Press release

 

How statistical mechanics may contribute to our understanding
of the way epidemics spread

Paris, June 11, 2003

 

At a time when we are faced with so many questions about severe acute respiratory syndrome (SARS), we must also ask ourselves if the epidemic has reached its peak or not. Research carried out by Frédéric Van Wijland(1) of the "Laboratoire de Physique Théorique" (Theoretical Physics Laboratory, CNRS, Université Paris Sud 11)(2) , shows how a theoretical field can make an original contribution to this issue. This research makes it possible to increase our in-depth knowledge of phenomena that appear to be highly unrelated. For the researcher, there is a common point between sandpile avalanches, autocatalytic reactions and how an epidemic spreads. All of these phenomena are studied within the framework of "nonequilibrium" systems. In other words, there is no general theory that describes the different macroscopic states on the basis of microscopic data.

We are dealing here with concepts of thermodynamics that focus on the macroscopic aspect of the behavior of matter (defined by scientists in a "system") on the basis of its properties (temperature, energy, etc.) or the microscopic aspect, by analyzing the mechanisms of atoms in motion to understand the relationship between these variables. Just as in the case of equilibrium systems (characterized by continuous and constant variables), we can observe phase transitions or changes in the state of "nonequilibrium" systems as well. And in the case of an epidemic, it is a transition between a state where the epidemic maintains itself indefinitely and a state where it dies out without the possibility of resurfacing again. We refer here to the transition of an active state (the epidemic in action, the avalanche in progress, etc.) towards an absorbing state (complete disappearance of the infectious agent, inert sandpiles, etc.).

One of the most commonly used modelling methods for epidemics consists of dividing the population into two groups: infected (I) and healthy (H). In the case of an infectious agent that is not deadly, the infected individuals recover their health spontaneously after a predetermined amount of time (I-->H). The infection of a healthy individual takes place during an encounter with an infected individual (I+H-->I+I). In this problem, we can observe that the epidemic will only spread if the population is big enough and that it will die out on its own in the opposite case. Once it has disappeared, it can never again come back. The scientist wanted to understand exactly what happens when the population reaches the threshold where the epidemic is on the verge of disappearing. It turns out that in the case of this population density regime, certain types of behavior emerge, known as universal in that they do not depend on the details of the problem itself (speed at which individuals move from one place to another, contamination rate, recovery speed, etc.). Other collective behaviors also emerge in this regime that are difficult to understand. Basically, everything must converge when the epidemic dies out. In a certain sense, a piece of information passes through the population, whereas contact occurs from one individual to the next in the contamination process; the idea of complexity is demonstrated here. Frédéric Van Wijland has identified these universal characteristics inherent in the propagation of an epidemic and has shown that they are the same for other phenomena: chemical autocatalysis problems (that lead to the poisoning of a surface, for example) or the sandpile avalanche. The universal characteristics are found in the way that the epidemic is spread or dies out, a system in which individuals speak to each other from one end to the other of the system.

The researcher determined an underlying critical point that, regardless of the precise definitions of these different models, makes them remarkably similar. This critical point is identical for the epidemic, the self-organized criticability of the sandpile and some chemical reactions. The direct consequence of its existence is that many do not depend on the microscopic details that define the systems being studied. The idea of universality, which we already knew was applicable to nonequilibrium phenomena, takes on a new meaning here since it bridges the gab between several widely unrelated fields, particularly epidemiology and self-organized criticability. These results cast doubts on beliefs that have been widespread for a good decade now.

1 - Physical Review Letters 89 (2002), entitled "Universality Class of Nonequilibrium Phase Transitions with Infinitely Many Absorbing States".
2 - Researcher at the Laboratoire de Physique Théorique of Orsay (http://qcd.th.u-psud.fr/), Frédéric Van Wijland is also a member of the research federation, "Matière et Systèmes Complexes" (Complex Systems and Matter, CNRS, Université de Paris 7), the physics center that will be located at the new site of the Université de Paris 7 "Tolbiac - Grands Moulins" in Paris (http://hogarth.pct.espci.fr/~jbf/msc/).




Researcher contacts:
Frédéric Van Wijland, Laboratoire de Physique Théorique
Tel: 00 31 30 253 28 33 or 00 31 30 271 12 36
E-mail: frederic.van-wijland@th.u-psud.fr
Alessandro Vespignani, Laboratoire de Physique Théorique
Tel: +33 1 69 15 82 12


CNRS Mathematics and Physical Sciences Department contact:
Frederique Laubenheimer
Tel: +33 1 44 96 42 63
E-mail: frederique.laubenheimer@cnrs-dir.fr

Press contact:
Magali Sarazin
Tel: +33 1 44 96 46 06
E-mail: magali.sarazin@cnrs-dir.fr


Paper: cond-mat/0209202
From: Frederic.Van-Wijland@th.u-psud.fr (van wijland)
Date: Mon, 9 Sep 2002 11:19:18 GMT (10kb)