Meteorological indicators of dengue epidemics in non-endemic Northwest Argentina

TL;DR

This study develops a stochastic model integrating climate, social, and mobility factors to understand dengue outbreaks in a non-endemic city in Argentina, revealing meteorological influences on epidemic patterns and potential for forecasting.

Contribution

The paper introduces a coupled ecological and epidemiological model that incorporates climate data and human mobility to explain dengue outbreaks in a non-endemic region.

Findings

Climate variables like rainfall and temperature influence mosquito populations.

The model reproduces observed outbreak patterns from 2009 to 2016.

Proxies such as vector-to-host ratio and reproductive number relate to epidemic size.

Abstract

In the last two decades dengue cases increased significantly throughout the world. In several regions dengue re-emerged, particularly in Latin America, where dengue cases not only increased but also occurred more frequently. It is therefore necessary to understand the mechanisms that drive epidemic outbreaks in non-endemic regions, to help in the design of control strategies. We develop a stochastic model that includes climate variables, social structure, and mobility between a non-endemic city and an endemic area. We choose as a case study the non-endemic city of San Ram{\'o}n de la Nueva Or{\'a}n, located in Northwest Argentina. Human mobility is intense through the border with Bolivia, where dengue transmission is sustained during the whole year. City population was modelled as a meta-population taking into account households and population data for each patch. Climate variability was considered by including rainfall, relative humidity and temperature time series into the models. Those climatic variables were input of a mosquito population ecological model, which in turn is coupled to an epidemiological model. Different hypotheses regarding people's mobility between an endemic and non-endemic area are tested, taking into account the local climatic variation, typical of the non-endemic city. Simulations are qualitatively consistent with weekly clinical data reported from 2009 to 2016. Our model results allow to explain the observed pattern of outbreaks, that alternates large dengue epidemics and several years with smaller outbreaks. We found that the number of vectors per host and an effective reproductive number are proxies for large epidemics, both related with climate variability such as rainfall and temperature, opening the possibility to test these meteorological variables for forecast purposes.