Area-to-point Poisson kriging analysis for lung cancer in Perth areas
Faculty of Computing, Health and Science
School of Engineering
This paper provides an analysis of lung cancer incidence in the Perth metropolitan area during the period 1990-2005. The Statistical Local Areas (SLAs) in Perth vary in size and shape, and the incidence case count follows an inhomogeneous Poisson process. For this reason the geostatistical methodology areato-point (ATP) Poisson kriging is adopted. For lung cancer in Western Australia (WA) incidence rates are higher for males than for females and it is more common among old people than younger people. In this paper, sex is first taken into consideration for geostatistical cancer analysis. Crude rates, age-adjusted rates and age-sex-adjusted rates are compared. ATP Poisson kriging estimates based on the two adjusted rates are quite similar. However, the ATP Poisson kriging variance based on the age-sex-adjusted rates is much lower than that based on the age-adjusted rates. Slight difference exists between the semivariogram of these two adjusted rates. The age-sex-adjusted rates will facilitate the correlation between lung cancer and population distribution and composition. Higher cancer rates with low variance are normally found close to the Perth metropolitan area. These areas are not only densely populated but also the average age is relatively higher than in other areas.