Background:There is some evidence that there are variations across geographical areas in breast cancer incidence rates in Western Australia.
Aim:To investigate the breast cancer incidence distribution over WA and if there was any observed incidence greater than expected.
Method:A spatial scan statistic using SaTScan was selected to complement descriptive statistical methods to identify areas (approximate locations) with apparently increased incidence of breast cancer in WA. This technique addressed the limitations of small numbers of breast cancer cases and low populations, accounted for the pre-selection bias and multiple testing inherent in a cancer cluster investigation, and adjusted for confounding factors such as age, race, and accessibility/remoteness. Demographic data and age-specific breast cancer incidence rates for women for 5,508 locations (Statistical Area Level 1) for 2011-2013 were obtained from the Australian Bureau of Statistics and WA Cancer Registry.
Results: The total number of cases for 3 years was 4,326 cases with annual cases 159.9 /100,000 women. Two significant female breast cancer clusters identified: i) the most likely cluster (5 SA1 locations) with 4.42 times (P value of <0.008) higher incidence ratethan the rest of the State; ii) the secondary cluster (9 SA1 locations) with 3.5 times (P value of <0.02) higher incidence rate that the rest of State. There were also 20 non-significant clusters.
Conclusions/Implications: The scan statistic is an important addition to the public health toolbox as a screening tool for evaluating which disease cluster is probably a chance occurrence and which cluster merits further investigation.