Outdoor Artificial Light at Night and Insomnia-Related Social Media Posts

Abstract

Importance: Understanding the outcomes of artificial light at night (ALAN) on insomnia is crucial for public health, particularly in rapidly urbanizing regions. However, evidence of the association between ALAN exposure and insomnia is limited, despite the large number of people exposed to ALAN.

Objective: To explore the association between outdoor ALAN exposure and insomnia among the Chinese population.

Design, setting, and participants: This cross-sectional study used social media data from Weibo (Sina), a social media platform, and satellite-derived nighttime light images. The study period spans from May 2022 to April 2023. The study encompasses 336 cities across China's mainland, providing a comprehensive national perspective. Data include insomnia-related posts from the platform users, representing a large and diverse population sample exposed to varying levels of ALAN.

Exposure: Outdoor ALAN exposure (in nanowatts per centimeters squared per steradian [nW/cm2/sr]) was measured using satellite-derived nighttime light images at a spatial resolution of 500 m.

Main outcomes and measures: The incidence of insomnia among residents at the city level was measured by the number of insomnia-related posts on social media. Multiple linear regression models were used to estimate the association between ALAN exposure and population insomnia, adjusting for population characteristics and meteorological factors at the city level.

Results: The study included data from 1 147 583 insomnia-related posts. Daily mean ALAN exposure across the 336 cities ranged from 3.1 to 221.0 nW/cm2/sr. For each 5 nW/cm2/sr increase in ALAN exposure, the incidence of insomnia increased by 0.377% (95% CI, 0.372%-0.382%). The association was greater in less populated cities and under extreme temperature and poor air quality conditions. The observed exposure-response functions between ALAN exposure and insomnia demonstrated an upward trend, with steeper slopes observed at low exposures and leveling off at higher exposures.

Conclusions and relevance: This study provides evidence of the association between increased ALAN exposure and higher incidence of insomnia. These findings expand the current knowledge on adverse health outcomes of ALAN exposure and emphasize the potential health benefits of well-planned artificial nighttime lighting in China and other developing countries in the early stages of city planning.

Figure 1.. Daily Mean Spatial Distribution of Artificial Light at Night (ALAN) Exposure in China at 500-m Resolution, May 2022 to April 2023

ALAN was measured as nanowatts per centimeters squared per steradian (nW/cm²/sr). Inset shows islands in the South China Sea.

Figure 2.. Descriptive Statistics and Pearson Correlation Coefficient of Artificial Light at Night (ALAN) and Insomnia-Related Social Media Posts

A, Average ALAN exposure as nanowatts per centimeters squared per steradian (nW/cm²/sr) within all samples and different subgroups. B, Average daily insomnia-related social media posts in all samples and different subgroups. C, The total cumulative number of observation days in all samples and different subgroups. D, The Pearson correlation between ALAN exposure and insomnia-related social media estimations in all samples and different subgroups.

Figure 3.. Exposure-Response Curve of Associations of Incidence of Insomnia With Artificial Light at Night Exposure

To estimate the exposure-response curves between artificial light at night exposure as nanowatts per centimeters squared per steradian (nW/cm²/sr) and the incidence of insomnia, exposure was fit as a smoothing term in a multivariable adjusted model using a spline with 3 nodes. Shaded area denotes 95% CI. The exposure of daily artifical light at night at the city level ranged from 0 to 396.55 nW/cm²/sr among all days in the study period.

Figure 4.. Subgroup Analyses of the Association Between Insomnia Incidence and Each 5 nW/cm²/sr Increase in Artificial Light at Night Exposure

A, Results of subgroup analysis by different city categories, divided into large cities (population >5 million), medium cities (population ≥3 million to ≤5 million), and small cities (population <3 million). B, Results of subgroup analysis by different time periods, where weekends and holidays were determined on the basis of the notices issued by the General Office of the State Council regarding holiday arrangements. C, Results of subgroup analysis by seasons, divided into spring (March to May), summer (June to August), autumn (September to November), and winter (December to February of the following year). D, Results of subgroup analysis by different temperature conditions and air quality. Temperature stratification was based on the specific temperature distribution for each city, with the range between the 25th and 75th percentiles defined as comfortable temperature, whereas the remainder was categorized as uncomfortable temperature. Days were classified as polluted when the air quality index (AQI) exceeded 101 and as nonpolluted when it was 100 or below.