A cancer cluster is the occurrence of more than the expected number of cancers within a group of people, a geographic area, or a period of time. Cancer clusters may be suspected when people report that several family members, friends, neighbors, or coworkers have been diagnosed with the same or related cancer(s). Differentiating between cancers that have occurred by chance and those that might have a common cause is often difficult. With their knowledge of diseases, environmental science, and statistics, epidemiologists try to distinguish actual cancer excesses from excesses that are due only to chance. Epidemiologists generally suspect that an excessive number of cancer cases is a true cluster if it involves a large number of cases of a specific type of cancer (rather than several different types), a rare type of cancer (rather than common types), or an increased number of cases of a certain type of cancer in an group that is not usually affected by that type of cancer. In addition, successful investigations have often involved high-level and relatively well-defined exposures.

Often, a first step in cluster evaluation is the calculation of the probability that the increase in the number of cases is due to chance alone. To perform such a calculation one would need to ascertain the number and type of cancers, count the number of person-years at risk, and calculate the expected number of cancer cases. Even when the increase in the number of cases appears to be unusual, interpretation is difficult, due to the Texas Sharpshooter Fallacy. The name comes from a story about a Texan who fired several shots at the side of a barn, and then painted a target centered on his hits and claimed to be a sharpshooter. In the context of cancer clusters, the issue is the definition of the cluster in time and space and the existence of implicit multiple comparisons.

In this talk, I will discuss how to approach the investigation of cancer clusters and how to address uncertainty due to small numbers, poorly identified study populations, and vague definitions of exposure and disease. True stories of such investigations, often performed in a politically-charged environment, will serve as examples.