Skip to main content


Malaria Elimination in Haiti


The Quantitative Spatial Ecology, Evolution, and Environment (QSE3) IGERT is an NSF-funded program at the University of Florida. QSE3 involves students and faculty from ten departments. Our program focuses on the critically important and conceptually unifying theme of spatial dynamics, covering topics such as evolution and spread of emerging pathogens; causes and consequences of shifting species distributions; and conservation of species in patchy habitats. To tackle these issues, graduate students need an arsenal of tools from a variety of fields, including mathematics, biology, geography, and statistics. Each cohort of 5 students comprises students from each of these and related departments. In seeking to train scientists who embrace a new philosophy about quantitative tools, we also prepare individuals to speak to colleagues from different disciplines, and who can function as part of intellectually diverse teams by bringing different tools to bear on shared problems. To integrate our philosophy into practice, each cohort is required to team with an outside client to define a research project, which may be peripheral to students own research interests. These projects range from collaborations with departments and faculty on campus to problems brought to us from outside organizations. The students work together formally in a workshop for an entire year on the project, and continue to work on manuscripts and offer presentations beyond their workshop year.

We are now working on our second workshop. The students in the second cohort are working on a project for the Center for Disease Control to determine the feasibility of malaria eradication in Haiti. The early part of the workshop year was spent finding, gathering, and organizing available data, consulting the literature for reasonable parameter estimates, and developing theoretical models. The students began with a theoretical model of malarial dynamics using the two-patch Ross-MacDonald model. The two-patch model incorporates the role of human movement and spatial heterogeneity, where transmission rates (e.g., rate of spread among humans within a patch) may differ between patches. The two-patch model yields the reproduction number, R0, a threshold value that indicates persistence or extinction of the disease. The two-patch model, although simplified, enables testing the sensitivity and elasticity of model parameters. Performing a sensitivity of R0 under different parameterizations of the model, helps identify which patch to target for control measures and what type of control measure to implement. For example, if the extrinsic incubation period is longer than the average mosquito lifespan, control measures targeting the mosquito biting rate are more effective; conversely, if shorter, then targeting mosquito death rate are more effective. Students are now working on extending the simple two-patch model to a multipatch theoretical model, and then incorporating the empirical data for human movement in Haiti.

The mutipatch model provides a more realistic representation of malarial dynamics of Haiti, with numerous patches on the island that differ in transmission rates. The multipatch model facilitates examination of interaction between heterogeneity in transmissibility and connectivity in a simulated multi-patch system. The preliminary results showed that there is an ‘optimum’ level of connectivity that maximizes disease prevalence in a heterogeneous system. The optimum level occurs when the level of connectivity allows for desynchronization of the individual dynamics within patches. These results will then form the basis to test the 3rd stage of the project when they actually apply the multi-patch model to malaria dynamics in Haiti.

Human movement has a profound impact on the spread of malaria, and on the reintroduction of it to areas that have experienced extinction of the disease. Rates of human movement between patches also influence the optimal control measures. Results from the initial two-patch model indicated, counterintuitively, that control measures may work best when applied to a patch with lower transmission rates when immigration from patches with high transmission rates is high. So the representation of human movement will have a strong impact on model results and on malarial dynamics in reality. Human movement can occur on a variety of time scales ranging from a long time scale (permanent relocation of households) to short time scale (a trip to the market). Movements at different scales require different types of data, ranging from census data to cell phone data. Likewise, more local movements will maintain malaria within a patch, but longer movements may reintroduce malaria into patches that had experienced extinction. The students continue to acquire these data, as well as additional covariate data that are likely to affect how people move from place to place (i.e., from rural to urban areas, etc.). They are developing statistical and GIS models to accommodate these different types of data. Once the models of human movement are developed to provide relevant results, they will be merged with the two-patch and multi-patch test models, and eventually into the final model that represents the malarial dynamics of Haiti.

The students organized a meeting in Gainesville with members of the CDC and the Clinton Health Access Initiative to present their work to date and to receive feedback on their proposed work and guidance toward the future. They hoped to ensure that their work would be useful to the clients. The visitors were impressed with their progress, provided avenues for additional types of data, and provided suggestions toward the future. The meeting went very well and helped set the stage for ongoing activities after the conclusion of the workshop.

Address Goals

Our IGERT exemplifies the education and training goals of the NSF IGERT program by working in multidisciplinary teams on several levels. From the time that students join the program, they are exposed to students and faculty from the ten departments, including math, biology, geography, and statistics, and related areas. During our weekly colloquium, most of the program’s students, as well as some affiliated with the program though not fellows, join together to discuss multidisciplinary research. The students have previously conducted mini-workshops in colloquium, to teach their tools to others and to learn the tools of other disciplines. They also have studied proposals and results of previously funded multidisciplinary research, and they have worked together to resolve each others research problems using a multidisciplinary approach. They are required to take courses outside of their home discipline, and then are called upon during their workshop year to call on those skills to work with students, faculty, and even outside clients, to resolve a problem by bringing their entire suite of skills to the problem. The students who are fellows in our program will be well equipped to handle problems that extend beyond the boundaries of traditional, within-disciple research.

Our second cohort of students are in their workshop year. They have developed models that have gone beyond a simple application of a proven model to a real-world problem. They have developed novel methods to approach epidemiological problems that can be applied to disease eradication. Their extension of modeling of malarial dynamics to multiple patches with heterogeneous transmission rates and accounting for human movement represent progress in the field of disease dynamics. Two of our workshop students have received funding to continue their theoretical work on the optimal means of control of malaria in spatially and temporally heterogeneous environments, beginning in Fall 2011, following the end of the workshop year.