An update of serial interval estimates for COVID-19: a meta-analysis

Abstract

Background: Serial interval (SI) is one of the most important parameter for COVID-19 modelling purposes as it is related to the reproduction rate of the infection. The first meta-analysis of serial interval were performed with a range of uncertainty in the estimate. This meta-analysis aimed to reduce the uncertainty estimates by assessing publications over a longer period. Methods: A literature search was performed for articles published between 1st December 2019 and 15th February 2022. It retrieved 117 eligible studies containing some 80 for 90 serial interval estimates. A random effects model was used. Heterogeneity was checked. To detect a publication bias, a funnel plot was performed using an Egger’s test. Results: For alpha variant, the serial interval was estimated at 5.17 days (95% CI = 4.87 – 5.47) with a significant heterogeneity (I2 = 97.1%). The meta-analysis did not exhibit evident publication bias (Egger’s test = −0.55, p = 0.58). The meta-analysis allowed for reducing uncertainty in estimating the serial interval, although subgroup analysis did not reduce it sufficiently and showed that studies using a gamma distribution of serial intervals exhibited the highest estimate of 5.6 days. Compared to the other variants of concern, alpha serial interval estimate was bigger than delta, 4.07 days, and omicron, 3.06 days. Conclusion: The meta-analysis was carried out as a real-time monitoring of this parameter to make a choice and a rapid assessment of the control measures implemented, and the effectiveness of the vaccination campaign. The meta-analysis was unable to provide a suitable estimate of serial intervals for COVID-19 modelling purposes although its uncertainty was reduced. Furthermore, serial intervals estimate for alpha variant was close to earlier reports and lower than previous publications, respectively. Another limitation is, that meta-analysis of COVID pandemic studies in principle contains and produces itself a significant source of heterogeneity.