Associations between COVID-19 risk, multiple environmental exposures, and housing conditions: A study using individual-level GPS-based real-time sensing data

Few studies have used individual-level data to explore the association between COVID-19 risk with multiple environmental exposures and housing conditions. Using individual-level data collected with GPS-tracking smartphones, mobile air-pollutant and noise sensors, an activity-travel diary, and a questionnaire from two typical neighborhoods in a dense and well-developed city (i.e., Hong Kong), this study seeks to examine 1) the associations between multiple environmental exposures (i.e., different types of greenspace, PM(2.5), and noise) and housing conditions (i.e., housing types, ownership, and overcrowding) with individuals’ COVID-19 risk both in residential neighborhoods and along daily mobility trajectories; 2) which social groups are disadvantaged in COVID-19 risk through the perspective of the neighborhood effect averaging problem (NEAP). Using separate multiple linear regression and logistical regression models, we found a significant negative association between COVID-19 risk with greenspace (i.e., NDVI) both in residential areas and along people's daily mobility trajectories. Meanwhile, we also found that high open space and recreational land exposure and poor housing conditions were positively associated with COVID-19 risk in high-risk neighborhoods, and noise exposure was positively associated with COVID-19 risk in low-risk neighborhoods. Further, people with work places in high-risk areas and poor housing conditions were disadvantaged in COVID-19 risk.