Practical considerations for measuring the effective reproductive number, Rt

• DOI: 10.1371/journal.pcbi.1008409
• Publication Date: 2020-12-10T18:56:16Z
• Authors: Katelyn M. Gostic; Lauren McGough; Edward B. Baskerv...; Sam Abbott; Keya Joshi; Christine Tedijanto; Rebecca Kahn; Rene Niehus; James A. Hay; Pablo M. De Salazar; Joel Hellewell; Sophie Meakin; James D. Munday; Nikos I. Bosse; Katharine Sherrat; Robin N. Thompson; Laura F. White; Jana S. Huisman; Jérémie Scire; Sebastian Bonhoeffer; Tanja Stadler; Jacco Wallinga; Sebastian Funk; Marc Lipsitch; Sarah Cobey
• Keywords: 0303 health sciences; Models, Statistical; QH301-705.5; SARS-CoV-2; Basic Reproduction Number; COVID-19; Computational Biology; Article; 3. Good health; 03 medical and health sciences; Perspective; Humans; Biology (General)

Estimation of the effective reproductive numberRtis important for detecting changes in disease transmission over time. During the Coronavirus Disease 2019 (COVID-19) pandemic, policy makers and public health officials are usingRtto assess the effectiveness of interventions and to inform policy. However, estimation ofRtfrom available data presents several challenges, with critical implications for the interpretation of the course of the pandemic. The purpose of this document is to summarize these challenges, illustrate them with examples from synthetic data, and, where possible, make recommendations. For near real-time estimation ofRt, we recommend the approach of Cori and colleagues, which uses data from before timetand empirical estimates of the distribution of time between infections. Methods that require data from after timet, such as Wallinga and Teunis, are conceptually and methodologically less suited for near real-time estimation, but may be appropriate for retrospective analyses of how individuals infected at different time points contributed to the spread. We advise caution when using methods derived from the approach of Bettencourt and Ribeiro, as the resultingRtestimates may be biased if the underlying structural assumptions are not met. Two key challenges common to all approaches are accurate specification of the generation interval and reconstruction of the time series of new infections from observations occurring long after the moment of transmission. Naive approaches for dealing with observation delays, such as subtracting delays sampled from a distribution, can introduce bias. We provide suggestions for how to mitigate this and other technical challenges and highlight open problems inRtestimation.