Intelligent approaches to improved disease prevention
Climate change and the global movement of people and goods have increased the risk of disease threats. The COVID-19 pandemic also highlighted the need for improved surveillance and epidemic intelligence systems for the early detection, monitoring and assessment of emerging infectious diseases. A key challenge is that current disease surveillance systems rely on routinely collected data, such as passive surveillance, to monitor disease occurrence and spread and to design appropriate responses. New approaches that incorporate artificial intelligence (AI), machine learning and the analysis of big data from a range of sources could help improve preparedness by better understanding the drivers of disease emergence and enabling the development of more accurate models.
Responding to infectious disease threats
A good example of this is the EU-funded MOOD(opens in new window) project. Through collecting and mining data using machine learning, this initiative has developed a unique digital platform designed to boost Europe’s capacity to detect and respond to infectious disease threats through a ‘One Health’ approach. “Our aim was to bring complex data science and modelling tools directly into the hands of those doing research and risk assessment for science-based public health decisions,” explains project coordinator Elena Arsevska from the French Agricultural Research Centre for International Development(opens in new window) (CIRAD) in France. MOOD employed mathematical, statistical and data science-powered methods in the platform’s infrastructure, and integrated machine learning algorithms to generate predictive risk maps for diseases. In addition, the consortium incorporated the PADI-web tool, which uses natural language processing to automatically scan, extract and analyse disease-related information from online media to support event-based surveillance.
Data science for predictive modelling
The end result is an open access platform(opens in new window) that brings together environmental, climatic, host distribution and disease data. The aim is to help public and animal health professionals make faster and better-informed decisions about emerging disease threats. For example, the platform currently offers disease risk mapping outputs for various pathogens, including the West Nile virus, tick-borne encephalitis in humans, avian influenza in birds, and antimicrobial resistance in food-producing animals. “We will add more diseases to the platform in the future, depending on European needs,” says Arsevska. Over 20 partner institutions across Europe and the United States worked closely to ensure the platform met real-world, i.e. end user, needs. “MOOD’s strength lies not just in the technology, but in its application,” notes Arsevska.
Real-world impact on public health strategies
The project carried out specific case studies to demonstrate the platform’s real-world impact on public health strategies. Each study incorporated data integration, predictive modelling and the ‘One Health’ perspective, reflecting the interdependence of human, animal and environmental health. These case studies addressed infectious diseases representing different transmission routes, temporal dynamics and geographical distribution across Europe in the context of climate change. For example, MOOD researchers showed that habitat, land-use and wildlife data can enhance local-level risk mapping in Europe for tick-borne encephalitis. In the case of the West Nile virus, the team combined indicator and event-based surveillance data to map the risk of disease occurrence and model the force of infection using longitudinal data from surveillance. The avian influenza case study integrated traditional outbreak reporting with genomic data to assess transmission risks from wild birds to domestic poultry. This was also used to map hotspots for disease occurrence. The case study on antimicrobial resistance revealed critical gaps in intersectoral data integration, demonstrating how resistant pathogens in livestock pose uneven risks across regions. The findings highlighted the urgent need for more coordinated ‘One Health’ surveillance strategies. During the COVID-19 pandemic, MOOD works helped assess virus spread, evaluate the impact of mobility restrictions(opens in new window) and inform interventions for public health agencies in Europe. “MOOD helped future preparedness by strengthening collaborative platforms, research consortia and modelling networks to foster data and code sharing and effective collaboration between academia, decision makers and data providers,” adds Arsevska. To build on these results and secure sustainability beyond the project’s lifetime, MOOD has established an international non-profit association to maintain and promote the platform, while forging new collaborations and exploring funding opportunities.
