Large-scale Lassa fever outbreaks in Nigeria: quantifying the association between disease reproduction number and local rainfall

Shi Zhao; Salihu S. Musa; Hao Fu; Daihai He; Jing Qin

We choose to use the average reproduction number (R) rather than the instantaneous reproduction number, commonly denoted by R t [22, 23] , as the measurement of the LASV transmissibility. The R t is a series of time-dependent reproduction numbers, namely, timedependent effective reproduction numbers, which can be estimated by a renewable equation [20, 22, 23, 33, 34] . The factors that affect the changing dynamics of R t, include (i) the depletion of the susceptible population [28], (ii) the change (usually it is the improvement) in the unmeasurable disease control efforts (e.g., contract tracing) and local awareness of the outbreak [29], and (iii) the natural features of the pathogen (e.g., its original infectivity and other interepidemic factors) [28, 29, 31]. The estimated R in this work is a point estimate to summarize LASV transmissibility over a whole epidemic. Hence, the temporal changes of the susceptible population in (i), above, and local disease awareness and control related to (ii) do not affect R. With respect to point (iii) and other heterogeneities of outbreaks in different . CC-BY 4.0 International license author/funder. It is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/602706 doi: bioRxiv preprint regions, we account for this issue by including the "region" dummy variables in the LRER model in Eqn (3). These dummy variables serve as random effects to offset regional heterogeneities in LF epidemics. Therefore, we can then quantify a consistent effect (the β in Eqn (3)) of the lagged rainfall on the LASV R estimate.