A spatiotemporal statistical model of the risk factors of human cases of H5N1 avian influenza in South-east Asian countries and China.

OBJECTIVES: This article aims to quantify the risk factors associated with the human cases of H5N1 avian influenza in South-east Asian countries and China; a dangerous region for this disease that has the potential for a pandemic outbreak. STUDY
DESIGN: A statistical model with time and spatial dimensions was built to capture the international spread patterns of this disease.
METHODS: The grid search method was used to fit the model with 2004-2006 data. The grid search approach is a simple procedure that allows the fit of any function to data.
RESULTS: This study found that: (1) when the number of domestic H5N1 human cases increases by one person in a certain time period, the chance that the country will have a human case in the next period increases by 22.10%; (2) when the number of human cases in a neighbouring country increases by one person in a certain time period, the chance that the country will have a human case in the next period increases by 1.62%; (3) when the number of avian cases in a neighbouring country increases by one, the chance that the country will have a human case increases by 0.02%; (4) as the human population increases by one unit, the chance that the country will have a human case increases by 0.10%; (5) when the quantity of imported poultry increases by 1000 metric tons, the chance that the country will have a human case increases by 0.03%; (6) when the outbreak of the disease among domestic birds increases by one, the chance that the country will have a human case increases by 0.19%; and finally (7) when the number of birds destroyed increases by 1000, the chance that the country will have a human case decreases by 0.30%.
CONCLUSIONS: These findings shed new light on the spatiotemporal characteristics of the epidemic, and thus need to be taken into consideration in interdisciplinary and scientific discussion of the disease.