How Our Social Connections Shape the Spread of Disease: What New Research Reveals

In today’s interconnected world, a local outbreak can become a global crisis with remarkable speed. Planes link distant cities in hours, social media connects millions in seconds, and our daily routines weave together countless invisible threads of contact. A recent scoping review led by WorldPop PhD researcher Zhifeng Cheng and published in Infectious Diseases of Poverty explores how these threads – our social ties, travel patterns, and physical interactions – shape not just the spread of disease, but also the behaviours that determine how communities respond.

The research highlights that we live within overlapping networks that constantly influence one another. Social networks, whether formed through friendships, workplaces, or online platforms, are powerful engines of behavioural contagion. People often make health decisions not in isolation, but in response to the habits and beliefs of those around them. Studies show that smoking, exercise, eating patterns, and risk perception can all ripple through networks in ways similar to infectious pathogens. The same networks that help good information spread can also amplify misinformation. During COVID-19, misleading narratives sometimes spread faster than many public health messages – fuelling vaccine hesitancy and undermining trust in science-based guidance.

While social networks shape perception and behaviour, mobility and contact networks determine the physical pathways of disease. Daily commutes move pathogens within towns, while long-distance travel connects distant communities. Air travel in particular acts as a bridge, linking outbreaks thousands of kilometres apart. But these networks are not uniform. A small number of individuals – whether because of their job, social habits, or biology – can infect far more people than average. These “super-spreaders” play an outsized role in shaping the trajectory of epidemics, making it even more important to understand how real-world interactions differ from the assumptions built into traditional epidemiological models.

For decades, modelling relied largely on top-down averages, treating people as if they mixed evenly within a population. But human behaviour is far more complex. To capture this complexity, researchers are turning to innovative modelling approaches that simulate epidemics from the ground up. One promising direction involves multiplex networks, which treat social influence, physical contact, and information flow as interconnected layers. A person who learns about a disease through their social network may change their mobility patterns or adopt protective behaviours, which in turn alters transmission dynamics. These feedback loops can help explain why two communities facing similar outbreaks may experience very different outcomes.

Even more ambitious is the emergence of generative agent-based models. Powered by advances in artificial intelligence, these models use autonomous agents that can perceive their environment, interpret information, and make decisions in ways that more closely resemble human reasoning. Instead of following fixed rules, these agents adapt as situations evolve, mirroring how real people change behaviour in response to new information, social pressure, or fear.

These tools offer exciting possibilities, but they also face challenges. Integrated models require detailed data that capture social relationships, movement patterns, and individual behaviour -data that are difficult to obtain ethically and consistently. Combining multiple data sources introduces new technical hurdles, and ensuring models reflect the experiences of diverse populations is essential for avoiding blind spots that could worsen inequalities. There is also a growing recognition that models must capture what happens after illness: how people reintegrate socially, whether they follow medical advice, and how stigma or support affects recovery.

Understanding these intertwined networks – social, spatial, and behavioural – brings us closer to predicting and managing outbreaks in a world where the next crisis may be only a flight, a message, or a moment away.