Associating nighttime lights with population health, with consideration of environmental and socio-economic confounders

Background: The relationship between nighttime lights and human health is being discussed in a growing number of studies, with a vast majority of results supporting an association. However, the impact of potential confounding has been insufficiently addressed in existing studies, undermining the reported associations. Objective: This study aimed to assess the relationship between exposure to nighttime lights and population health outcomes in the USA with a careful adjustment of potential confounders. Methods: This ecological study analyzed county-level data from 3104 counties in the USA. Nighttime light exposure data were derived from the DMSP-OLS Nighttime Lights Time Series. The outcomes were county-level health indicators including life expectancy at birth, age-specific mortality risk, and 21 cause-specific mortality rates. Covariates included county-level data on population demographics, socio-economics, healthcare, and environmental variables. Univariate regression and multivariate regression modeling were conducted to analyze the crude and confounding-adjusted association between the exposure and outcomes. Results: After adjusting for confounders, no significant associations were found between nighttime light exposure and the population health outcomes except mortality rates from mental and substance use disorders, and transport and unintentional injuries. Substantial changes in the direction and magnitude of the association estimates were identified before and after adjusting for the potential confounders, particularly the urbanization and socio-economic factors. Conclusions: This national study provides evidence that the relationship between nighttime light exposure and population health outcomes is complex and often confounded by environmental and socio-economic factors. Policy efforts should focus on mitigating unnecessary nighttime illumination to potentially reduce associated health risks.