The Promise of AI-driven Early Warning Systems in Public Health

The resilience of health systems is being challenged across the world, as disasters triggered by climate change—flooding, wildfires, tropical disease outbreaks—increasingly threaten public health. Meanwhile, on the heels of COVID-19, pandemics continue to loom as an emergent threat, worsened by environmental degradation and globalised travel and supply chains.

At Trilateral’s Health cluster, we’re supporting the development of AI-powered Early Warning Systems (EWS), tools which identify patterns and threats in troves of data too large for humans to sift through, to enable public authorities to anticipate and proactively manage public health crises and infectious diseases.

The evolution of AI-driven EWS

AI-powered EWS fuses cutting-edge technology with the principles of crisis management. According to the United Nations Office for Disaster Risk Reduction (UNDRR), an effective EWS consists of four pillars: disaster risk knowledge, detection and monitoring of hazards, warning dissemination and communication, and preparedness to respond. Integrating AI into EWS achieves these aims by enabling authorities to gain insights from stores of data that are generally unused. These data-driven insights allow AI tools to create risk models that accurately predict the potential impact of disasters.

Machine learning algorithms can analyse historical data, climate patterns, and socio-economic factors to develop models capable of forecasting the intensity, duration, and geographical spread of hazards. This predictive modelling not only enhances the precision of early warnings but also aids in the strategic allocation of resources, ensuring that response efforts are targeted where they are most needed. Adopting these systems shifts crisis management from a reactive to a proactive model, enabling authorities to anticipate issues, manage resources, and coordinate responses across territories.

Case study: AI-driven EWS in the CLIMOS project

Trilateral’s work in the EU-funded CLIMOS project contributes to the development of an ethically sound and legally compliant EWS that uses AI data insights to anticipate public health threats related to Sand Fly Borne Diseases (SFBD), which are driven by climate change. In the CLIMOS context, an effective EWS must anticipate the likelihood of a surge in SFBD, pinpointing when and where it may occur and identifying the relevant authorities for notification. The precision of these predictions hinges on amalgamating real-time satellite data, which monitors climate impacts on the environment, with continuous ground sensor data.

The tool incorporates key information from the Intergovernmental Panel on Climate Change (IPCC) risk framework to make predictions over the long, medium, and short-term:

• hazard (the likelihood of adverse events),
• exposure (the human and animal populations in the area the hazard may occur), and
• vulnerability (the likelihood of exposed elements to suffer adverse effects)

The complexity of this task necessitates the use of AI to generate detailed, high-resolution land use maps for climate and geolocated information on sand fly populations. Here, Trilateral Research assumes a pivotal role, ensuring the EWS is developed in compliance with ethical and legal principles. This involves, for example, balancing the need for satellite surveillance while safeguarding the privacy of individuals in targeted areas and preventing adverse social and economic effects. Trilateral also endeavours to integrate these systems into the broader context of public healthcare management.

For all their promise, CLIMOS and similar EWS projects face significant hurdles related to data sharing and access. Health data is sensitive, revealing personal information such as disease history and genetic makeup. No single regulatory framework exists defining safe data sharing practices, making data controllers reluctant to share information for fear of violating ethical, privacy, and legal safeguards that may jeopardize patients’ rights.  Trilateral’s health cluster recently issued a series of recommendations to facilitate data sharing in healthcare technologies. You can read them, alongside a more detailed breakdown of data-related challenges, here.

The data held by hospitals, regional health authorities, and governments could transform collective responses to public health threats. Trilateral’s health cluster is contributing to the development of AI-powered EWS tools which can mine these datasets to facilitate proactive, coordinated responses to complex societal challenges. Along the way, we’re working to ensure innovative technology responds to society’s needs and concerns, producing outputs in compliance with the highest ethical, legal, and regulatory standards. To learn more about our work, visit https://innovation.trilateralresearch.com/.