SYNTHETIC CANCER VARIABLES AND THE CONSTRUCTION AND TESTING OF SYNTHETIC RISK MAPS

Cancer cluster investigations are usually univariate in nature; they f ocus on a particular cancer, such as leukaemia, and attempt to determi ne whether excess risk is associated with a suspected cancer-causing a gent. Although several causes of death (such as leukaemia, lymphoma, H odgkin's) may be considered, the approach is univariate because the ca uses of death are analysed sequentially and independently of one anoth er. This approach is consistent with a one-cause one-effect model. Rar ely, however, is the action of a carcinogen manifested at only one bod y site, and correlations among causes of death are the norm rather tha n the exception. A multiple effects model is therefore appropriate, an d the multivariate nature of cancer mortality data should be exploited when exploring geographic pattern in cancer risks. This paper describ es such an approach. We construct maps based on a principal components analysis of cancer mortality rates from different geographic areas. T he resulting principal components are called synthetic cancer variable s (SCVs), and maps of the SCV scores are synthetic risk maps (SRMs). T hese maps quantify geographic variation in cancer risk at several body sites simultaneously, and may be analysed for (1) spatial structure a nd (2) geographic association with potential risk factors. As an examp le, we use synthetic risk maps to determine whether high-risk counties in Illinois cluster near nuclear facilities. Much work remains to be done, but synthetic cancer risk maps appear to be a useful tool for qu antifying geographic pattern and multivariate structure in cancer mort ality.