# Variability in the Population Diffusion Patterns of SARS‐CoV‐2 by Exposure Setting and Its Roles in Driving Epidemic Dynamics

## TL;DR

This study examines how SARS-CoV-2 spreads through different settings in Hong Kong and how these patterns affect epidemic growth.

## Contribution

The study identifies specific exposure settings that drive epidemic spread and quantifies their impact on transmission dynamics.

## Key findings

- Neighbourhood outbreaks were the largest and fastest, with an average size of 11.9 and daily generation of 1.18 cases per cluster.
- Social activity clusters had the highest spillover rate, breeding 3.73 onward clusters compared to 0.18 for residential clusters.
- Clusters in neighbourhood and social activity settings had the greatest impact on raising the effective reproduction number (Rt).

## Abstract

Identifying transmission events that trigger epidemic spread is paramount for informing outbreak control. This study characterised the population diffusion patterns of SARS‐CoV‐2 across exposure settings and evaluate their ramifications in epidemic growth.

In Hong Kong, COVID‐19 clusters delineated through case‐based surveillance during the pandemic period were classified into eight exposure settings: residence, home gathering, neighbourhood, workplace (office)/school, workplace (non‐office), daily activity, social activity and healthcare. Diffusion patterns characterised by outbreak size, speed and likelihood of spillover (cases seeding a new cluster) were compared among settings. With different clusters emerging, the lagged effect on effective reproduction number (R

t
) was evaluated.

Between January 2020 and January 2022, some 2800 clusters involving 14,202 cases were identified over five epidemic waves precipitated by outbreaks occurring in daily activity (wave I/III), social activity (wave II/IV) and neighbourhood (wave V—Omicron). Adjusted for variations by epidemic wave, the largest and fastest spread was observed in neighbourhood, averaging a size of 11.9 and daily generation of 1.18 cases per cluster. Spillover was the most common for social activity clusters with each of which normally breeding 3.73 onward clusters, compared to 0.18 for residential clusters. A cluster emerging in neighbourhood, social activity and daily activity was estimated to raise the R

t
by 0.021–0.025, 0.013–0.024 and 0.008–0.015, respectively, on the ensuing 7 days.

Neighbourhood and social activity outbreaks were inclined to induce epidemic spread, warranting the need for prioritised mitigation and targeted implementation of precautionary measures during both epidemics and peak season of respiratory infection.