Artificial intelligence as a tool for enhancing the performance of public health emergency operation centres (EOC)

Defined by the WHO, an emergency operation centre (EOC) is a designated physical space centralising the coordination of emergency response activities.1 Modern EOCs extend their utility to preparedness activities and functions as early warning platforms, monitoring potential emergency indicators in a ‘watch’ mode. Beyond surveillance, EOCs facilitate the on-ground coordination of response actions and the development of information necessary for decision-making. Covering preparedness, watch, response and occasionally recovery, EOCs’ critical functions highlight the importance of continuous operation in mixed modes, avoiding traditional siloed management approaches. This contemporary perspective underscores the necessity for EOCs to remain operational permanently, adapting to the complexities of modern emergency management.2

In the dynamic environment of public health emergency management, the integration and analysis of various categories of information are paramount. This process is meticulously structured around different phases, with each phase demanding specific data inputs that range from epidemiological trends to strategic decisions and policy formulations. At the core of this information ecosystem are baseline datasets, including geographical, demographic and geopolitical data, along with a comprehensive list of hazards. These foundational datasets serve as a reference point, allowing for a nuanced understanding of the emergency context and informing targeted responses.

During the preparedness phase, information flow is constant and categorised based on the frequency of updates. Daily and weekly updates encompass a wide array of critical data points. Epidemiological data, offering insights into disease surveillance trends such as new cases, hospitalisation, recoveries and deaths, are crucial for monitoring the health landscape. Parallelly, resource inventories keep a tab on the availability of medical supplies, personal protective equipment (PPE), medications and vaccines, ensuring that the response mechanisms are well equipped.

Healthcare capacity metrics, including bed occupancy rates, the availability of isolation units and ICU capacity, provide a snapshot of the system’s readiness to handle surges in cases. Laboratory testing data, detailing testing capacities, turnaround times and rates of positive tests, further enhance the operational readiness by ensuring timely diagnosis and intervention. The availability and distribution of healthcare workers and volunteers are tracked rigorously, alongside the schedules and completion status of emergency preparedness drills and training sessions, ensuring that human resources are primed for emergency response.

On a monthly or as-needed basis, the focus shifts towards more strategic and infrastructural readiness aspects. Risk assessments are updated to reflect changes in vulnerability, risk profiles and potential impacts, providing a basis for informed decision-making.3 Infrastructure readiness encompasses the operational status of emergency response facilities, communication systems and transportation networks, ensuring that physical and technological infrastructures can support response efforts. Stakeholder coordination, documented through records of meetings, agreements and mutual aid compacts, emphasises the collaborative nature of emergency management. Public communication plans, detailing strategies for risk communication, public engagement and educational material readiness, are crucial for maintaining public trust and compliance.

Annually or on an ad hoc basis, the emphasis is on policy and planning documents, budget and funding, and training and development programmes. These categories encompass emergency operation plans, legal documents, mutual aid agreements, financial readiness and long-term training schedules, ensuring a holistic preparedness framework that is both resilient and adaptive. As the situation escalates into the response phase, real-time or continuous information becomes the lifeline of operations. Situation reports provide ongoing updates on the status of the emergency, guiding the tactical aspects of the response. Real-time tracking of resource distribution and logistic support, along with immediate casualty reports, ensures that resources are optimally allocated and that the health status of affected individuals is promptly addressed.

Daily and weekly updates continue to play a critical role, with a focus on emergency communications, health system status, volunteer and workforce management, epidemiological updates, health system status, public health interventions, and supply chain status. This granularity of information allows for a responsive and flexible management strategy that can adapt to evolving circumstances. As events unfold, media monitoring, strategic decisions and after-action reviews become integral, offering insights into public perception, the effectiveness of response strategies and lessons learnt for future improvement. Through this comprehensive approach to information management, public health EOC (PHEOC) can navigate the complexities of emergency response, from the preparedness phase through to response and recovery, ensuring that decisions are data-driven, timely and effective. This narrative underscores the significance of a structured information framework that not only facilitates the coordination of response efforts but also enhances the decision-making process, ultimately leading to a more resilient public health system.

The PHEOC must ensure that the flow of information is accurate, timely and secure, providing an evidence-based foundation for operational decisions. Proper categorisation and prioritisation of information are critical for an effective response to any public health emergency.

Artificial intelligence (AI) stands as a transformative force in the operational effectiveness of a PHEOC throughout the public health emergency management cycle, which covers the preparedness, response and recovery phases. In the preparedness phase, AI delves into historical data, employing predictive analytics to foresee potential emergency scenarios and disease outbreaks, thereby strengthening the PHEOC’s state of readiness. It harnesses simulation and optimisation techniques for modelling resource allocation, ensuring that resources and personnel are distributed efficiently in anticipation of potential crises. Moreover, AI customises training for healthcare workers by identifying predictive skill shortages and adapting to individual learning needs, optimising preparedness at a personal level. As an emergency unfolds, AI’s role becomes even more critical. It processes streams of real-time data, from epidemiological intelligence to logistics, offering instant, actionable insights. Its ability to recognise patterns in disease transmission aids in pinpointing hotspots, which is crucial for directing targeted intervention.

AI also supports decision-makers by integrating diverse information streams, enhancing situational awareness and suggesting evidence-based actions for prompt response. In the aftermath of an emergency, AI swiftly assesses the impact on public health and the healthcare infrastructure, guiding the recovery phase with data-driven assessments. Predicting long-term trends, AI facilitates the crafting of recovery plans that ensure efficient utilisation of resources and expedite the restoration of healthcare services. Moreover, by examining data from the entire disaster management process, AI distils insights for future improvement, thereby feeding into a continuous cycle of learning and preparedness enhancement. Beyond these phase-specific applications, AI acts as a central data management system that unifies and analyses information throughout the disaster management cycle. It automates routine reporting and generates alerts for critical situations, providing early warning and facilitating timely interventions. With each cycle, the AI system refines its predictive models and recommendations, honing its ability to inform decision-making.

The AI–human interface in a PHEOC is designed to enhance decision-making, data analysis and response strategies during public health emergencies. AI can analyse large datasets to predict outbreak trajectories or the impact of interventions, with results displayed in user-friendly formats like dashboards for quick understanding. It identifies trends, anomalies and provides evidence-based recommendations, helping prioritise resources and strategies. The interface also allows interaction with AI simulations for emergency planning and integrates AI into communication tools to ensure critical information reaches the right people. To maintain transparency and fairness, the interface should explain AI recommendations, monitor biases and allow for review of AI-generated content. Overall, integrating AI into the PHEOC interface empowers public health professionals by improving efficiency and decision-making.

In essence, AI transforms the vast volumes of data into strategic foresight and operational guidance, significantly bolstering the PHEOC’s mission to protect public health. Through AI, the disaster management cycle evolves into a dynamic, intelligent process, increasingly capable of mitigating risks and promoting public health resilience.

The holistic approach regarding the flow and time of information collected by the EOC can be presented in figure 1. The EOC serves as the nerve centre during public health emergencies, and its efficacy is heavily reliant on the systematic collection and processing of various types of information. This collection is not just an isolated activity but a web of interactions that integrate geographic and demographic data, public health statistics, social, economic and political data, as well as an inventory of potential hazards. In the preparedness phase, the EOC consolidates geographical and demographic data to understand the area and population at risk, crafting baseline frameworks for risk analysis.

The integration of AI into PHEOC operations represents a paradigm shift in public health emergency management. Through intelligent data processing and predictive analytics, AI enables a proactive, data-driven and strategic approach to disaster management (figure 1). It propels the cycle of preparedness, response and recovery towards increased efficiency, fostering a resilience that is adaptive to the needs of public health. AI’s transformative impact ensures that PHEOCs can effectively manage large volumes of data, translating complex information into strategic action and operational guidance.