Abstract B28: Spatial analysis of socioeconomic status and breast cancer stage in Louisiana

Background: Breast cancer is the most common cancer among Louisiana women; in 2007 approximately 3,800 cases and 730 deaths will be reported. Recent data from the Louisiana Tumor Registry indicates that the age-adjusted mortality rate for black women in LA shows significant excess when compared to white women in LA (36% higher) and black women throughout the U.S. (16% higher).

Methods: Women diagnosed with breast cancer between 1996 and 2004 were identified from the Louisiana Tumor Registry database. Information about breast cancer diagnosis, such as stage, race (black, white only), age and the woman's address of residence at the time of diagnosis were ascertained. Geographic and population data for LA census tracts was obtained from the US Census Bureau's population files. Reference databases for geocoding were built from the newly released Census Bureau 2007 TIGER/line shapefiles. ArcSDE address locators for each Louisiana parish were built from these reference databases and patient addresses geocoded to census tract and latitude and longitude. The Spatial Scan Statistic developed by Kulldorf was utilized to identify geographic areas with significantly elevated rates of women diagnosed with late stage breast cancer in Louisiana between 1996 and 2004. Breast cancer incidence rates were modeled using SatScan. The identified areas were mapped using ArcGIS. The identified areas were characterized using area based measures derived from the 2000 U.S. Census data.

Results: Results from the SatScan Poisson models for white & black women revealed three clusters of census tracts with significantly elevated or decreased breast cancer incidence as estimated by the spatial scan statistic. Two clusters were statistically significant in terms of an excess risk of late stage breast cancer. One cluster was associated with an almost seven fold increase rate of late stage breast cancer among white women relative to white women in the other regions of the state. This region may be characterized be a younger female population with a greater proportion of white females below poverty, than the state of LA overall

Conclusions: This approach has profound implications for identifying vulnerable populations at the highest risk for late stage disease and to guide future breast cancer control efforts. Further, the methodology may be applicable to other cancers that are amenable to large scale screening efforts. Identifying communities with lower rates of screening or advanced stage cancers are critical first steps toward cancer control.