Analytics to Eliminate Dengue

Dengue elimination is a real possibility with the use of new technologies, but how do communities know it is really gone? Embedded within a unique collaboration, this Solution will develop a novel data analytics system and dashboard, to better understand what surveillance, from cases of dengue to mosquito data, are best used to understand local elimination of dengue. In its initial stage this specific funding will facilitate a ‘proof of concept’ and act as a springboard to gain support from MIT Solve partners in dashboard development. The proof of concept will illustrate how useful different data streams are for assessing dengue elimination, and be compared to their resource costs. The dashboard development will help us understand what requirements stakeholders have in a dashboard, and how this can be successfully developed. By supporting development of this Solution, limited resources in public health can be put to best use to eliminate dengue.

Dengue is a viral disease that is largely endemic in the tropics, where >40% of the world's population are at risk. Each year, ~390 million dengue infections occur, resulting in ~25,000 deaths. Prevention includes reducing the habitat of the mosquito vector, Aedes aegypti, with limited effectiveness, until recently. Introduction of Wolbachia-carrying mosquitoes have been shown to naturally prevent mosquito borne diseases such as dengue. 

Our partners, the World Mosquito Program (WMP) have shown Wolbachia to be highly effective. In Yogyakarta (Indonesia), a randomised controlled trial (RCT) illustrated a 77% reduction in dengue incidence over 3 years. Wolbachia is so effective that elimination of dengue is now possible, but the steps towards this being economically manageable is not yet clear. Significant resource was put in place to monitor the RCT, which is not feasible across low and middle income settings.

The problem we seek to solve is what data are required to monitor Wolbachia introduction in the short and long term to enable elimination of dengue disease, how should data collection be carried out to monitor dengue, and what digital tools can be used to facilitate ownership of this in local communities? 

Our Solution is the development of a novel data analytics algorithm and dashboard that will identify surveillance targets for meeting dengue elimination. This algorithm and platform are unique because we are focussed on achieving elimination as opposed to the classical epidemiological task of identifying cases.

The algorithm will require refinement of the statistical approaches used. In this proof of concept stage we will apply previously used approaches to data that has already been collected in Indonesia, and review and critique other available methods. We anticipate that this seed funding will illustrate the utility of the approach and the ability of team members to work together, and help secure future funding to provide the necessary resource for further algorithm development.

This seed funding will also act as a platform for dashboard development as we welcome support and engagement from Solve partners. In this initial stage we will work with Solve partners and in-country stakeholders to understand stakeholder needs for a dashboard, providing criteria that the dashboard needs to meet. This work will provide the scaffolding for future dashboard development, where we aim to secure further resource to achieve this.

The primary audience are health workers in dengue control and elimination, especially where Wolbachia has been introduced. Our partners, the World Mosquito Program (WMP) have projects in 10 countries; Brazil, Colombia, Mexico, Indonesia, Sri Lanka, Vietnam, Australia, Fiji, Kiribati, New Caledonia and Vanuatu, where dengue is endemic. The Indonesian project is most relevant; here detailed data are collected as part of a randomised controlled trial (RCT) meaning that the algorithm and platform can be refined and validated. Wolbachia has been released into the wider community where identifying suitable surveillance to assess elimination is warranted.

More broadly, the algorithm and platform will provide a novel analytics framework that illustrates the power of Wolbachia in preventing dengue, but also illustrates best practice in illustrating elimination of an infectious disease. We envisage that the platform will serve many common requirements (production of elimination targets, inputting data, and visualising data for stakeholders), whilst being adaptable for specific needs. We are confident that the unique partnership between WMP and local communities will act as a springboard to develop this tool, and engagement with MIT Solve can only strengthen this.

There are many effective interventions against infectious diseases, and recent advances have illustrated that Wolbachia use to eliminate dengue has now joined this group. The problem is that the scientific trials have been very resource intensive and this is not scalable for widespread deployment of Wolbachia – resource is always limited in public health This Solution aims to support the widespread deployment by identifying what data are best suited for widespread use and developing the analytics platform to support its use. Affected communities will consequently have better engagement with current progress in dengue control and elimination.

The pilot phase has been selected because resource is required to develop the algorithm and dashboard, and we aim to consolidate out partnership between WMP, Aus Vet and LSHTM. The PI has already illustrated the proof of concept through application of the framework to poliovirus in humans (O’Reilly et al. Epi & Infe (2020) doi:10.1017/S0950268820001004). We are confident that the partnership between WMP and LSHTM will be successful, as WMP already collaborate with Prof Nick Jewell (at LSHTM and co-PI on the project) regarding the RCT, and the PI has previously collaborated with WMP on a previous project (estimating burden and cost-effectiveness of Wolbachia deployment across Indonesia, O’Reilly et al. BMC Med (2019) doi.org/10.1186/s12916-019-1396-4). 

Use of statistical algorithms to assess dengue elimination is innovative; the classical approach in epidemiology is to count cases but here we count the absence of cases and other measures that contribute to assessing disease elimination. 

In this pilot phase our Solution aims to develop and validate the algorithm before being used in public health; it would be unethical to deploy an algorithm to assess elimination before first assessing the accuracy of the approach. With further support, even within the first year, we will aim for the algorithm and dashboard to serve a number of engaged communities with Yogyakarta, and obtain feedback of the utility of the dashboard. After 5 years, with successful development we would expect the dashboard to serve large communities in Indonesia, and with translation can be expanded to other countries where Wolbachia is in use.

In the first year, the primary focus of the Solution team will be to develop the algorithm, and validate in silico and using data from Indonesia. Platform development will begin, with the aim of having a beta-version for data input and visualisation.

In years 2-3, the algorithm will be applied to data within the platform and outputs generated. At this stage, engagement with end users will be critical to assess the requirements, provide training, and refine the platform to the users needs. We also expect tangible assessments of elimination within specific settings, thus providing a “case use” for a wider audience (stakeholders in public health). It is also anticipated that in this period Wolbachia will have been scaled up to larger communities within and beyond the 10 countries where it is already in use. The algorithm may require refinement to specific settings in this latter stage, and better account for dengue importations. We envisage that as the Solution develops, wider use of the targets and platform will occur across Wolbachia areas, illustrating suitable scaling of this assessment tool.

Note that these impact goals are reliant on sustained funding of the research beyond the current round of funding.

The LSHTM is a world-leading centre for research, postgraduate studies and continuing education in public and global health. LSHTM has a strong international presence with over 3,000 staff and 4,000 students working in the UK and countries around the world, and an annual research income of £180 million. LSHTM is one of the highest-rated research institutions in the UK, and was named University of the Year in the Times Higher Education Awards 2016. Our mission is to improve health and health equity in the UK and worldwide; working in partnership to achieve excellence in public and global health research. 

Within the Team there are individuals based at LSHTM (x2), Aus Vet (x1), and WMP (x2). With further resource we would like to expand this team.

The PI has a strong background in quantitative epidemiology and mathematical modelling, and is experienced in developing relevant algorithms for improving analysis of public health problems. The research team at LSHTM includes Professor Nicholas Jewell, who is an internationally renowned professor in biostatistics, who has partnered with the WMP for several years to design the RCT in Indonesia. 

World Mosquito Program (WMP - https://www.worldmosquitoprogram.org) is a not-for-profit initiative that exists to protect the global community from mosquito-borne diseases. There are two regional hubs (Vietnam and Australia) and projects within 11 countries where Wolbachia-carrying mosquitoes have been released to reduce the incidence of local dengue infection. Cameron Simmons (Exec Team & Regional Director Oceania) is actively involved with the project based in Indonesia.

As part of the Solution, we will actively engage with health workers based in Indonesia and other settings. Working with affected communities is integral to Wolbachia use; it is community volunteers who rear and release Wolbachia mosquitoes and trap mosquitoes to monitor introduction. We will leverage the connections within WMP to understand how the platform and algorithm can be used by health workers to improve data collection and assessment of dengue elimination.  

All Team members value diversity in the workplace. Should this Solution be further resourced we will form a Steering Committee of stakeholders which will be balanced in terms of diversity, and will aim to provide appropriate steer in the direction of the programme of work.

MIT-Solve will help support the development of an innovative algorithm and platform, and the mentoring to strengthen the Team’s expertise in software and platform development. The Solution has a clear potential to help expedite informative data collection that informs dengue elimination, but requires investment in analysis and development of the platform. The involvement of MIT-Solve with the Solution will be highly valued; Members have considerable experience in scale-up of data driven interventions, and many have software development expertise that would complement our Solution.   

This Solution leverages data science to better inform estimates of dengue disease elimination from data.