Friday 18 May 2012
Article published
in CEA Techno(s) n° 86

GLOBAL EPIDEMICS

Air travel's key role in transmission

How does the structure of the world's air network influence the spread of global epidemics? To answer this question, specialists in statistical physics at the CEA and other laboratories have created models of this network and simulated disease outbreaks. Their tool is unlike any other in the world and makes it possible to compare the different strategies for combating epidemics.

Why do pathologies that develop in one
area of the world reach certain countries
faster than others? How are such pathologies
propagated throughout the world over
the course of a few weeks? To find out, the
air travel network is a good place to start, as
it's the only one linking all countries, easily
moving large numbers of people across the
oceans. For example, there are one million
travellers per year for the Paris-New York
link alone.


Public health authorities are well aware of this,
multiplying the precautions taken in planes
and airports during epidemics. But what
they lack is a tool to predict the dynamics of
transmission and to respond appropriately.
Where will the epidemic hit first? How much
time will officials have to take the necessary
action? It may now be possible to bridge such gaps in knowledge using the modelling tool
designed by French researchers (CEA, CNRS,
INSERM) and their American counterparts
(Indiana University) in the fields of statistical
physics and epidemiology.


«Using data provided by IATA*, we have
described the main characteristics of the
global air network. This includes the major
airlines, which serve 3000 airports in 200
countries, and their yearly passenger
numbers», explains Marc Barthélémy, who
heads up this research at the CEA. «We can thus
understand how this network
functions and most importantly,
its heterogeneity. That
is, certain regions are served
by a greater number of flights
than others.»


In a second step, the researchers
added to this model a
dynamic process ? the spread
of an epidemic (with its incubation time and
its probability of infecting new subjects).
This led to an extraordinary predictive tool,
validated with data from the SARS epidemic.
This tool shows how epidemics spread from
their points of origin, with varying intensities
and time frames in different countries.


For each outbreak
site, the infection's
preferential means
of transmission are
highlighted.


For each starting point, the infection's
preferential means of transmission are
highlighted, which relate to the air network's
heterogeneity. For example, in the event of
a pandemic, Madrid's main concern would
be travellers arriving from Paris and London.
For Paris, the major source of contamination
would be Hong Kong. In the United States
(see maps on facing page), where the airline
network is very dense, an outbreak from Asia
would first attack the East and West Coasts
before spreading throughout the country.



A cooperative versus a «selfish» strategy



The researchers also set out to measure
the effectiveness of travel restrictions for
combating epidemics. Their conclusions
are compelling: «Even by reducing air
traffic 50?60%, which would not be economically
viable, we would have little impact
on the number of people infected», states
Barthélémy. «There are too many loopholes
and too many possible pathways for an
epidemic to spread.»


So how can we respond? A more cooperative
strategy has been tested. Instead of
restricting air traffic, the richest countries
would give 10?20% of their antivirals to a
pool managed by the WHO**, which would
then redistribute these medicines to the
hardest hit and most vulnerable countries
(i.e. the poorest nations) to slow the spread
of disease during development and
production of a vaccine.


With this strategy, the epidemic is attacked
at its source and thus significantly slowed.
«One of the positive
findings with this model
is that fighting infections
where they start is more
effective than trying to
control air traffic, whose
role in transmission is
just too great», notes
Barthélémy.


The researchers have published all their
findings, making them freely accessible.
They are now focusing on transmission at
the city level and have already collaborated
with a transport manager for an urban area
of 12 million inhabitants. They are also ready
to cooperate with health officials to help
refine prevention policies.



* International Air Transport Association


** World Health Organisation




Trying to slow epidemics by reducing air traffic to contaminated zones is ineffective; even with a reduction of 50–60%, there would be little impact on the number of people infected. 

  • Robust model, validated on a real epidemic
  • Open-source tool (requires expertise in statistical physics)
  • Applicable to other types of networks and geographical entities: continents, countries, regions, cities, etc.