Investigating setting-specific superspreading potential and generation intervals of COVID-19 in Hong Kong.

Abstract

Superspreading is an important feature of COVID-19 transmission dynamics, but few studies have investigated this feature stratified by transmission setting. Using detailed clustering data comprising 8647 COVID-19 cases confirmed in Hong Kong between 2020 and 2021, we estimated the mean number of new infections expected in a transmission cluster ( <math xmlns="http://www.w3.org/1998/Math/MathML"> <msub><mrow><mi>C</mi></mrow> <mrow><mi>Z</mi></mrow> </msub> </math> ) and the degree of overdispersion (k) by setting. Estimates of <math xmlns="http://www.w3.org/1998/Math/MathML"> <msub><mrow><mi>C</mi></mrow> <mrow><mi>Z</mi></mrow> </msub> </math> ranged within 0.4-7.1 across eight settings, with highest <math xmlns="http://www.w3.org/1998/Math/MathML"> <msub><mrow><mi>C</mi></mrow> <mrow><mi>Z</mi></mrow> </msub> </math> in the close-social indoor setting that an average of seven new infections per cluster was expected. Transmission was most heterogeneous (k = 0.05) in retail setting and least heterogeneous (k = 1.1) in households, where smaller k indicates greater overdispersion and superspreading potential. Point-estimates of the mean generation interval (GI) ranged within 4.4-7.0, and settings with shorter mean realized GIs were associated with smaller cluster sizes. Here, we show that superspreading potential and generation intervals can vary across settings, strengthening the need for setting-specific interventions.