Estimating the generation interval for COVID-19 based on symptom onset data

Tapiwa Ganyani; Cecile Kremer; Dongxuan Chen; Andrea Torneri; Christel Faes; Jacco Wallinga; Niel Hens

The most well-known method to estimate the serial interval distribution from line list data is the likelihood-based estimation method proposed by Wallinga and Teunis [8] . In 2012, Hens et al. [9] proposed using the Expectation-Maximisation (EM) algorithm to estimate the generation interval distribution from incomplete line list data based on the method by [8] and allowing for auxiliary information to be used in assigning potential infector-infectee pairs. Te Beest et al [10] used a Markov chain Monte Carlo (MCMC) approach as an alternative to the EM-algorithm, to facilitate taking into account uncertainty related to the dates of symptom onset. In this paper, we use an MCMC approach to estimate, next to the serial interval distribution, the generation interval distribution upon specification of the incubation period distribution. We compare the impact of differences amongst previous estimates of the incubation period distribution for COVID-19 and analyse data on clusters of confirmed cases from Singapore (January 21 to February 26) and Tianjin, China (January 14 to February 27).