Elsevier established a free COVID-19 resource centre in January 2020, offering information on the novel coronavirus in English and Mandarin. The centre is hosted on Elsevier Connect, and Elsevier grants permission to freely reuse and analyze its research in PubMed Central and other public repositories. A study estimates the serial interval of COVID-19 infections, defined as the time between illness onset in an infected person (infector) and a secondary case (infectee). Using data from 28 infector-infectee pairs, the median serial interval was estimated at 4.0 days (95% credible interval: 3.1-4.9) when considering right truncation, and 4.6 days (3.5-5.9) when using only the most certain pairs. This suggests that a significant portion of transmission occurs before illness onset, and the serial interval is shorter than that of SARS, indicating potential bias in using SARS-based calculations. The study used a Bayesian approach with doubly interval censored likelihood to estimate the serial interval, considering the exponential growth phase of the epidemic. The results show that the serial interval is shorter than the incubation period, implying that pre-symptomatic transmission is likely. This has implications for contact tracing and isolation strategies, as many transmissions may occur before symptoms appear, making containment via isolation alone challenging. The study acknowledges potential biases in data collection due to varying criteria for illness onset across jurisdictions. However, it addresses data quality by including updated information and analyzing both certain and probable pairs. The findings suggest that the median serial interval of COVID-19 is close to or shorter than its incubation period, highlighting the need for improved public health strategies to account for pre-symptomatic transmission.Elsevier established a free COVID-19 resource centre in January 2020, offering information on the novel coronavirus in English and Mandarin. The centre is hosted on Elsevier Connect, and Elsevier grants permission to freely reuse and analyze its research in PubMed Central and other public repositories. A study estimates the serial interval of COVID-19 infections, defined as the time between illness onset in an infected person (infector) and a secondary case (infectee). Using data from 28 infector-infectee pairs, the median serial interval was estimated at 4.0 days (95% credible interval: 3.1-4.9) when considering right truncation, and 4.6 days (3.5-5.9) when using only the most certain pairs. This suggests that a significant portion of transmission occurs before illness onset, and the serial interval is shorter than that of SARS, indicating potential bias in using SARS-based calculations. The study used a Bayesian approach with doubly interval censored likelihood to estimate the serial interval, considering the exponential growth phase of the epidemic. The results show that the serial interval is shorter than the incubation period, implying that pre-symptomatic transmission is likely. This has implications for contact tracing and isolation strategies, as many transmissions may occur before symptoms appear, making containment via isolation alone challenging. The study acknowledges potential biases in data collection due to varying criteria for illness onset across jurisdictions. However, it addresses data quality by including updated information and analyzing both certain and probable pairs. The findings suggest that the median serial interval of COVID-19 is close to or shorter than its incubation period, highlighting the need for improved public health strategies to account for pre-symptomatic transmission.