Description and analysis of the poultry trading network in the Lake Alaotra region, Madagascar: Implications for the surveillance and control of Newcastle disease

Madagascar's 36.5-million-head poultry industry holds a foremost place in its economy and the livelihood of its people. Unfortunately, regular Newcastle disease outbreaks associated with high mortality causes high losses for smallholders and threatens their livelihood. Therefore, Madagascar is seeking concrete, achievable and sustainable methods for the surveillance and the control of Newcastle disease. In this paper, we present and analyze the results of a field study conducted in Madagascar between December 2009 and December 2010. The study area was the Lac Alaotra region, a landlocked area in the north-eastern part of the country's center. Poultry trading is suspected of playing a major role in the spread of avian diseases, especially in developing countries characterized by many live-bird markets and middlemen. Therefore, the goals of our study were to: (i) describe and analyze smallholders' poultry trading network in the Lake Alaotra region using social network analysis; (ii) assess the role of the network in the spread of Newcastle disease; and (iii) propose the implementation of a targeted disease surveillance based on the characteristics of the poultry trading network. We focused our field study on the harvesting of two data sets. The first is a complete description of the poultry trading network in the landlocked area of Lac Alaotra, including a description of the poultry movements between groups of villages. The second set of data measures the occurrence of outbreaks in the same area by combining a participatory approach with an event-based surveillance method. These data were used to determine the attributes of the network, and to statistically assess the association between the position of nodes and the occurrence of outbreaks. By using social network analysis techniques combined with a classification method and a logistic model, we finally identified 3 nodes (set of villages), of the 387 in the initial network, to focus on for surveillance and control i