Researchers Model of Avian Flu Outbreak, Impact of
A carefully chosen combination of public health measures,
if implemented early, could stop the spread of an avian
flu outbreak at its source, suggest two international
teams of researchers in Nature (August 3) and
Science (August 5).
The researchers used computer modelling to simulate what
might happen if avian flu were to start passing efficiently
between people in Southeast Asia. They found that antiviral
treatment is a critical component of any multi-pronged
The computer simulations are part of the Models of
Infectious Disease Agent Study (MIDAS) research network
funded by the National Institute of General Medical
Sciences (NIGMS), a component of the National Institutes
of Health. The overall goal is to develop computational
models of disease spread that will aid the development
of effective control strategies.
"These new models illustrate how the fundamental features
of infectious disease spread can be captured to predict
possible outcomes and the potential impact of interventions,"
said Jeremy M. Berg, Ph.D., director of NIGMS. "As these
modelling approaches develop, they will offer policymakers
and researchers powerful tools to use in strategic planning."
The H5N1 strain of the avian flu virus, found in birds
throughout Southeast Asia, has infected a number of
species, including domestic poultry, pigs, and people.
Scientists fear that a genetic exchange between bird
and human flu viruses or the accumulation of H5N1 mutations
could soon make efficient person-to-person transmission
The avian flu strain represents a particular threat
because it is so deadly, said Neil M. Ferguson, D.Phil.,
a computational biologist at Imperial College in London
and lead author of the Nature paper. "A large
percentage of animals and people infected with this
virus have died," he explained. "The consequences of
an H5N1-based pandemic could be catastrophic."
With bird flu continuing to spread in Southeast Asia,
the MIDAS network decided to model a hypothetical human
outbreak of H5N1 in this region.
"The pressing questions are if and how we can contain
an outbreak of avian flu at the source before it becomes
a pandemic," said Ira M. Longini, Jr., Ph.D., a biostatistician
at the Emory University Rollins School of Public Health
in Atlanta and lead author of the Science paper.
To enhance reliability, both models were based on detailed
data for Thailand, such as population densities, household
sizes, age distribution, and distances travelled to
work. The models also included information about the
flu virus, such as the possible contagiousness of an
infected person. Ferguson and Longini noted that actual
contagiousness would not be known before an outbreak.
Although the models differed in the specific scenarios
they simulated and the intervention strategies they
tested, the general conclusions were similar and confirm
current knowledge of how diseases spread: Preventing
a pandemic would require a combination of carefully
implemented public health measures introduced soon after
the first cases appear.
The model presented in Nature simulated 85 million
people living in Thailand and bordering regions of neighbouring
countries. It tested the effectiveness of giving courses
of antiviral treatment to everyone, socially or geographically
targeting who received them, and combining these drug-sparing
approaches with other interventions, such as restricting
The results suggest that an international stockpile
of 3 million courses of flu antiviral drugs, combined
with other interventions, could contain a pandemic.
Treating infected individuals and everyone in their
vicinity, along with closing schools and workplaces,
could have more than a 90 percent chance of stopping
the spread of a pandemic virus, according to the model.
Ferguson emphasised that successful containment would
depend on the early detection of the first cases and
the rapid implementation of public health measures.
The model described in Science simulated 500,000
people living in rural Southeast Asia and relied on
information about how those individuals move within
their communities. Containment strategies included giving
antiviral medication to people in the same social networks,
vaccinating before an outbreak with a vaccine that is
not well matched to the strain that emerges, quarantining
the houses or neighbourhoods of infected people, and
combinations of these approaches.
Giving a low-efficacy vaccine to just half the population
before the start of a pandemic would greatly enhance
the success of other containment strategies, according
to the model. Longini reported that a combination of
targeted antiviral treatment and quarantine introduced
two weeks after the first case had the potential to
successfully contain disease spread, resulting in less
than one case per 1,000 people.
Both models demonstrated that the need for additional
public health measures greatly increased as the virus
became more contagious. "Each measure can have a significant
effect, but it can't contain spread on its own," said
Ferguson, adding that targeted antiviral treatment was
a crucial component of all combined strategies.
While the researchers said that implementing such a
combination of approaches would be challenging and require
a co-ordinated, international response, they offered
this good news: The models show that containing
an avian flu pandemic at its source is feasible.
Because computer models cannot capture all the complexities
of real communities and real outbreaks, the MIDAS researchers
will continue to refine their simulations and test different
scenarios as new information becomes available. By developing
a collection of models, they can compare and contrast
different interventions, leading to more accurate predictions.
Other researchers involved in this work represent Johns
Hopkins Bloomberg School of Public Health; the University
of Hong Kong; the Ministry of Health in Thailand; and
INSERM, the French National Institute of Health and