The threat of emerging infectious diseases has stimulated the search for techniques to prevent and control communicable disease spread [1
]. Simulation models have emerged as key tools in examining trade-offs between multiple health interventions, and in aiding the control of communicable diseases [2
]. While properly parameterized and calibrated models can inform decision making, building such models is challenging because critical parameters are difficult to measure precisely, including the structure and dynamics of contact networks among population members, which shape the spread of both pathogens and risk behaviors.
Data collected by contact tracing [3
] and self-reporting [4
] has provided some important insights into network structure for many notifiable illnesses. Unfortunately, even for the best models, contact data depends heavily on unreliable self-reporting data collection methodologies [5
] which omit much detail and place a substantial recording burden on participants [4
]. Because of the less tangible character of contacts involved, determining contact network structure for air-borne pathogen spread requires collection of additional information on casual contacts [4
]. While self-reported measures can provide insight, some leading studies have noted the desirability of employing automated data-collection approaches to capture higher fidelity contact frequency and duration information [4
Some early work in the linking of health and micro-contact data has been reported [6
]. The work of [6
], with their single-day tracking of several hundred high-school students, forms a particularly important basis and methodological framework for the integration of micro-contact data and disease simulation models. However, in both these cases, limited health information was collected and a stylized infection model was utilized to extract the impact of contact network dynamics on the spread of infection.
As the first influenza pandemic in decades, the H1N1 pandemic – whose initial outbreak was described in April 2009 – served as a catalyst for research into control of emerging infectious diseases. Within the study site of Saskatoon (a Midwestern Canadian city of approximately 250,000 people) H1N1 first emerged in Spring 2009, and followed the typical summer quiescence, and Autumnal re-emergence. By mid-October, cases of H1N1 began a notable rise [8
]. At the same time, vaccination initiated in a staged fashion. Mass vaccination proceeded aggressively from early November through December 18. Vaccination data suggest that approximately 50% of the city population was vaccinated [9
]. Aided by the staged vaccination process, reported cases of influenza in 2009–2010 peaked unusually early (mid-November). Low numbers of influenza cases were reported in December 2009 and thereafter. Most circulating influenza transmission in Saskatchewan over this period was drawn from the H1N1 strain [8
In anticipation of the significance of the 2009–2010 influenza season, the co-authors had launched a previously-described [10
] pilot study in the City of Saskatoon to electronically collect contact patterns between 36 participants in addition to their influenza-related health status information. Each participant was requested to carry a proximity sensor at all times during the study period, as well as to fill out a sequence of weekly health surveys via a web browser. The study started on November 9th 2009 and finished on February 9th 2010, collecting all contacts between 36 individuals for 92
days. It recorded a total of approximately 265,000 thirty-second proximity time slots between individuals cross-linked to weekly self-reported health status and contact history.
In this work, we sought to integrate rich contact micro-data collected in [10
] with an adaptation of a well-grounded individual-level Canadian transmission model [11
], and population-level statistics on the infection rates for the province where the outbreak occurred in an agent-based model. Our study objectives were threefold: to assess the effectiveness of incorporating contact micro-data with models of infectious disease, to identify features within empirical contact patterns that exerted disproportionate impacted on infection spread, and to validate these findings against self-reported health status information.
] we opt for a smaller study population (36 as opposed to 788) but much longer duration (92
days as opposed to a single day) allowing us to evaluate the evolution of contact patterns and disease with the health state of the individuals throughout the flu season. Using collected health data, we can compare the results of our simulation to the infection rate in the province and amongst the participants. In this paper we make the following contributions:
1. A novel methodology for integrating disease, population level, and micro-contact data into a coherent agent-based simulation framework, validated by comparison with the actual health status of the study population;
2. A comparison of metrics for measuring the risk associated with contact and contact-duration, culminating in a novel measure: log time-degree (LTD);
3. A demonstration of the utility of micro-contact data during an epidemic outbreak based on both empirical and simulation results;
4. A preliminary investigation into the role of dynamic network structure on the spread of disease, and the impact of vaccination on that structure.