Towards a participatory European social contact observatory

To understand how a virus spreads through a population, scientists must first understand how people move through their lives. It is not enough to know that a pathogen exists; researchers need to map the invisible web of human interaction that carries it from one person to another. These interactions, often called social contacts, are the moments when people stand close enough to share the air they breathe, whether in a crowded office, a classroom, or a family home. For decades, the best way to map this web was to ask thousands of people to keep detailed diaries of every person they met, recording the time, place, and age of each encounter. While these surveys provided a solid foundation for predicting outbreaks, they were expensive, difficult to organize, and usually happened only once or twice a decade. This meant that when a new crisis struck, such as a pandemic, the data often felt outdated, failing to capture how people quickly changed their behavior to stay safe.

A team of researchers from across Europe has now tested a different approach, one that turns the daily act of reporting illness into a way of tracking social life. By tapping into an existing network of volunteers who already report their flu-like symptoms online, the scientists asked these same people to also record their daily contacts. The goal was to see if this continuous, low-cost method could produce reliable maps of human interaction that are as accurate as the old, heavy-duty surveys, but with the added ability to update in real time. The study, which analyzed over 110,000 contact reports from more than 18,000 volunteers in Italy, Belgium, and the Netherlands, suggests that this method works. It found that the patterns of who meets whom in these digital reports closely match the patterns seen in traditional studies, while also revealing how factors like having a fever or being on holiday change the way people interact.

The researchers built their study on the InfluenzaNet, a long-standing digital platform where citizens in twelve European countries voluntarily report their health status. For years, these volunteers have been sending weekly updates about symptoms like coughs and fevers, helping public health officials track the spread of respiratory illnesses. The team realized that this same group of engaged citizens could be asked to do one more thing: keep a simple diary of who they met the day before. Starting in late 2023 and continuing through 2024, volunteers in three countries were invited to fill out a contact survey every two months. They reported the number of people they had spoken to or been in close proximity with, breaking these down by age group and setting, such as home, work, or school. Because these volunteers were already part of the system, the researchers could link their contact reports directly to their symptom histories and personal details, creating a rich picture of how health and behavior are intertwined.

When the team compared the contact maps generated by these volunteers with the gold-standard surveys from the past, the results were reassuring. The new data showed the same fundamental structures that scientists have long observed: people tend to mix most frequently with others of their own age, and the patterns of interaction in schools, workplaces, and homes looked familiar. The digital volunteers, much like those in the old paper-based studies, showed that children and young adults have the most contacts, while older adults have fewer. The study also confirmed that these patterns shift predictably with the calendar; people reported significantly fewer contacts during holidays and weekends, and more during the workweek. This alignment with established data suggests that the digital method captures the true shape of social life, even though it relies on volunteers rather than a random sample of the entire population.

Beyond confirming that the method works, the study uncovered how specific health conditions alter social behavior. By linking the contact diaries to the volunteers' symptom reports, the researchers could see what happens when someone feels sick. They found that people who reported symptoms of influenza, particularly those with a fever, tended to have fewer contacts than those who felt well. This reduction in social mixing was most pronounced in Belgium and the Netherlands, where individuals with fever reported significantly fewer interactions. However, the relationship was not always straightforward. In some cases, people with milder symptoms that did not qualify as full-blown flu still reported high numbers of contacts, suggesting that mild illness does not always stop people from going about their daily lives. This nuance is something that traditional, one-off surveys often miss, as they cannot easily track the same person's health and behavior over time.

The study also highlighted the power of this approach to track changes as they happen. Because the data is collected continuously, the researchers could watch how the "transmission potential" of the population shifted over time. They observed that the density of social contacts fluctuated, rising and falling in ways that reflected the changing seasons and public health conditions. This ability to monitor the social landscape in near real-time offers a new tool for public health officials. Instead of relying on a static map of how people interact, which might be years old, they could potentially access a living stream of data that shows how behavior is changing right now. This could help predict how fast a virus might spread during a new outbreak or how effective a new policy, like a holiday lockdown, might be at slowing it down.

While the study points to a promising future for digital surveillance, the authors are careful to note its limitations. The volunteers who participate in these online networks are not a perfect mirror of the general population; they tend to be more health-conscious and better educated than the average person. The researchers used statistical adjustments to correct for these differences, but the data still reflects the habits of a specific group of engaged citizens rather than a random cross-section of society. Additionally, the study relied on people remembering and reporting their own interactions, which can be imperfect. Despite these caveats, the findings suggest that digital participatory surveillance is a viable way to gather high-quality data on social contacts. It offers a way to keep the maps of human interaction up to date, providing a clearer, more timely view of the social currents that drive the spread of disease.