DengueME: A Tool for the Modeling and Simulation of Dengue Spatiotemporal Dynamics

Abstract

The prevention and control of dengue are great public health challenges for many countries, particularly since 2015, as other arboviruses have been observed to interact significantly with dengue virus. Different approaches and methodologies have been proposed and discussed by the research community. An important tool widely used is modeling and simulation, which help us to understand epidemic dynamics and create scenarios to support planning and decision making processes. With this aim, we proposed and developed DengueME, a collaborative open source platform to simulate dengue disease and its vector's dynamics. It supports compartmental and individual-based models, implemented over a GIS database, that represent Aedes aegypti population dynamics, human demography, human mobility, urban landscape and dengue transmission mediated by human and mosquito encounters. A user-friendly graphical interface was developed to facilitate model configuration and data input, and a library of models was developed to support teaching-learning activities. DengueME was applied in study cases and evaluated by specialists. Other improvements will be made in future work, to enhance its extensibility and usability.

Keywords: Aedes aegypti; DengueME; dengue; framework; model; modeling; simulation; spatiotemporal.

Figure 1

Exponential growth of publications on dengue (blue circles) and specifically on dengue models (red squares) in the last two decades (search done in the ISI-Web of knowledge database) [63].

Figure 2

DengueME architecture [63].

Figure 3

Diagram showing the modeling process using the DengueME Visual Development Environment [63].

Figure 4

Graphical interface of the DengueME Visual Development Environment (VDE) [63].

Figure 5

Compartmental model SIR-SI, where the human population is represented by three compartments: susceptible (Sh), infectious (Ih) and removed (Rh); and the mosquito population by two compartments: susceptible (Sv) and infectious (Iv).

Figure 6

Modeling and simulation using the DengueME VDE: (top left panel)—Wizard to support the creation/selection of models; (right panel)—Graphical interface for parameterization; (bottom left panel)— Execution of the model and visualization of the simulation results.

Figure 7

Sensitivity analysis using DengueME: (a) sensitivity of the predicted epidemic curve to variations in the biting rate parameter and; (b) recovery rate parameter. The Y-axis is the proportion infected in the human population. The black dashed line is the model output with default values (as used by Nishiura (2006) [106]).

Figure 8

Study area, Ilha do Governador. (a) Satellite image (Google Earth); (b) map of census tracts; (c) simulated map generated by DengueME. Highlighted blue areas are the sites of application of adulticide [63].

Figure 9

Using DengueME to simulate locally-applied chemical interventions. Comparison between the the vector population in the study area with and without the application of adulticide. (a) Global impact considering the entire study area; (b) local impact in a 100 × 100 m area [63].

Figure 10

Using DengueME to simulate dengue spread from commercial to residential areas of a city. Panels show three moments of a simulated epidemic in the study area. (a) Beginning of the simulation with a few hotspots; (b) propagation waves from the hotspots; (c) overall dissemination and pockets of immunity [63].

Figure 11

DengueME output. Time series of (a) susceptible, infected and recovered humans and (b) susceptible and infected mosquitoes, generated by DengueME simulations [63].

Figure 12

User’s evaluation of the DengueME graphical interface: (a) support for learning, using and understanding dengue models; (b) applicability potential for supporting decision making, researching and teaching [63].