EIOS Collaboration

Collaboration

Collaboration

The Epidemic Intelligence from Open Sources (EIOS) initiative is more than just the system; it is a global effort to transform public health intelligence (PHI) by connecting open-source information, technology, and people. As a WHO-led initiative, EIOS sits at the center of a collaborative ecosystem, enabling the timely sharing of knowledge and insights to prevent, prepare for, respond to, and recover from health emergencies. By integrating advanced information technology tools, artificial intelligence, and public health expertise across geographies, domains and sectors, EIOS strengthens collaborative surveillance and enables early detection and risk assessment of health threats, embodying a One Health, all-hazards approach.

To achieve its vision, the EIOS initiative bridges collaborative innovation in technology with capacity building through PHI Training. Our Technology collaboration projects drive advancements in innovative tools within and around the EIOS system, while our PHI Training collaboration projects focus on building the global workforce needed to use these solutions effectively. Together, these components form the foundation for enhanced public health intelligence, where cutting-edge technology meets a well-trained global network of experts.

Technology

The EIOS system processes and categorizes information from tens of thousands of publicly accessible online sources, offering a comprehensive interface for PHI analysts to collect, analyze, and act on critical information. The system continuously processes hundreds of thousands of articles daily, categorizing and enriching content across hundreds of themes and topics while extracting key information such as locations and entities for analysis. During high-volume events, such as the COVID-19 pandemic, the system faced a significant surge in demand, with throughput increasing tenfold in the early days of the crisis. At this scale, quickly identifying trends, emerging topics, and actionable insights is challenging, particularly in the context of mis- and disinformation. To ensure the system remains scalable and effective during future crises, several collaborative initiatives aim to enhance its ability to handle increasing data volumes while maintaining high performance and enhanced support for PHI analysts.

The core of the EIOS system is the product of a longstanding collaboration between WHO and the European Commission's Joint Research Centre (JRC). To enable JRC to focus on advanced research, WHO has partnered with Adappt.ai to gradually assume the system's maintenance, operations, and further development. This collaboration aims at ensuring seamless integration of new solutions developed by other EIOS partners.

WHO and JRC are both receiving generous support from the Health Emergency Preparedness and Response Authority (HERA) of the European Commission.

EIOS System v.2

The EIOS system is provided by WHO as a service to an increasing number of users and requires further enhancement to cope with the additional usage and additional requirements of a growing user base. The development of a new version, “EIOS V2,” is currently ongoing, with new features and functionalities based on user feedback. The initial release will focus on a new multilingual user interface with a series of new features and functionalities including shared communications across communities, translated UI, and other requested improvements. A series of advanced AI powered analysis capabilities will be released in future iterations.

Collaborators: JRC, Adappt

Show less Show more

Abstractive Summarization

WHO will improve the efficiency and effectiveness of public health intelligence by developing, implementing, and continuously improving performance of advanced analytics algorithms, including clustering and abstractive summarization. This includes evaluating the performance and potential benefits of new language models and summarization algorithms, evaluating different approaches for multilingual summarization. This will result in (i) an improved algorithm and software component for abstractive summarization, (ii) enabling the function of abstractive summarization of clusters of news articles and other text items vs. only single articles, and (iii) abstractive summarization in at least one additional language.

Collaborators: WHO IT Teams, WHO Project Management Office

Show less Show more

Anomaly Detection

The goal of this collaboration is to develop an automated alerting mechanism that can identify unusual or unexpected events, and alert users to relevant changes. Anomalies may manifest themselves in different ways, for example changes in known distributions, or previously unknown relationships between data elements. Various detection algorithms and approaches are currently being explored and combined including the use of large language models.

Collaborators: Adappt

Show less Show more

Enhanced Web Scraping

A collaboration is now underway to enhance the web scraping functionality of EIOS. The development will expand the number of different data formats that can be scraped, including images and pdf documents, ease the configuration and set up of new sources, enhance the maintainability and source management. The new web scraping platform allows also to extract indicators like number of suspected or confirmed cases or number of vaccinations carried out. This will facilitate monitoring of key indicators which are published in different formats, for example during large scale events. A rich set of configurations, scraping and analysis techniques are available ranging use of DOM (Document Object Model) and Natural Language Processing to the use of large language models.

Collaborators: Adappt

Show less Show more

Semantic Search

One of the cornerstones of the automated analysis in the EIOS system is the identification of topics of interest known as categories in the EIOS system. This is currently achieved through Natural Language Processing based on lexical rules defined in multiple languages. Going forward this will be complemented with semantic search capabilities for discovering not only individual concepts represented by their keywords but also the contextual information to identify the “meaning” of the identified concepts more precisely. Given the large body of text of publicly available information this will be achieved using a combination of vector embeddings to focus on most relevant information combined with Large Language Models allowing the user to interact and fine tune filter and search criteria using natural language prompts.

Collaborators: Adappt

Show less Show more

Graph-Powered Category Editor

EIOS depends on highly effective content tagging with ontologies which are referenced from and used by lexical rules and leverage a significant Graph Database. With thousands of epidemiological and hazard-related base terms, and in many languages, the full potential of the graph can extend to billions of terms. This effort is intended to develop a solution that provides a multilingual approach for graph management, with vector embedding model for checking of terms, as well as for the addition and suggestion of missing terms.

Collaborators: Adappt

Show less Show more

Recommender System and Collective Intelligence

This work package aims at data analysis and modelling actions users are carrying out while using the EIOS system to develop a measure of information “interestingness” at global and community/team level. AI and ML algorithms are being developed and tested allowing specific relevancy scoring of information derived from system (patterns of behavior) and user provided feedback and preferences. The resulting scores will be integrated in the EIOS portal allowing the user to filter and search for recommended information tailored to their needs, benefiting from the work carried out by users across all communities.

Collaborators: JRC

Show less Show more

Speech-to-Text Sources from Radio

While “print” media on the Web is a significant source of information for the EIOS initiative, there is an identified gap associated with a variety of issues, including internet penetration. Radio is able to fill this gap, both for media and non-media sources. The addition of a speech-to-text component in the EIOS system will increase the reach of the system and enhance the ability to detect signals in vulnerable areas that are not otherwise captured. WHO is working with the UN Office for Information and Communications Technology to adapt the Unite Wave platform for public health intelligence, allowing to capture, identify and transcribe radio broadcasts for the detection of health threats

Collaborators: UN Office for Information and Communications Technology (UN OICT)

Show less Show more

EIOS Noise Reduction

A supervised learning approach is being developed to identify and de-emphasize text items that are not relevant for a specific user, so called "noise". Based on examples annotated by public health analysts, algorithms will compute a relevance score. Improvements to the system's user interface will allow users to filter their feed based on the relevance score. This will lead to more efficient monitoring of text items.

Collaborators: Robert Koch Institute (RKI)

Show less Show more

Other Projects

Our many partners and collaborators are constantly working on piloting new projects, such as a PHI Knowledge Graph, News Article Credibility Detection, Misinformation Classification Systems, among others. Please check back on this page in the near future for more information on these projects.

Show less Show more

Training

Effective public health intelligence depends on a skilled and adaptable workforce capable of transforming and interpreting vast amounts of raw data into actionable insights. Training and capacity-building initiatives are essential to empower professionals with the competencies needed to respond to health threats in real time. The EIOS initiative recognizes the critical role of training in fostering a unified global approach to public health intelligence. Through collaborations with partners such as the WHO Academy, the Robert Koch Institute, the US Centers for Disease Control and Prevention , and others, we develop and deliver tailored training programs, competency frameworks, and innovative learning approaches. These efforts ensure that public health professionals are equipped to navigate evolving challenges with confidence, promoting data-driven decisions and timely action across a One Health, all-hazards framework.

PHI Capacity Building Collaborations

In collaboration with the Robert Koch Institute, we are working to strengthen public health intelligence capacity by designing, developing, and piloting a comprehensive training program on the Foundations of Public Health Intelligence. This training builds on the expertise of the EIOS training program, providing opportunities to gain skills on PHI and on-the-job learning. The objective is to support countries in conducting PHI activities and enhancing the early detection and response to public health threats.

Collaborators: Robert Koch Institute (RKI), WHO Regional Office for the Eastern Mediterranean (EMRO)

Show less Show more

PHI Training Working Group

The Working Group on Public Health Intelligence (PHI) Training, established by the EIOS Core Team at the EIOS Coordination Group Meeting in March 2023, brings together experts from leading institutes from all over the world. This group supports the development and strengthening of PHI competencies and training beyond EIOS by providing strategic recommendations, advising on priorities, and fostering collaboration among stakeholders. Key activities include reviewing and revising an environmental scan of PHI activities to identify country-specific training needs, developing and evaluating training materials, sharing lessons learned, and promoting the initiative to expand the network of stakeholders. These efforts aim to ensure the ongoing success of PHI training programs and support the broader vision of enhancing global public health intelligence capabilities.

Collaborators: US Centers for Disease Control and Prevention (CDC), Ministry of Health Brazil, Robert Koch Institute (RKI), Africa Center for Disease Control (Africa CDC), European Center for Disease Control (ECDC), Food and Agriculture Organization (FAO), WHO Public Health Intelligence Unit

Show less Show more

Respiratory Pathogen Surveillance Collaboration

The EIOS Initiative, in collaboration with the US CDC and the WHO Global Influenza Programme, works to enhance Public Health Intelligence (PHI) capacities for detecting and responding to influenza and other respiratory pathogens with epidemic and pandemic potential. Through pilot activities in selected countries across the WHO African and South-East Asia regions, the initiative aims to implement a coordinated approach that integrates different sources of information. This work involved conducting landscape analyses to map existing surveillance mechanisms, assessing gaps, and identifying best practices. Building on these findings, the initiative supported the development of national workflows, standard operating procedures (SOPs), and enhanced systems for timely information exchange between traditional and non-traditional surveillance networks. By leveraging the use of the EIOS System at national level and extensive SOPs, this work aims to achieve integration among diverse information sources. By fostering early detection, comprehensive risk assessment, and effective public health responses, this initiative strengthens the foundations of global preparedness for respiratory disease outbreaks.

Collaborators: Centers for Disease Control and Prevention (CDC), WHO Global Influenza Programme (GIP)

Show less Show more

WHO Academy Collaboration for Capacity Building

The Intelligence Innovation and Integration Unit collaborates closely with the WHO Academy to enhance the capabilities of the global public health workforce in pandemic and epidemic preparedness and response. This partnership focuses on developing high-quality, tailored training programs, fostering innovation in lifelong learning, and addressing critical capacity gaps in public health intelligence and surveillance. In context of the EIOS initiatives and linking these to broader PHI training, key initiatives include the creation and dissemination of EIOS-related courses, competency frameworks, and curricular guidance, as well as exploring advanced analytics and simulation-based learning approaches. By aligning expertise and leveraging shared infrastructure, the collaboration aims to strengthen public health decision-making and foster a future-ready surveillance workforce capable of navigating evolving threats and innovations – both within and beyond the EIOS initiative.

Collaborators: WHO Academy

Show less Show more

Building and Strengthening Public Health Intelligence Competencies

Recent emergencies highlighted the catastrophic impact that a public health emergency can have at all levels, from sub-national to global. Responding effectively to a pandemic requires public health professionals that can provide up-to-the-minute information to officials and the public at large, prompting effective, confident, and data-driven decisions and actions. As such, there is a critical need for public health personnel who are able to transform an abundance of raw data from a multitude of sources into useful information that can be communicated in a timely manner to the appropriate Member State or organization to support a prompt response. Developing and sustaining Public Health Intelligence (PHI) capacity requires people with these professional competencies. This work built on the gaps identified through EIOS System Trainings among others. PHI workforce development resources are being integrated within the EIOS Training curriculum. As the Public Health Intelligence community grows and evolves, training and capacity building for a unified all-hazards, One Health approach are essential.

Collaborators: Centers for Disease Control and Prevention (CDC), Training Programs in Epidemiology and Public Health Interventions Network (TEPHINET), Robert Koch Institute (RKI)

Show less Show more