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

Here, E(•) represents the expectation function and j is the region index corresponding to different regions (states). Term c j is the interception term of the j-th region to be estimated and it is variable from different regions, serving as the baseline scale of transmissibility in different states. The term t denotes the cumulative lag in the model, and 〈rainfall j,t 〉 represents the average monthly rainfall of the previous t months from the turning point, τ, of the j-th region. The reproduction numbers, R j , are estimated for different epidemics from the selected growth models. The regression coefficient, β, is to be estimated. Hence, the term (e β -1)*100% is the percentage change rate (of R), which can be interpreted as the percentage change in transmissibility due to a one unit (mm) increase in monthly rainfall. The framework of the regression is based on the exponential form of the predictor to model the expectation of transmissibility (e.g., R); this framework is inspired by previous work [28] [29] [30] [31] . To quantify the impacts of local rainfall, we calculate the percentage change rate with different cumulative lags (t) from 0-11 months and estimate their significant levels. Only the lag terms (t) with significant estimates are presented in this work. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/602706 doi: bioRxiv preprint