To the Editor:
CitationGoodman et al. (2012) conclude that extensive efforts to find causes of community cancer clusters have not been “successful”. They base their measure of success strictly on whether these studies have improved our understanding of cancer etiology. We contend that the authors’ arguments have major flaws, notably in two areas: 1) their selective definition of a cancer cluster and 2) their misguided assessment of whether community cancer cluster investigations meet their measure of success.
The authors reference the Centers for Disease Control and Prevention’s (CDC) definition of a cancer cluster (CitationKingsley et al., 2007) as a perceived aggregation of cancer cases within a group of people in a geographic area over a defined period of time. In fact, Kingsley et al. define a cancer cluster as a greater-than-expected number of cancer cases within a group of people in a geographic area over a defined period of time. Identifying a cluster based on perception is quite different from evaluating the relevant data to determine whether a pattern of cancer appears unusual. The National Cancer Institute further defines a suspected cancer cluster as being more likely to represent a true cancer cluster (rather than coincidence) if it involves: (1) a large number of cases of one type of cancer, (2) a rare type of cancer, (3) a type of cancer in an age group not usually affected by that cancer, or (4) cancers of similar causes occurring over a short period of time (CitationNCI 2006). Goodman et al. consider a cancer cluster to be confirmed based on statistical significance alone, which is just one measure to consider in evaluating cancer incidence data, particularly in small area analyses. Statistical significance does not necessarily mean biological significance.
Experience has shown that cancer cluster investigations are valuable for several reasons:
They allow for education of the public on basic cancer facts: according to the American Cancer Society, one in two men and one in three women will develop some form of cancer in their lifetime.
They provide an opportunity to explain that the term cancer is often used to describe a multitude of diseases, most of which have different etiologic factors.
They allow for education of the public on the principles of environmental exposure: a source of contamination without a pathway for exposure does not increase risk.
They allow for promoting the importance of cancer screening and other primary and secondary interventions.
They often uncover other factors beyond the environment that may explain an excess of a particular type of cancer (e.g. screening rates that are significantly different from what is typical).
The use of readily-available cancer registry data allows state health department staff to evaluate risk factor patterns to determine if, for example, age at diagnosis, cancer subtype, and tobacco history information are consistent with what would be expected for a particular cancer type, based on national and state statistics and the epidemiological literature. Even if an environmental culprit is not found, evaluating the pattern of cancer in an area can better inform the public on the community’s health status, risk factors for particular types of cancer, and the status and potential exposure opportunities associated with a site or property of concern due to contamination.
The authors also contend that cancer cluster investigation methods have essentially not changed in the past 20 years. This also is incorrect. In May 2010, the Council of State and Territorial Epidemiologists (CSTE) and CDC convened a technical working group to update CDC’s 1990 Guidelines for Investigating Clusters of Health Events. (Note: one of the commenters is among a group of public health scientists updating CDC’s cancer cluster investigation protocol and co-author of the referenced Costas et al. paper.) The updated guidelines developed by the working group are imminent. These updates, along with the use of Geographic Information System and statistical software (such as SaTScan), have improved the science of cancer cluster investigations significantly.
The authors imply that unless a causal link is established, the effort is worthless. We argue quite the contrary. Responding to concerns about community cancer patterns is a responsibility of state health officials and our experience has shown that it is a valuable service. Although not in the same way as animal bioassays or clinical trials, we believe that investigating what the authors refer to as cancer clusters both contributes to the public’s understanding of cancer etiology and informs them about cancer prevention and control.
Declaration of interest
The authors of this letter and their employment affiliation are as shown at the end of the letter. The letter was prepared during the normal course of their employment; however, the authors have sole responsibility for the writing and content of this letter and the opinions expressed. As noted, one of the authors is assisting CDC in updating its cancer cluster investigation protocol.
References
- Costas K, Knorr RS, Condon SK. (2002). A case-control study of childhood leukemia in Woburn, Massachusetts: the relationship between leukemia incidence and exposure to public drinking water. Sci Total Environ 300:23–35.
- Goodman M, Naiman JS, Goodman D, LaKind JS. (2012). Cancer clusters in the USA: what do the last twenty years of state and federal investigations tell us? Crit Rev Toxicol 42:474–490.
- Kingsley BS, Schmeichel KL, Rubin CH. (2007). An update on cancer cluster activities at the Centers for Disease Control and Prevention. Environ Health Perspect 115:165–171.
- NCI (National Cancer Institute). (2006). NCI Fact Sheet: Cancer clusters. [Online] Available at: http://www.cancer.gov/cancertopics/factsheet/Risk/clusters. Accessed on 28 September 2012.