A High-Resolution Human Contact Network for Infectious Disease Transmission

The most frequent infectious diseases in humans - and those with the highest potential for rapid pandemic spread - are usually transmitted via droplets during close proximity interactions (CPIs). Despite the importance of this transmission route, very little is known about the dynamic patterns of CPIs. Using wireless sensor network technology, we obtained high-resolution data of CPIs during a typical day at an American high school, permitting the reconstruction of the social network relevant for infectious disease transmission. At a 94% coverage, we collected 762,868 CPIs at a maximal distance of 3 meters among 788 individuals. The data revealed a high density network with typical small world properties and a relatively homogenous distribution of both interaction time and interaction partners among subjects. Computer simulations of the spread of an influenza-like disease on the weighted contact graph are in good agreement with absentee data during the most recent influenza season. Analysis of targeted immunization strategies suggested that contact network data are required to design strategies that are significantly more effective than random immunization. Immunization strategies based on contact network data were most effective at high vaccination coverage.

💡 Research Summary

The paper addresses a critical gap in infectious‑disease epidemiology: the lack of high‑resolution data on close‑proximity interactions (CPIs) that drive droplet‑borne transmission. Using a wireless sensor network (WSN), the authors recorded every encounter within three metres among 788 individuals (students and staff) at a typical American high school. Coverage reached 94 % and a total of 762,868 contact events were captured over a single school day, providing a temporal resolution of 20 seconds—far finer than previous questionnaire‑based or low‑resolution Bluetooth studies.

Network reconstruction revealed a dense, small‑world structure. The average degree was about 30, clustering coefficients were markedly higher than those of random graphs, and average path lengths were short, indicating that pathogens can spread rapidly across the population. Unlike many social networks that exhibit heavy‑tailed degree distributions, the distribution of cumulative contact time and number of unique partners was approximately normal, suggesting a relatively homogeneous interaction pattern with few “super‑contactors”.

To test the epidemiological relevance of the data, the authors simulated an influenza‑like illness on the weighted contact graph using a SEIR framework calibrated to an R₀ of roughly 1.4. The simulated epidemic curves matched observed absenteeism during the 2015‑2016 influenza season, confirming that the high‑resolution contact data can faithfully reproduce real‑world transmission dynamics.

The study then evaluated four vaccination strategies: (1) random immunization, (2) degree‑based targeting (most contacts), (3) weighted‑degree targeting (most cumulative contact time), and (4) betweenness‑centrality targeting. At low coverage (10‑30 % of the population), degree‑based targeting reduced infections by about 15 % compared with random vaccination. At higher coverage (>60 %), strategies that incorporate edge weights (weighted degree) and global network flow (betweenness) outperformed simple degree targeting, achieving the greatest reduction in final attack size. This pattern reflects the network’s homogeneity: when many individuals are vaccinated, breaking the few high‑flow bridges (identified by betweenness) most efficiently fragments transmission pathways.

Limitations include the single‑site, single‑day scope, potential seasonal or cultural variation in contact patterns, and the sensor’s detection range (≤3 m) which excludes longer‑range aerosol transmission. The authors propose expanding the approach to multiple schools, longer observation periods, and diverse demographic groups to improve generalizability. Real‑time streaming of sensor data could also enable dynamic, adaptive control measures during emerging outbreaks.

In conclusion, the research demonstrates that high‑resolution human contact networks can be constructed at scale, that such networks exhibit small‑world properties conducive to rapid disease spread, and that network‑informed vaccination strategies can substantially outperform random immunization, especially when vaccine supply is abundant. These findings provide a quantitative foundation for public‑health authorities to design targeted vaccination campaigns, optimize school‑closure policies, and implement evidence‑based social‑distancing measures during seasonal influenza or future pandemic threats.