Statistical model enables improved disease management in Brazil
2 mins to read
Dr Theo Economou and a research team at the University of Exeter have developed a statistical data modelling framework for prediction of disease burden, and thus warn different regions of Brazil against the threat of major diseases.
Predicting and warning against disease outbreak in Brazil, relies heavily on the availability of data related to the number of disease cases. Conventional data modelling approaches for estimating and predicting disease occurrence are designed for data at annual/seasonal time scales. However, on smaller time scales (e.g. weeks), which are more relevant to real-time warnings, the disease data is subject to reporting delay. This can be caused by pending laboratory confirmation and logistic and infrastructure difficulties.
Dengue fever is a mosquito-borne disease
The reliability of any public warnings and health resource allocation is dependent on the amount of delay.
In 2015, the warning system InfoDengue was launched in an effort to reduce Dengue risk in Rio de Janeiro. However, the underlying method for correcting the delays in the data was too basic.
Research led by Dr Theo Economou (Senior Lecturer in Statistics at Exeter University) and Dr Leo Bastos (Oswaldo Cruz Foundation, Brazil) resulted in the development of a statistical data model for real-time correction of delay in disease data.
The framework had the following properties:
- Computational feasibility for use in a warning system updated in real-time
- Full quantification of uncertainty in the correction
- Ability to capture: structured variability in the delay mechanism; temporal evolution of the disease (e.g., seasonal cycles); and unobserved effects (e.g. policy changes).
With the teams work, a new improved version of InfoDengue was launched in 2017 and is currently being used to produce warnings for Dengue, Chikungunya and Zika in 6 major Brazilian states (788 cities).
The vertical axis are weekly counts of dengue cases across the whole of Rio de Janeiro. The statistical model (green) corrects the delayed dengue counts (red) and how that compares to what actually happened (black).
The success that InfoDengue had on warnings for outbreaks led to the production of a second national warning system for Severe Acute Respiratory Infection (SARI). The national warning system InfoGripe is currently protecting more than 200 million people and is being used by state health authorities by mandate of the federal government. More recently, the system has also been used for capturing the true burden from the Covid (SARS-Cov-2) pandemic, although its efficacy has not yet been quantified.
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
Oswaldo Cruz Foundation (Fiocruz), Brazil
Meet our researchers
Dr Theo Economou
Senior Lecturer in Statistical Science
Theo is an applied statistician with experience in applying statistical models to solve problems in a variety of areas, including environmental sciences, hydroinformatics, epidemiology and public health.