes, smoking, and life expectancy-relative to the poverty rate, a standard single-variable measure of neighborhood deprivation. We used spatial regression models to account for autocorrelation and ordinary least squares regression to compare explanatory power across measures. While indices correlate with one another and with poverty, their associations with health outcomes and explanatory power vary considerably. Some indices exhibit stronger associations with health compared to poverty, suggesting they may offer more nuanced insights into neighborhood health inequities. Conversely, other indices explained less variation in health outcomes, questioning their utility in guiding interventions. Our findings underscore that indices are not interchangeable and highlight the need for careful selection of indices based on specific public health contexts and outcomes. Our study suggests simpler measures like poverty may, in some cases, be equally or more effective while offering a more interpretable and practical benchmark for targeting resources.