LEAPS: Modelling the decisions behind pandemic response

When the COVID-19 pandemic swept across the world, governments were forced to make decisions at extraordinary speed, often with incomplete information and under intense public and media pressure. Scientific collaboration accelerated at unprecedented speed, vaccines were developed in record time, and global attention focused on public health as never before. Yet the crisis also exposed gaps in coordination, communication and decision-making that complicated the response across many countries.

The LEAPS project – Learning from Epidemics and Pandemics to Strengthen Preparedness and Response Systems – was created to address precisely that challenge. Rather than developing vaccines or diagnostics, the project takes a step back to examine how societies respond to health crises, how decisions are made under pressure, and how lessons from past outbreaks can be captured and used to improve preparedness for the future.

At its core is a simple: the experience of COVID-19 should not fade into institutional memory, but instead be transformed into practical tools that help policymakers respond more effectively to future pandemics .

“We want to capture the learnings,” explains Nico Vandaele, of the Access-To-Medicines Research Centre at the KU Leuven and project co-coordinator. “And in those learnings, we have both the good and the bad. The good things we want to keep in policies for the future, but the bad things we find – things which went wrong or which could be improved – should be remembered and changed for next time.”

Learning challenges
For many observers, one of the most surprising aspects of the pandemic was how quickly previous experience seemed to fade from institutional awareness. Earlier outbreaks – from H1N1 influenza to Ebola – had already revealed weaknesses in global preparedness, yet many countries still struggled to coordinate their response when COVID-19 emerged. Part of the explanation for this delay lies in the distance between crises themselves.

“It was 100 years ago since we had the big pandemic,” says Catherine Decouttere, research manager at the Access-To-Medicines Research Centre of the KU Leuven and project co-coordinator. “That is too long in a human life. We work in generations of about 30 years, so it was simply too far away from people’s experience.”

The perception that pandemics were unlikely to affect Europe also played a role. According to Decouttere, many societies had grown accustomed to a sense of stability that made such threats seem remote. “I have a feeling that the European population was almost reckless,” she reflects. “There was a sense that nothing can really happen. It was only when the videos from Bergamo in Italy started circulating that people really woke up.”

At the policy level, knowledge about past outbreaks often existed but remained concentrated among a small group of experts. As time passes, that knowledge risks being lost as individuals move on or retire. “You had people who understood the situation very well,” Vandaele explains. “They knew what had happened and why certain decisions were taken. But the question is how you sustain that understanding for the future. You have to extract that knowledge from the experts’ mental models and capture it so that it can be used by others – and LEAPS aims to do precisely that.”

A systems perspective
What makes LEAPS different from many other pandemic-related research initiatives is its focus on systems rather than individual technologies or medical interventions. The project brings together experts in epidemiology, surveillance, public health, social sciences and systems engineering to examine how different elements of crisis response interact. Instead of looking only at how diseases spread biologically, the project explores how decisions, behaviour, communication and governance shape the trajectory of an outbreak.

“The whole idea is about connecting different scientific disciplines and making them available to policymakers,” Decouttere says. “We want to create transparency and include the learnings of the past in the way decisions are made.”

The central tool enabling this approach is systems modelling, a method that integrates multiple streams of information into a single analytical framework. These models combine epidemiological data with social and behavioural factors, allowing researchers to explore how decisions in one area can create ripple effects elsewhere. “If you push something in a system, it will bounce back,” Decouttere explains. “You might implement a measure because it looks correct from a biomedical perspective, but you also have to understand what it leads to – how people react, what pressures arise, and what unintended consequences appear. In that sense it may cause unexpected negative consequences, which are best known when interventions are designed.”

This perspective emerged partly from Decouttere’s own research experience. Her earlier work on immunisation systems in Rwanda revealed that technical solutions alone rarely explain gaps in vaccination coverage. “We initially thought the problem was the availability of vaccines in remote areas,” she says. “But it turned out that much of the issue was behavioural – people didn’t necessarily know when they needed repeated (booster) vaccination or why it mattered.”

That lesson proved equally relevant during COVID-19. Understanding how these dynamics played out in practice became a central part of the LEAPS research to understand how pandemic decisions unfold in practice. The LEAPS team has conducted extensive interviews and workshops with policymakers, health officials and scientists involved in the COVID-19 response. These discussions help reconstruct the sequence of decisions taken during the crisis, the information available at the time, and the pressures that shaped those decisions as they were taken.

“We talked to a lot of stakeholders in the countries involved,” Decouttere explains. “We wanted to understand how decisions were made, what information was used, and how those conversations unfolded.”

From this qualitative evidence, researchers construct causal loop diagrams that map the relationships between different factors in the system. These diagrams are then translated into quantitative simulation models capable of exploring how different decisions might affect the trajectory of a crisis. The project focuses on four countries – Belgium, France, Denmark and Greece – each of which experienced the pandemic in different ways. By comparing their responses, researchers can examine how national policies influenced outcomes.

“Neighbouring countries did not always respond in exactly the same way,” Decouttere notes. “That allows us to see the effect of different policies and how they influenced the effectiveness of measures.”

Once built, the models are validated by the same stakeholders who contributed to the research. “We bring the models back to them and ask: does this reflect what you experienced?” says Vandaele. “If the numbers come out and someone says, ‘no way, this never happened’, then we refine the model. We try to reach consensus about whether it captures reality.”

This is important because LEAPS is modelling not only disease dynamics, but decision dynamics. Its aim is to reflect how choices are actually made in crisis conditions, how they are shaped by timing, uncertainty, pressure and imperfect information, rather than how decisions might appear on paper after the event. Grounding the models in the experience of those directly involved helps ensure that they capture the real explicit and implicit mechanics of policymaking, making them more useful as tools for future preparedness.

From COVID-19 to the next crisis
Capturing the past, however, is only the first step. So, while the initial modelling focuses on COVID-19, the project is designed to prepare for future outbreaks. The real purpose of the models is to allow policymakers to explore how similar decisions might unfold under different circumstances – including

the next epidemic or pandemic, which may look very different from COVID-19.

To test the flexibility of their approach, the LEAPS team is extending the models to other disease scenarios, currently including influenza and West Nile virus. Each of these diseases introduces new complexities. Influenza represents a potential zoonotic threat, where viruses circulate between animals and humans, while West Nile virus highlights the role of environmental factors such as climate change through mosquito vector migration in shaping disease transmission.

“Disease X could be anything,” Decouttere says. “So we start with COVID because it is in our collective memory, but then we look at other types of diseases that might emerge in the future.”

The models can also incorporate new technologies that were not widely available during the early stages of the pandemic. For example, advances in genomic surveillance could allow scientists to detect new variants earlier, while targeted screening strategies could help contain outbreaks before they escalate.

“If we had detailed metagenomic surveillance across Europe, we might see variants emerging earlier,” Decouttere explains. “We cannot predict everything, but we could have a much closer view of what is happening.”

Understanding the human factor
One of the most distinctive features of the LEAPS approach is its emphasis on human behaviour. Pandemic responses are shaped not only by scientific evidence but also by political pressures, economic constraints and public trust. Ignoring these factors can lead to policies that look effective on paper but fail in practice. “Behavioural aspects can be part of the modelling,” Vandaele says. “It is not only technological or medical elements. Citizens, healthcare workers, scientists and policymakers all have behaviour that influences what happens.”

Vaccine hesitancy is a clear example of this dynamic. During COVID-19, public attitudes toward vaccination were influenced by communication strategies, social media narratives and perceptions of trust. “These mechanisms have many feedback loops,” Decouttere explains. “Trust, information, social media – they all interact. And if you ignore them, you miss a very important part of the system.”

By incorporating these behavioural dynamics into their models, LEAPS hopes to provide policymakers with a more realistic picture of how societies respond to interventions in any crisis. Measures such as vaccination campaigns, lockdowns or targeted screening do not operate in isolation. Their effectiveness depends on how people perceive risk, the trust they place in authorities and how they adjust their behaviour over time. Capturing these feedback loops allows the models to reflect not only the spread of a disease, but also how societal responses ultimately shape the course of an outbreak.

A tool for decision-makers
Ultimately, the goal of LEAPS is not to replace political decision-making but to support it. Scientists can provide evidence, models and scenarios, but final decisions remain the responsibility of governments. “What we do as scientists is propose possibilities,” Vandaele says. “We put information on the table. Policymakers then combine that with their own judgement and political considerations.”

The models offer a way to explore potential decisions before they are implemented in the real world. “Compared to real-life experiments, modelling is extremely cheap,” Vandaele notes. “You can test scenarios on your computer without causing damage in the real world.”

To make these tools accessible, the project is developing an online platform that will allow policymakers and practitioners to interact with the models through a user-friendly interface. “We want an interactive learning environment,” Decouttere says. “People can try different scenarios and see the consequences. That helps them understand the complexity of the system.”

Although health policy remains largely the responsibility of individual countries, pandemics do not respect national borders. One of the ambitions of LEAPS is, therefore, to explore how better coordination across Europe could strengthen crisis response. Decisions taken in one country – whether related to travel restrictions, testing strategies or vaccination policies – inevitably influence what happens elsewhere, particularly in a highly interconnected region such as Europe. By modelling these interactions, the project allows researchers and policymakers to explore how different national strategies may affect one another.

In some cases, stronger coordination, such as aligning certain measures or sharing critical data and resources, could improve overall outcomes. In others, countries may need to respond differently depending on their circumstances, healthcare capacity or stage of the outbreak. “Coordination does not mean that everyone must take exactly the same decision,” Vandaele explains. “But decisions should at least be aligned so that they do not work against each other.”

Understanding those interdependencies is one of the key objectives of the modelling work. By simulating different policy combinations across multiple countries, LEAPS aims to highlight where cooperation can reduce barriers and ultimately strengthen Europe’s collective resilience to future health crises.

This perspective reflects a broader shift in thinking about preparedness. Instead of focusing solely on individual threats, governments are increasingly considering how societies can respond to a wide range of emergencies, from pandemics to infrastructure disruptions. “The methodology can also be used for other crises,” Vandaele says. “It could be applied to situations like nuclear contamination or disruptions to water systems. The same systems approach can help decision-makers understand complex risks.”

Looking ahead
As the LEAPS project progresses, its immediate objective is to deliver a proof-of-concept platform that allows policymakers to explore pandemic scenarios and learn from past experiences. But its longer-term legacy may lie in changing how governments think about preparedness. By combining scientific evidence, behavioural insights and systems modelling, the project aims to make decision-making more transparent, more informed and more resilient to uncertainty.

“We want to bring science closer to decision-making,” Decouttere says. “And to make sure that what we know – the lessons from the past – is visible and accessible to the people who have to act.”

Whether the next global outbreak arrives in five years or twenty, the hope is that societies will be better equipped to respond. As Vandaele puts it, the goal is simple: “If we fill a portion of the gap that we observed during COVID, then we are adding value for the future.”