Mapping the growing global burden of dengue to help countries plan for the next decade of dengue control

Dengue is one of the fastest growing global infectious diseases with an 8-fold increase in reported cases since 2000. Because a large proportion of dengue virus infections are asymptomatic and the quality of disease surveillance is constantly changing, computational models have been central to estimating the global burden of dengue. However, a key limitation of current models is that they do not accurately estimate how this burden changes year-on-year. This prevents model-based burden estimates of dengue having the same impact they do in the fields of HIV, TB and malaria where they are actively used to track progress towards international targets. This year, new goals for reducing the global burden of dengue between 2021-2030 will be set. Data gaps and limitations of current modelling approaches constraint our ability to track country progress, understand how existing interventions are working and suggest how control programmes need to change to meet these new targets.

In this fellowship I will develop novel detailed global dengue models that allow a new generation of predictions of the past present and future global burden of dengue to be made. These models can be used to answer three main aims:

Aim 1: How has the global growth in dengue burden changed since 2000?
By pairing a new global dengue database with geostatistical and mathematical modelling techniques I will generate unbiased burden estimates of the growth in dengue burden 2000-2020.

Aim 2: How effective have policy changes to reduce dengue deaths been?
Through an analysis of over 4 million individual-level patient records in São Paulo (Brazil) and the Philippines, I will investigate how changes in treatment seeking, diagnosis and clinical management have reduced the risk of dengue death. This will measure the impact of current interventions on mortality and identify where further gains could be made.

Aim 3: By how much will dengue burden grow 2021-2030 in each country?
By projecting the models from Aim 1 into the future taking into account changes in climate, urbanisation and growing levels of immunity to dengue, robust predictions of dengue burden 2021-2030 can be made. A dengue control feasibility assessment will then identify the specific barriers each country faces in meeting their 2021-2030 targets.

This fellowship aims to improve our understanding of how the global burden of dengue is changing at a pivotal time. Aim 1 will focus on developing a new generation of dengue models to estimate past changes in incidence 2000-2020 to understand why we were unable to contain dengue expansion. Aim 2 will answer key questions about the effectiveness of current efforts to reduce dengue deaths. Aim 3 will estimate future growth in dengue incidence allowing countries to decide how to best address this growing problem. All the data, models and predictions will be hosted on a new dedicated website that allows researchers and government officials to explore these estimates in detail and with full transparency. Beyond dengue, these models will also provide fundamental insights into how modern emerging infectious diseases are capable of uncontained global spread and help design new strategies against future pandemics.

Dengue is one of the fastest growing uncontained emerging infectious diseases. Computational models play an important role in estimating disease burden because of the limitations of epidemiological data. This year, new goals for reducing the global burden of dengue between 2021-2030 will be set, but current models are insufficient for tracking progress. In this fellowship I will develop the next generation of global dengue models that take into account the spatio-temporal variation in multiple types of epidemiological data to predict the past, present and future global burden of dengue.

A systematic search will assemble incidence and prevalence data at a sub-national level with geospatial models used to make predictions for data deficient areas. A Bayesian multi-serotype dengue transmission model fit to dengue incidence and prevalence data will then give consensus estimates of how dengue burden has changed 2000-2020. These predictions will give insights into how dengue was able to spread so rapidly and help design strategies to contain future arboviral pandemics.

With data on >4 million individual-level dengue patient records I will characterise spatio-temporal trends and causal determinants of delayed treatment seeking and misdiagnosis. Using multivariate regression approaches I will then assess the contribution of each of these to the risk of severe disease and death after accounting for relevant confounders. This will give new insights into how to prevent dengue deaths through non-therapeutic approaches.

Projecting these models using future climate, urbanisation and immunity scenarios will allow new predictions of future changes in dengue burden. A national feasibility assessment will then rank countries in their ability to meet 2021-2030 goals based on epidemiological, operational and financial barriers. This work can be used to define target product profiles of novel interventions and identify under or over-achieving control programmes.