Cancer cluster detection with incidence registry data

B86

Purpose: The role of desktop cluster detecting programs in the development of new surveillance models to identify communities at risk, leading to the implementation of tailored, community-level interventions to lower cancer burden, was investigated.
 >Methods: A cluster detection focused test, the Lawson & Waller Test Statistic, was used to detect clustering around hazardous waste site point sources, which were on the Environmental Protection Agency (EPA) National Priority List (NPL) for clean-up. Using publicly available data, we identified 38 sites currently on the NPL that are contaminated with potential breast cancer carcinogens.
 >Results: Seven of 38 NPL sites were identified as having higher rates of invasive breast cancer around the sites; four of which were statistically significant (p = 0.001). Differences in the chemical contaminants, hazard scores, contaminated media, and breadth of contamination and clean-up were identified. The average hazard scores for the sites contaminated with benzene, chromium, polycyclic aromatic hydrocarbons (PAHs), and vinyl chloride were higher for the sites with elevated breast cancer rates. The average hazard score for sites contaminated with pesticides were lower for the sites with elevated breast cancer rates. The sites identified with elevated breast cancer risk were more likely to be contaminated with PAHs and vinyl chloride, and the sites without elevated risk were more likely to be contaminated with pesticides.
 >Conclusions: Cancer registries allow for the systematic analysis of data to identify burden, trends, to generate hypotheses about cancer risk and etiology, and to target high-risk groups in need of appropriate interventions. The growing ease and capability of desktop geographic information systems (GIS) software permits the combination of disparate data sources for analysis, such as community based risk assessment. However, interpretation of cluster analysis results is problematic. These preliminary results must be interpreted with caution because the results do not include an evaluation of the demographic and behavioral features of the underlying population.
 >Although breast cancer incidence has an inverse relationship with socioeconomic status, populations living in poverty and near a contaminated site may have higher risk for breast cancer. This study has public health significance because our approach may provide a screening tool for proactively addressing potential public environmental concerns and identifying high risk communities in need of further investigation.