Using multi-sourced big data to correlate sleep deprivation and road traffic noise: A US county-level ecological study

Background: Road traffic noise is a serious public health problem globally as it has adverse psychosocial and physiologic effects (i.e., sleep). Since previous studies mainly focused on individual levels, we aim to examine associations between road traffic noise and sleep deprivation on a large scale; namely, the US at county level. Methods: Information from a large-scale sleep survey and national traffic noise map, both obtained from Government's open data, were utilized and processed with GIS techniques. To examine the associations between traffic noise and sleep deprivation, we used a hierarchical Bayesian spatial modelling framework to simultaneously adjust for multiple socioeconomic factors while accounting for spatial correlation. Findings: With 62.90% of people not getting enough sleep, a 10 dBA increase in average sound-pressure level (SPL) or SPL of the relatively noisy area in a county, was associated with a 49% (OR: 1.49; 95% CrIs:1.19–1.86) or 8% (1.08; 1.00–1.16) increase in the odds of a person in a particular county not getting enough sleep. A 10% increase in noise exposure area or population ratio was associated with a 3% (1.03; 1.01–1.06) or 4% (1.04; 1.02–1.06) increase in the odds of a person within a county not getting enough sleep. Interpretation: Traffic noise can contribute to variations in sleep deprivation among counties. This study suggests that policymakers could set up different noise-management strategies for relatively quiet and noisy areas (i.e., different limiting SPLs) and incorporate geo-spatial noise indicators, such as exposure population or area ratio. Furthermore, urban planners should consider urban sprawl patterns differently.