Wednesday, May 12, 2010
This week, New York State released the first set of maps that allow individuals to map cancer cases and align them with various locations, like power plants and waste storage facilities (map). While data and maps can be powerful and fascinating tools (my personal favorite this week is the Die-Hard baseball fan graphic), the New York maps fail to account for some key data points.
In the past, this type of map was used to identify potential locations of “cancer clusters” – areas where people felt the number of cancer cases was particularly high. But, when researchers looked closer, nearly all of these cancer clusters were explained by other factors that also tended to cluster in these areas. For example, cancer is a disease of age – age is one of the strongest risk factors for the most common cancers. As a result, areas where there is a high prevalence of older adults, will have more cases of cancer. Similarly, breast cancer rates are higher in Ashkenazi Jewish women, so a town or county that has a high number of Ashkenazi Jewish women, like certain suburbs of New York City, should be expected to have a higher rate of breast cancer than a town or county further upstate where there are fewer Ashkenazi Jewish women.
Cancer is also a disease that develops over many years, unlike infectious diseases like the flu or acute incidents like salmonella poisoning. Mapping of diseases can be extremely valuable in those instances as it can identify a possible geographic location for the outbreak (i.e., a grocery store selling tainted chicken or the mapping of the 2003 SARS outbreak). But in the case of cancer, the location where an individual lives at the time of diagnosis may have little to do with the relevant exposures or risks for when the cancer developed. As we noted in an earlier CNiC post, we are increasingly finding a role for early life exposures in cancer risk. Cancer maps like those from New York provide little to no context for such risks.
As a result, any mapping of this kind is a useful tool in the context of a much larger and more detailed investigation of cancer rates, but should be used with extreme caution, if at all, otherwise.
For more on this, check out the National Cancer Institute's Cancer Cluster Page