Model Results

Download Report

Transcript Model Results

How can Modeling Help
in Emerging Epidemics?
John Grefenstette, PhD
Public Health Dynamics Lab
Health Policy & Management
Pitt Public Health
Dec 5, 2014
Modeling Ebola
Outline




Current status
Near term projections
Potential for spread outside Africa
Perspectives
Ebola Outbreak
WHO Situation Report (12/3/14)
•
17,145 reported cases of
Ebola virus disease (EVD)
•
6070 reported deaths
•
Case fatality rate across the
three most-affected countries
in all cases with a recorded
definitive outcome is 72%
•
Case fatality rate is 60% in
hospitalized patients
•
Case incidence is slightly
increasing in Guinea, stable or
declining in Liberia, and may
still be increasing in Sierra
Leone.
http://www.who.int/csr/disease/ebola/situation-reports
http://www.cdc.gov/vhf/ebola/outbreaks/2014-west-africa/casecounts.html
Models of Epidemics
 Key parameter: R0 = the expected number of secondary cases infected by
an index case in an entirely susceptible population




Estimate R0 from doubling rate at start of outbreak
Use mathematical models to compute expected incidence
Identify possible interventions (eg isolation, contact tracing, safe burial)
Use models to explore alternative interventions to estimate costs and
benefits
JH Jones, Stanford Univ, 2007
Comparison with other diseases
Infection
R0
Ref
Ebola (2014)
1.71 – 2.02
Influenza (1918)
1.8
Mills et al., 2004
Measles
18.8
Anderson & May, 1991
Pertussis
3.8 – 5.6
Anderson & May, 1991
SARS
2.7 – 3.6
Wallinga & Teunis, 2005
Smallpox
4 – 10
Anderson & May, 1991
WHO Ebola Resp Team, 2014
Important notes:

Transmission of Ebola requires close contact with infected bodily fluids.
Uncertainties about current Ebola data:

Undetected / unreported cases

Asymptomatic infections

Level of prior immunity in the population

Cross-border mobility patterns
Early Projections
Projections made based on observation through 8/26/14
Liberia: With No Intervention
Liberia: With Intervention
 Used sparse data available at the time
 Criticized for being too alarmist (Nature, Nov 6)
 "As these measures [isolation, safe burial practices] are rapidly
implemented and sustained, the higher projections presented in this
report become very unlikely."
Meltzer et al. Centers for Disease Control and Prevention (CDC). Estimating the future number of cases in the Ebola epidemic--Liberia and Sierra
Leone, 2014-2015. MMWR Surveill Summ. 2014 Sep 26;63 Suppl 3:1-14. PubMed PMID: 25254986.
Model of Ebola Transmission
Susceptible
Exposed
Infectious
Hospitalized
Dead but
not buried
Fig 5: Gomes et al. PLOS Currents Outbreaks. 2014 Sep 2.
No longer infecting
Based on its models, the CDC projects that stopping the epidemic
requires isolating up to 70% of patients in treatment centers or other
settings that reduce transmission, assuming that burials are handled
safely.
Human Mobility Models
The rapid spread of the virus may
be driven by local and regional
travel.
Epidemiological models of the
spatial spread of Ebola rely on
estimates of the volumes and
flows of traffic between
populations
Models can estimate human
mobility patterns from mobil call
data records.
Mobility can be a target of control
measures.
Wesolowski et al. PLOS Currents Outbreaks. 2014 Sep 29
Potential for International Spread
Modeling Methods
• Analyze air flight
networks, passenger
flows and
destinations
• Evaluate risks to
destination countries
Figure 2 Final traveller destinations, passenger volumes * and scheduled non-stop flights†
departing Guinea, Liberia, and Sierra Leone
*From Sept 1, 2013, to Dec 31, 2013. †From Sept 1, 2014, to Dec 31, 2014.
Model Results:
• Highest risk to Ghana and Senegal in Africa
• Risk to UK and France combined is about 8 times risk to U.S.
• Travel screening helps reduce risk of international spread
• Exit screening (alone) more effective than entrance screening (alone).
Bogoch et al. The Lancet (2014).
Gomes et al PLOS Currents Outbreaks. 2014 Sep 2
Summary
Contribution from early Ebola models
 Projections of the number of cases
 Potential effects of interventions
 Analysis of local mobility patterns
 Likelihood of international dissemination
Topics for future models
 Causes of this outbreak:
– Social disruption, climate change, urbanization, behaviors?
 Efficiency of response:
– Within Africa
– Within U.S.
Questions?
More info on epidemic modeling:
www.phdl.pitt.edu