Research

[rev_slider research]
Mathematical modeling of infectious and non-infectious diseases of global importance

Our primary interest here is the application of mathematical and computer science techniques that assimilates data with mechanistic models to reveal the dynamics and progression of major vector-borne, environmentally-mediated and emerging chronic diseases, ranging from neglected tropical diseases, such as lymphatic filariasis, onchocerciasis and schistosomiasis, to malaria, dengue, cholera, melioidosis and diabetes. The goal is to uncover generic and specific transmission structures and processes, and their configuration across space and time, in order to gain a better understanding of factors underpinning the emergence, spread and persistence of disease transmission under different socio-ecological settings. The ultimate goal of this research is to use these results to learn the best strategies to disrupt transmission and achieve the sustainable control of these diseases. We couple mathematical modelling with Bayesian calibration and model updating methodologies, and computational tools spanning data discovery, application of machine and deep learning to reveal patterns in data and modelling results, and take advantage of hardware advances to exploit grid-based computational speed ups to execute this work. Our present partners include the Ministries of Health of Tanzania and Uganda, the Carter Center, The Gates Foundation, WHO, IBM Research, and the University of South Florida.

Geospatial modeling of disease distributions

A major theme of the laboratory’s current interest is mapping and undertaking spatial analysis of disease niches and the landscape and socio-environmental factors that influence distribution of disease over space and time. This work is highly multidisciplinary and leverages multidimensional data encompassing the environment, socio-demographics, land use/land cover, economic activity (GDP), social sensitivity and adaptive capacity of communities to model spatio-temporal disease risk. Techniques from geographic information systems, remote sensing, models of economic activity and land use, climate modelling and spatial analysis using both species distribution modelling (SDM) and Bayesian spatial models implemented in INLA, and mathematical models of disease transmission, are used conjointly in executing this theme. Current work focuses on malaria spatial risk modelling in Africa, including modelling the effects of climate, landuse and socio-economic change on intervention outcomes, as well as the landscape ecology and control of melioidosis in Malaysia.

Social epidemiology

Social factors, ranging from cultural and behavioral variables to history, politics and institutional arrangements, in conjunction with ecological factors and geography, govern the transmission of diseases in human communities. Our research is two-pronged: first, to investigate the specific social determinants of diseases, and second to use the information gained to derive interventions and governance structures that can modify these factors to bring about sustainable and beneficial change. We use constructs from social epidemiology, anthropology, theories of social change, sustainability theory and complexity science to investigate this problem, focusing on developing and testing new  governance models grounded in social science for sustainably controlling neglected tropical diseases working with partners from Tanzania and the London School of Economics, and obesity and diabetes working with partners from India.

Policy analysis

Our major goal here is two-fold. First to develop a platform-strategy for system dynamics modelling of diseases, and second to learn control policies from the resulting platform-based model simulations. We are working with the Center for Research Computing and iCeNSA, ND, and IBM Research to develop the computational platform that will allow automated data discovery, coupling of models to data and running of intervention simulations feeding into spatially-explicit policy analysis tools, for guiding the learning of control policies that will take a fuller account of deep uncertainty and dynamically complex multi-dimensional societal issues. This work is based on leveraging mathematical modelling, software engineering and decision analysis tools, ranging from exploratory modeling and analysis, to using machine learning and robust multi-objective optimization methods for learning disease control strategies. Addressing spatial elements to the control problem is a major and challenging novelty.