Please join us for a GLOBES event on Monday November 20th at 12:00 pm as part of the GLOBES mini-grant series as Rachel Oidtman gives a talk on the role of environmental conditions and importation on dengue epidemics in China.
Disentangling the relative roles of environmental conditions and importation in driving interannual variation in dengue epidemics in Guangzhou, China
R.J. Oidtman*1, Z. Huang2 , S. Lai2,3, Y. Li2 , R.C. Reiner4 , A.J. Tatem3,5, A.S. Siraj1 , H. Yu2 , T.A. Perkins1
1 University of Notre Dame, USA, 2 Center for Disease Control and Prevention, China, 3 University of Southampton, UK, 4 University of Washington, USA, 5 Flowminder Foundation, Sweden
In 2014, the city of Guangzhou, China experienced a dengue virus epidemic with nearly 40,000 locally transmitted, symptomatic cases. In the nine years prior to 2014, the number of reported, locally transmitted cases in each year ranged 0-1,279. At the same time, the relative roles of different factors capable of driving such extensive interannual variability in epidemic size remain unclear. To determine whether elevated dengue virus importation or anomalous environmental conditions drove this unusually large epidemic, we incorporated weather data, mosquito surveillance data, and imported case data from 2005-2015 into a stochastic TSIR model. We used adaptive Markov chain Monte Carlo methods to estimate the joint posterior distribution of model parameters relating these putative drivers to time-varying transmission patterns under a Bayesian framework. To determine the extent to which environmental conditions or importation patterns drove these dynamics, we conducted a simulation experiment under a factorial design with all combinations of environmental conditions and importation patterns from different years. Using simulations of our fitted stochastic model, we validated the model by showing that it can reproduce general features of seasonal transmission, such as relative epidemic size across years. In the simulation experiment, in which all combinations of environmental conditions and importation patterns were considered, we found that interannual variation in epidemic size was roughly equally attributable to interannual variation in environmental conditions and importation patterns. Our results indicate that large seasonal dengue epidemics in China, such as 2014, result from a combination of both environmental conditions and importation patterns being unusually suitable for local transmission in a given year. In other years (e.g., 2015), it appears that low suitability for one factor (e.g., environmental conditions) offsets high suitability for another factor (e.g., importation patterns). These results have implications for policies related to surveillance and outbreak response.