Biomedical Engineering Reference
In-Depth Information
Figure 6
.
Dynamic host resistance
. Variation in individual disease resistance is modeled by varying
the probability of infection given exposure. Compared to a population with no resistance (
A
), a con-
stant resistance of .5 (
B
) (50% probability of infection given exposure) substantially reduces disease
propagation. Populations with interindividual variation in resistance enjoy even greater protection
(
C
), despite an equivalent population-wide average. In
C
, resistance is randomly realized on the
uniform interval 0-1, for a mean .5. Under these conditions, illness-reactive link dynamics can be
highly decisive as in (
D
), where uninfected individuals can also evade one visibly sick contact per
unit time. However, when host resistance depends in part on the number of social contacts realized
(
E
), protective effects of reducing exposure can be offset by increased individual vulnerability to
infection via remaining contacts (who may be infectious but not visibly sick).
diminishes in people with little social contact (9,10). Any attempt to decrease
disease exposure by reducing social interaction may thus be offset by increased
vulnerability to infection. In Figure 6E, healthy individuals withdraw from those
who are visibly sick just as in 6D, but they suffer a fractional reduction in dis-
ease resistance as a result. This dynamic creates the opportunity for explosive
acceleration of an epidemic, particularly in clustered populations where much of
one's social network may fall ill simultaneously. Total contact levels can be
maintained by redeploying links to new partners, but this increases the real con-
nectivity of the network and is counterproductive when the redeploying agents
are asymptomatically infected. The reemergence of explosive epidemic dynam-
ics in Figure 6E underscores how seemingly subtle aspects of social behavior
can have highly amplified impacts on epidemic behavior in the context of dy-
namic host networks.
3.2.3. Transmission of Resistance in Multilevel Networks
In addition to altering social contact with those already ill, host networks
also respond to disease by developing preventive interventions (e.g., safe sex,
antibiotics, and vaccines). However, the networks distributing such interventions
are often structured differently from those distributing disease. For example, the
socioeconomic network that controls access to antiretroviral medications is quite
incongruent with the network that currently transmits HIV. Such misalignments
can be analyzed by superimposing a second "intervention network" upon a
