Biomedical Engineering Reference
In-Depth Information
Figure 1
.
General architecture of the ActiveHost modeling system
. Each simulated epidemic is
based on a collection of hosts interacting with one another in ways that transmit disease. Agents
maintain a list of potential contacts that are realized according to probabilistic rules that constitute the
structure of the disease transmission network. Network dynamics occur as agents modulate contact
realization probabilities in response to simple, ecologically realistic rules, such as "interact with
friends more often than strangers," "avoid people who are sick," or "maintain a constant number of
social contacts." Transmitted pathogens provoke immune responses, replicate, and die with the host
after a defined period of infection if they cannot spread to a new host. Each simulated object main-
tains a set of relevant parameters (bulleted list items), and the overarching simulation object compiles
mortality trajectories, population-wide values of evolving parameters, and other outcomes over an
arbitrary number of stochastic repetitions.
ery). Hosts also have a natural lifespan, and the system includes a framework for
modeling evolutionary dynamics via age-dependent reproduction and noisy
transmission of physiologic and behavioral parameters from parent to child. The
aggregate collection of hosts is aged for a fixed number of time units, and the
primary outcomes monitored are the host population size, the pattern of realized
social linkages, and the frequency of various host states (sick, latently infected,
overtly infected, dead, immune, etc.).
3.
RESULTS
The present analyses focus on two ways in which real disease transmission
networks depart from the assumptions underlying conventional epidemiologic
