Enhancing Dengue Fever Modeling Through a Multi-scale Analysis Framework – A Case Study in the Central Valley of Costa Rica
MetadataShow full metadata
Dengue fever is the second most widespread tropical disease after malaria and affects populations of more than 100 countries (Derouich and Boutayeb 2006). It is considered one of the most severe viral diseases in terms of morbidity and mortality (Guzmán and Kourí 2004). Over the last decade, dengue fever has become the most wide-spread vector-borne disease in Costa Rica (CCSS 2008). However, only a few research studies have been conducted in Costa Rica to investigate the factors influencing the rates of dengue fever. While GIS and statistical analysis have been used in research studies, agent-based modeling has not been applied to the study of dengue. This study emphasizes how traditional macro level GIS analysis and the implementation of a micro-level dengue fever agent-based model can be merged into a novel framework for the study of dengue fever in the Central Valley of Costa Rica. One of the main objectives was to develop an agent-based model, which integrated GIS to simulate the spread of dengue fever disease in an urban environment, as a result of an individual’s interactions in a geospatial context. Precipitation, temperature, socio-economic and demographic variables were analyzed using these technologies to identify the factors affecting the rates of dengue fever in the study area. GIS was used to map dengue risk and the spatial distribution and vulnerability of dengue risk in the study area using geographically weighted regression. The Dengue Fever Agent Based Model (DFABM) was developed using the Java programming language and the open-source MASON simulator, a multi-threaded agent-based simulation platform. The DFABM represented daily movements and interactions of people, the environment, and the vector, relative to dengue cases. The simulation examined detailed data about each scenario to identify the significant events occurring during outbreaks. The data employed included the number of susceptible, exposed, and infected people. The locations (described by longitude and latitude) and temporal data describing infected individuals were also collected for analysis. The DFABM tracked the factors affecting dengue fever, including precipitation, temperature, and the most important demographic and socio-economic characteristics of the population in the study area. The research questions guiding this study were: Does a community-level dengue fever agent-based model (DFABM) produce results comparable (agree) to those produced by traditional macro-level GIS analysis? Does a community-based dengue fever agent-based model (DFABM) enhance traditional geographic information system analysis and could it aid in predicting future dengue fever outbreaks? The findings of the community-level ABM generated similar results to (they were in agreement with) the traditional GIS analysis technique. Likewise, the DFABM enhanced traditional methods of analysis and could aid in predicting dengue fever outbreaks. Therefore, the coupling of GIS and ABM was the optimal research design for the study of dengue fever in the Central Valley of Costa Rica.