Retrospective space-time analysis of H5N1 Avian Influenza emergence in Thailand
1 Center of Excellence for Vectors and Vector Borne Diseases, Faculty of Science, Mahidol University at Salaya, 999 Phutthamonthon 4, Nakhon Pathom 73170, Thailand
2 UMR 190, IRD, 44, Bd de Dunkerque 13572 Marseille Cedex 02, France
3 RS&GIS FoS, Asian Institute of Technology, PO Box 4, Klong Luang, Pathumthani 12120, Thailand
4 Centre International de Recherches Médicales de Franceville, BP 769, Franceville, Gabon
5 Laboratoire Espace, Santé et Territoires, Université Paris Ouest-Nanterre La Défense, 200 avenue de la République, 92001 Nanterre Cedex, France
International Journal of Health Geographics 2010, 9:3 doi:10.1186/1476-072X-9-3Published: 27 January 2010
The highly pathogenic avian influenza (HPAI) H5N1 virus remains a worldwide threat to human and animal health, while the mechanisms explaining its epizootic emergence and re-emergence in poultry are largely unknown. Data from Thailand, a country that experienced significant epidemics in poultry and has recorded suspicious cases of HPAI on a daily basis since 2004, are used here to study the process of emergence. A spatial approach is employed to describe all HPAI H5N1 virus epizootics from 2004 to 2008 and to characterize the pattern of emergence: multiple independent introductions of the virus followed by moderate local spread vs. very rare emergences followed by strong local spread and rare long range diffusion jumps. Sites where epizootics originate (by foreign introduction, local persistence, or long range jump) were selected from those to which the disease subsequently spreads using a filter based on relative date and position. The spatial distribution of these selected foci was statistically analyzed, and to differentiate environmental factors from long range diffusion, we investigate the relationship of these foci with environmental exposure factors and with rearing characteristics.
During each wave of epizootics, the temporal occurrence of cases did not show a temporal interruption of more than a week. All foci were globally clustered; i.e., more than 90% of cases had a previous case within a 10 km range and a 21 day period of time, showing a strong local spread. We were able to estimate 60 km as the maximum distance for the local farm to farm dissemination process. The remaining "emergent" cases have occurred randomly over Thailand and did not show specific location, clusters, or trends. We found that these foci are not statistically related to specific environmental conditions or land cover characteristics, and most of them may be interpreted as long range diffusion jumps due to commercial practices.
We conclude that only a few foci appear to have been at the origin of each HPAI epidemic wave, leading to the practical action that surveillance and control must focus on farm to farm transmission rather than on emergence or wild fauna.