social
network properties related to transmission
number of components - reachability
component = set of persons
who are all connected directly or indirectly
-
more/smaller components prohibit
transmission; linking components facilitate it
-
Colorado Springs study -
persons at high risk for HIV (Potterat et al.)
-
no significant HIV epidemic
in CO Springs, despite large amount of risky behavior and large number
of partners
-
infected persons in small
components
density - # observed ties/#
possible ties
-
higher density - greater
risk for epidemic transmission
-
heterosexual sexual networks
less efficient for transmission than homosexual/bisexual sexual networks
or injection networks, independent of mode of transmission
centrality
-
degree - how many ties to
others
-
betweenness - how many pairs
of others a person connects, directly or undirectly
-
closeness - how many
steps away from others
-
high centrality of each type
puts one at greater risk for infection
-
prostitutes = high centrality;
intersections/distribution points for transmission
geographic aspect of networks
-
regions of network often
correspond to geographic locations/areas
-
mid 1990s Baltimore syphilis
epidemic
-
prior to epidemic, high prevalence
areas = a few poor inner city neighborhoods
-
public and private housing
in inner city condemned, residents to different parts of city
-
formed new partnerships in
new areas, exporting syphilis to different parts of city
-
before relocation, most partnerships
within inner city; cyclical pattern of (re)infections
-
urban renewal, etc. can splinter
networks and promote epidemics
concurrency - whether partnerships
overlap in time
-
serial monogamy vs. multiple
ongoing, concurrent partnerships
-
concurrency increases density
-
sexual network structures
in some African societies tend toward concurrency
-
educational campaigns: "Love
faithfully," "One partner at a time"
-
to work, must hold to strict
serial monogamy -- no recurrent, cyclical partners