Insights into COVID-19 Epidemiology and Control from Temporal Changes in Serial Interval Distributions in Hong Kong

Background: The serial interval distribution is used to approximate the generation time distribution, an essential parameter to infer the transmissibility (Rt) of an epidemic. However, serial interval distributions may change as an epidemic progresses rather than remaining constant.Method: We examined detailed contact tracing data on laboratory-confirmed cases of COVID-19 in Hong Kong during the five waves from January 2020 to July 2022. We reconstructed the transmission pairs and estimated time-varying effective serial interval distributions using Bayesian inferential framework with a sliding window of 7-14 days. We used regression models to identify the factors of temporal changes in serial intervals and quantify their respective impacts. Finally, we assessed the biases in estimating transmissibility using constant over time-varying serial interval distributions.Findings: 2497 transmission pairs were identified for the ancestral strain of SARS-CoV-2 during the first two years of the COVID-19 pandemic in Hong Kong. We found clear temporal changes in mean serial interval estimates within each epidemic wave studied and across waves, with mean serial intervals ranged from 5.5 days (95% CrI: 4.4, 6.6) to 2.7 (95% CrI: 2.2, 3.2) days. The mean serial intervals shortened or lengthened over time, which were found to be closely associated with the temporal variation in COVID-19 case profiles and public health and social measures and could lead to the biases in predicting Rt.Interpretation: Accounting for the impact of these factors, the time-varying quantification of serial interval distributions could lead to improved estimation of Rt, and provide additional insights into the impact of public health measures on transmission.Funding Information: This study was supported by the Health and Medical Research Fund (project no. 20190712); the Collaborative Research Fund of the Research Grants Council of the Hong Kong Special Administrative Region, China (project No. C7123-20G); AIR@InnoHK administered by Innovation and Technology Commission, European Research Council (grant no. 804744); the Grand Challenges ICODA pilot initiative, delivered by Health Data Research UK and funded by the Bill & Melinda Gates Foundation and the Minderoo Foundation. Declaration of Interests: BJC received honoraria from AstraZeneca, Fosun Pharma, GSK, Moderna, Pfizer, Roche, and Sanofi. The authors report no other potential conflicts of interest.