Transcript ppt

Effect of Demographic and Spatial
Variability on Epidemics: A Comparison
between Beijing, Delhi, and Los Angeles
Jiangzhuo Chen
Joint work with Fei Huang, Maleq Khan, Madhav
Marathe, Paula Stretz, and Huadong Xia
NDSSL Technical Report 10-111
2010 International Conference on Critical
Infrastructures (CRIS2010)
September 21st, 2010
Network Dynamics & Simulation Science Laboratory
Our group members (NDSSL)
Work funded in part by NIGMS, NIH MIDAS program, CDC, Center of
Excellence in Medical Informatics, DTRA CNIMS, NSF, NeTs, NECO and OCI
(Peta-apps) program, VT Foundation.
Network Dynamics & Simulation Science Laboratory
Talk Outline
• Major contributions.
• Background:
– synthetic population and contact network;
– propagation of infectious disease on social contact
networks;
– public health interventions.
• Comparison study: Beijing, Delhi, Los Angeles.
– demographics of populations;
– structural properties of social networks;
– disease dynamics and intervention efficacy.
Network Dynamics & Simulation Science Laboratory
Our Contributions
A methodology for generating a coarse synthetic population and a
social contact network for any region in the world.
– From very limited census data and LandScan data.
– We applied the methodology to create Beijing and Delhi synthetic
populations and social networks.
A comparison study of three different urban regions: Beijing, Delhi,
and Los Angeles:
– demographic and spatial variations;
– structural properties of social contact networks;
– epidemic dynamics and public health intervention strategies.
Network Dynamics & Simulation Science Laboratory
Social Infrastructure
Synthetic Population & Contact Network
Data
synthetic
population
census
people
(demographics)
location
locations
activity survey
activities
Contact
Network
sublocation
model
Network Dynamics & Simulation Science Laboratory
weighted
edges
between
people
contacts with
durations
Disease Spread in a Social Network
• Within-host disease model: SEIR
• Between-host disease model:
– probabilistic transmissions along edges of
social contact network
– from infectious people to susceptible people
Network Dynamics & Simulation Science Laboratory
Public Health Interventions
• Pharmaceutical interventions: vaccination or
antiviral changes an individual’s role in the
transmission chain
– Lower susceptibility or infectiousness
• Non-pharmaceutical interventions: social
distancing measures change people activities and
hence the connectivity of social network
– Work closure, school closure, isolation, etc.
Network Dynamics & Simulation Science Laboratory
Social Infrastructure
Los Angeles
Data
US detailed
census
distributions
real locations
US activity
survey
Contact
Network
synthetic
population
people
(demographics)
locations
temlates
sublocation
model
activities
Network Dynamics & Simulation Science Laboratory
weighted
edges
between
people
contacts with
durations
Social Infrastructure
Beijing & Delhi
Data
distributions
limited census
data
LandScan
open map
US activity
survey
Contact
Network
synthetic
population
people
(demographics)
locations
temlates
sublocation
model
activities
Network Dynamics & Simulation Science Laboratory
weighted
edges
between
people
contacts with
durations
Generating Non-US Contact Networks
• Generate synthetic population
– create individuals with age and gender
• joint distribution from public census data
– create home, work, and school locations
•
•
•
•
type, size, coordinates
region boundary from map
location size from distributions (conditioning on location type)
location density based on LandScan data:
population density in each 30” x 30” grid cell of the whole world
– assign daily activities to each person
• start time, duration (in seconds), activity type
– for each activity, assign location and sublocation
• Generate social contact network
– edge between two people if they stay in same sublocation
simultaneously
Network Dynamics & Simulation Science Laboratory
Demographics
Region
Population size Avg. age
Avg. household size
Sex ratio (M/F)
Beijing
16 million
37.9
2.6
0.99
Delhi
13 million
25.6
9.1
1.22
Los Angeles
16 million
32.9
3.0
0.97
Network Dynamics & Simulation Science Laboratory
Structural Properties of Contact Networks
Network
No. of Nodes Avg. Deg Max Deg
Delhi
13 million
79.78
321
Beijing
16 million
66.77
313
Los Angeles
16 million
56.60
463
Network Dynamics & Simulation Science Laboratory
Disease Dynamics: Progression and Vulnerability
•Disease outbreaks earlier in Beijing, peaks higher in Delhi.
– denser population and stronger mixing
•Vulnerability has larger variation in Beijing population; is more
evenly distributed in Delhi population.
Network Dynamics & Simulation Science Laboratory
Epidemic Progression by Age Group: Beijing
Network Dynamics & Simulation Science Laboratory
Epidemic Progression by Age Group: Delhi
Network Dynamics & Simulation Science Laboratory
Epidemic Progression by Age Group: Los Angeles
Network Dynamics & Simulation Science Laboratory
Effectiveness of Interventions
Network Dynamics & Simulation Science Laboratory
Conclusion
• A model to generate synthetic social infrastructure for
studying epidemics and public health policies:
– works for any region of the world;
– flexible with data availability;
– richer data  more realistic results.
• Comparison study between Beijing, Delhi, and Los Angeles:
– generate Beijing and Delhi contact networks using our model;
– demographic and spatial differences affect contact network structure
– differences in network structure affect epidemic dynamics
– public health policies for preventing/containing pandemic should
adapt to differences across populations
Network Dynamics & Simulation Science Laboratory
Policy Implications
• Beijing needs more prompt public health interventions
– Disease outbreaks and peaks earlier in Beijing.
– Peak occurs within 3 months.
– School age people in Beijing are extremely vulnerable.
– If vaccines are not yet ready, it may be helpful to close schools.
– For normal seasonal flu, early vaccination is most effective.
• Vaccination works best in Los Angeles.
– Only 25% coverage is is enough to contain the epidemic.
– Disease outbreaks 3 months after index cases – more time to get vaccines
ready.
• Two options for Delhi: vaccination and school closure.
– About 2 month for preparing vaccines.
– Shorter epidemic duration means smaller social cost for school closure.
Network Dynamics & Simulation Science Laboratory
Thank You
Network Dynamics & Simulation Science Laboratory