JMIR Publications

ABSTRACT

Background:

Exposure risk was shown to have impacted individual susceptibility and epidemic spread of COVID-19 at population level. However, its dynamics across exposure settings and along adjustments in social distancing intervention were understudied.

Objective:

Utilising a participatory surveillance platform, this study aimed to examine the population’s trajectory of exposure risk in different settings and its association with SARS-CoV-2 infection over three consecutive Omicron waves in Hong Kong.

Methods:

In March – June 2022, individuals from randomly selected households were recruited for enrolment in a prospective population cohort. Through web-based monthly surveys, participants were asked to report any incident SARS-CoV-2 infection, COVID-19 vaccination uptake, the mode of work and activity frequency in daily and social settings in the previous month. Reflecting their exposure risk, the trend of the reported behaviours was differentiated into setting-specific trajectories using latent class growth analysis. Their associations with SARS-CoV-2 infection in overall and each Omicron wave (Wave I: February-April; Wave II: May-September; Wave III: October-December 2022) were evaluated using multivariable cox proportional-hazards model and Kaplan-Meier analysis.

Results:

Between March 2022 and January 2023, totally 33,501 monthly responses were collected from 5,321 individuals with 42% being male and a median age of 46 years (interquartile range 34-57 years). Against an expanding vaccination coverage from 82% to 98% for two doses and 20% to 78% for three doses from February to December 2022, the cumulative SARS-CoV-2 burden escalated from <0.2% to 25%, 32% and 44% towards the end of Wave I, II and III. Throughout, some 52% of the participants had worked regularly at workplace, while 29% and 19%, respectively, demonstrated a remote and assorted work pattern. For daily and social setting, four (Class 1-4) and five trajectories (Class 1-5) were identified correspondingly to delineate the varying dynamics of exposure risk. Individuals working regularly on-site (adjusted hazard ratio (aHR) 1.47), living in a larger household (aHR 1.12) were associated with a higher risk of SARS-CoV-2 infection in overall and contributed to more cases in Wave I. Having a greater exposure risk in daily (Class-4: aHR 1.46) and social setting (Class-4: aHR 1.52) also led to an increased risk of infection throughout and, respectively, in Wave II and III.

Conclusions:

In an infection-naïve population, SARS-CoV-2 transmission had predominantly initiated at workplace, inflated in household, and perpetuated in daily and social environments, as social distancing restrictions transitioned from a stringent to near-normal level. These patterns highlighted the phasic shift of exposure setting which could help calibrate an effective mix of social distancing interventions for future epidemic control instead of a complete lockdown.