Lung cancer mortality among women in Xuan Wei, China: a comparison of spatial clustering detection methods.

The identification of spatial clusters of lung cancer mortality can be a useful instrument in detecting locations with high risk of this disease. This study compared 2 methods for identifying spatial clusters of village-level women lung cancer mortality rates in Xuan Wei. One used a local indicator of spatial association to detect which groups of neighboring villages had lung cancer mortality rates that were significantly related to each other. The other was a spatial scan technique that calculated a maximum likelihood ratio of lung cancer deaths relative to the underlying population in order to identify the group of villages with relatively higher risk. As each technique based its cluster detection process on its own criteria, different clusters of villages were identified. However, the overlapping indicated that the 2 methods illustrated different components of the same clusters. These spatial analytic techniques were complementary to each other and can be used jointly rather than separately.