Synthesising data from multiple spatial scales and levels of detail to improve malaria transmission model predictions
Lead Research Organisation:
Imperial College London
Department Name: Infectious Disease Epidemiology
Abstract
In recent years the goal of malaria eradication has been put forward. Mathematical models of malaria transmission are valuable for understanding which control measures will be most effective, given the local conditions. I will statistically fit malaria models to a wide range of data sources. Then I will use data where the genetic make-up of the malaria parasites that people are infected with is known, in order to estimate how fast infection spreads through space on a local scale.
Together, these results will improve our ability to choose the most appropriate control measures. This will allow us to use limited resources more effectively, and hence have a greater chance of eliminating malaria altogether from some areas, or of controlling disease at a low level. The results will also help in designing strategies which aim to optimise the use of resources by targeting interventions at local foci of infection. The statistical methods developed will also be helpful for those modelling other infectious diseases who need to make use of diverse types of data, or who would like to use genetic data to estimate quantities that affect the spread of infection.
Together, these results will improve our ability to choose the most appropriate control measures. This will allow us to use limited resources more effectively, and hence have a greater chance of eliminating malaria altogether from some areas, or of controlling disease at a low level. The results will also help in designing strategies which aim to optimise the use of resources by targeting interventions at local foci of infection. The statistical methods developed will also be helpful for those modelling other infectious diseases who need to make use of diverse types of data, or who would like to use genetic data to estimate quantities that affect the spread of infection.
Technical Summary
Introduction
In recent years there has been a renewed effort to control malaria, with eradication as the ultimate goal. Mathematical models of malaria transmission are valuable tools to help us to understand which combinations of control measures will be most effective, and how to tailor control programmes to local conditions. However there has been limited work validating such models against data. Furthermore, malaria transmission is locally heterogeneous, but more needs to be known in order to effectively target local foci of infection (hotspots).
Aims
My aims are the following:
1) choose among possible models for acquisition of immunity to malaria;
2) improve these models‘ predictions of the impact of control measures;
3) estimate how fast malaria spreads through space at a local level;
4) determine which methods for detecting hotspots work best;
5) predict the impact of targeting of interventions at local hotspots of transmission.
Methodology
I will have access to many datasets with differing levels of detail and spatial scales. First, I will fit transmission models to the datasets that do not have control measures, to choose the best model for how various kinds of immunity develop. Then I will fit extensions of this model to datasets recording the impact of interventions.
I will develop and apply methods to use data of genotyped infections and their spatial location to estimate the rate at which malaria infections spread through space and time from person to person (via mosquitoes). I will then use simulation to compare methods for detecting hotspots, and combine these results to design and assess strategies which target hotspots.
Outcomes
Together, these results will improve the predictions of malaria models, allowing us to use limited resources more effectively, thus having a greater chance of eliminating malaria altogether from some areas, or of controlling disease at a low level. The results will also aid the design of strategies which aim to optimise the use of resources by targeting interventions at hotspots of infection.
The work will also result in methodological developments in a number of areas. It will require a comprehensive statistical framework for validating infectious disease models, combining methods that are not generally applied to complex biologically-based models. I will develop novel methods for estimating quantities relevant to infection processes from partially observed genetic data; and assess and develop methods to detect clusters of infectious disease transmission.
In recent years there has been a renewed effort to control malaria, with eradication as the ultimate goal. Mathematical models of malaria transmission are valuable tools to help us to understand which combinations of control measures will be most effective, and how to tailor control programmes to local conditions. However there has been limited work validating such models against data. Furthermore, malaria transmission is locally heterogeneous, but more needs to be known in order to effectively target local foci of infection (hotspots).
Aims
My aims are the following:
1) choose among possible models for acquisition of immunity to malaria;
2) improve these models‘ predictions of the impact of control measures;
3) estimate how fast malaria spreads through space at a local level;
4) determine which methods for detecting hotspots work best;
5) predict the impact of targeting of interventions at local hotspots of transmission.
Methodology
I will have access to many datasets with differing levels of detail and spatial scales. First, I will fit transmission models to the datasets that do not have control measures, to choose the best model for how various kinds of immunity develop. Then I will fit extensions of this model to datasets recording the impact of interventions.
I will develop and apply methods to use data of genotyped infections and their spatial location to estimate the rate at which malaria infections spread through space and time from person to person (via mosquitoes). I will then use simulation to compare methods for detecting hotspots, and combine these results to design and assess strategies which target hotspots.
Outcomes
Together, these results will improve the predictions of malaria models, allowing us to use limited resources more effectively, thus having a greater chance of eliminating malaria altogether from some areas, or of controlling disease at a low level. The results will also aid the design of strategies which aim to optimise the use of resources by targeting interventions at hotspots of infection.
The work will also result in methodological developments in a number of areas. It will require a comprehensive statistical framework for validating infectious disease models, combining methods that are not generally applied to complex biologically-based models. I will develop novel methods for estimating quantities relevant to infection processes from partially observed genetic data; and assess and develop methods to detect clusters of infectious disease transmission.