In 1917, the “Radium Girls” were plastered over the covers of newspapers all across the United States. They were factory workers painting watch dials with glow-in-the-dark paint; to get fine paintbrush tips, they often licked the brushes. In the process, they contracted radiation poisoning from ingesting deadly amounts of radium. The shockingly high number of cancer patients led to identification of radium paint as a carcinogen and led to stricter legislation of such chemicals.

Disease clusters in general can be due to a particular geological or environmental phenomenon; surprisingly enough, however, they can also be pure, random chance.

In this particular case, the cancer cluster — an unusually high incidence of cancer in a given area — had a very definable cause. Disease clusters in general can be due to a particular geological or environmental phenomenon; surprisingly enough, however, they can also be pure, random chance.

About 1 in 10000 children in the United States contract leukemia; we’d expect at least some of those to be near each other. Every year, the Centers for Disease Control (CDC) investigates hundreds of clusters, but in all except for about 5 to 15% of these cases, no environmental cause was found. In general, random events don’t distribute themselves evenly throughout a given region. The mathematical term is Poisson clumping, and you can see the same phenomenon by taking a jumble of coins or seeds, tossing them into the air, and seeing where they land. They aren’t going to land uniformly, and cancer instances follow the same mathematical pattern.

The CDC’s actual process for investigating cancer clusters (as well as most general types of disease clusters) is a four step process. First, the agency gathers information about the individuals involved in the cluster. How old are they? What areas do they frequent, and are there any commonalities? Investigators then compare their data to the state and national average, and if a significant excess of cases is found, the investigating organization examines whether an epidemiologic study can be conducted. For example, when the number of afflicted individuals isn’t very large, even in clear clusters, often there simply isn’t enough data to merit further investigation.

Finally, officials try to tie the clusters to environmental factors — toxic waste, water contamination, etc. This step most often fails, because even in clusters with a high number of affected individuals, there isn’t necessarily a defined cause.


Cancer data from the New York State Cancer registry. The distribution pattern is evidently non-uniform, yet the majority of clusters are not due to potential hazards.

In 2012, a group of investigators from various institutions published a study in Critical Reviews in Toxicology, asking whether state and federal investigations into cancer clustering had been successful. Their conclusions: It is fair to state that extensive efforts to find causes of community cancer clusters have not been successful. There are fundamental shortcomings to our current methods of investigating community cancer clusters. They go on to state that a greater national discourse on cancer and disease clusters was necessary to reduce false investigations.

Still, cancer clusters pose an interesting conundrum for researchers. The existence of a statistically significant number of cancer cases certainly can’t be ignored — the Radium Girls prove that. A repeated number of cancer clusters, all with an underlying environmental factor, could point to causes that require immediate attention.  More often than not, however, our own human tendency to search for patterns works against us, causing us to see impending doom instead of simple probability.

About The Author

Rohit Dilip