Beyond residential exposure: Mobility-oriented outdoor artificial light at night and its lagged associations with sleep

Residence-based assessments of outdoor artificial light at night (ALAN) risk contextual errors, potentially obscuring its relationships with health. Advances in fine-grained nighttime light remote sensing and mobility-oriented measurements enable more accurate individual-level outdoor ALAN exposure assessment. Leveraging a mobility-oriented framework to decouple residential from non-residential ALAN exposure and test their different associations with sleep efficiency, we analyzed 502 Hong Kong participants over 1286 nights through a cross-sectional survey. Sleep efficiency was measured using actigraphy and sleep diaries, while GPS-equipped smartphones tracked individual activity-travel trajectories. We implemented three spatiotemporally weighted outdoor ALAN exposure measurements: residential (R-ALAN), non-residential (NR-ALAN), and complete mobility-oriented (M-ALAN) outdoor ALAN exposure. Linear mixed-effects models estimated their associations with sleep efficiency, adjusting for sociodemographic, lifestyle, health status, and physical activity covariates. Results revealed that only NR-ALAN exposure was robustly associated with reduced sleep efficiency (β = -0.03, 95% CI: [-0.04, -0.01], p = 0.004), while R-ALAN exposure showed no statistically significant association. M-ALAN exposure showed statistically significant but less robust association (β = -0.03, 95% CI: [-0.05, -0.00], p = 0.032), suggesting that inclusion of R-ALAN exposure introduced contextual error that weakened model precision. Our findings indicate that outdoor ALAN encountered in non-residential activity spaces, rather than at residences, is the primary exposure pathway affecting sleep efficiency. This highlights critical contextual errors in traditional static assessments, hints at the potential lagged associations between outdoor ALAN exposure and sleep efficiency, and underscores the necessity of integrating mobility data to accurately capture environmental burdens, which provides a validated framework for future epidemiological research.