The role of land use patterns in limiting the spread of equine influenza in Queensland during the 2007 epidemic.

In 2007, an epizootic of equine influenza (EI) occurred in Australia, involving parts of the states of Queensland and New South Wales. Following an extensive control program, the disease was eradicated within 4 months, after infecting more than 75,000 horses on over 10,000 properties. In Queensland, examination of land use patterns revealed that the majority of infected premises (89.5%) were located in one of three land use classes viz. rural residential, residential-unspecified or grazing natural vegetation. All nine clusters of infection in Queensland were surrounded to some degree by parcels of land whose designated use precluded the presence of horses. In four clusters, these horse-free areas completely enclosed the cluster and spread was limited to a relatively small area whereas the remaining five clusters were not completely surrounded by horse-free buffers and spread was much more extensive. The cumulative incidence of infection with EI dropped from 41% of properties in rural-residential areas to 4% of properties located further than 5 km from rural-residential areas. This decrease was associated with a 10-fold increase in the median size of properties and a 5-fold decrease in the density of horses on those premises. There was a strong negative correlation between the cumulative incidence of infection and the median property size (R = -0.91, P < 0.05) and a positive correlation with the horse density on those properties (R = 0.93, P < 0.01).