Cancer clusters at random

Miles O’Brian posts this story to Boing Boing today and an Australian piece regarding cancer clusters. I contrast the ideas here.

Erin Brockovich is now focusing on cancer clusters.

Erin Brockovich: the real-life unhappy ending of Hinkley, California, and a tale of science for sale – Boing Boing.

The real-life Erin Brockovich has moved onto the national stage as a consumer advocate and now curates a crowd-sourced map of reported cancer clusters. It is a real eye-opener. And it makes you wonder why environmental regulators don’t do this kind of thing.

A few years ago, The Environmental Working Group did a study of U.S. tap water, and it found a chrome-plated, potentially carcinogenic mess. They tested tap water samples from 35 cities and found chromium-6 in 31 of them.

So, this article paints a scary picture. I don’t have enough information on this topic to have an opinion on what is valid or whether to be outraged or not. But I do have some interest in cancer clusters. Because often, they are no such thing.

Cancer clusters in the USA

via Cancer clusters.

Source: Critical Reviews in Toxicology: Cancer clusters in the USA. What do the last 20 years of state and federal investigations tell us?

+567 suspected cancer cluster sites were studied over the past 20 years.

+In 87 per cent of cases, a true increase in cancer rates could not be established, let alone a cause.

+In only 0.5 per cent of the cancer types was there some evidence of a link between the cancers of concern and the suspected carcinogen.

Clusters are part of randomness. But most people can’t grasp that because it feels wrong. In the case of cancer clusters we are dealing with SO MANY VARIABLES! It’s so difficult to wrap your head around them all – each person’s lifestyle and genetics, the multiple environmental factors, chance. Logic tells us that if there was a “cause”of the cluster, it should sift out of the data analysis. It would be obvious. But that’s not been the result of most investigations. Then, the people involved remain suspicious and concerned, unsatisfied. It’s hard to blame them when there is no clear resolution.

This quote from the SMH article was especially interesting:

Phyllis Butow, professor of psychology at the University of Sydney, is an expert in the psychological impact of disease. She says cancer brings with it a special fear. “Data suggests people catastrophise about risks they see as life-threatening; they search for the threat, in fact, and are more likely to see it,” she says.

“There is a lot of evidence that people find cancer more scary than other diseases [with] just as high mortality rates.”

CDC – Cancer Clusters – Home.

Cancer Clusters – National Cancer Institute.

  5 comments for “Cancer clusters at random

  1. Mr. Shreck
    March 14, 2013 at 10:59 AM

    Tangential, but maybe of interest to some: Cancer maps and the fear of environmentally caused cancers (and environmental and lifestyle diseases in general) figure prominently in Peter Carey’s excellently surreal novel Bliss.

    Cancer maps at National Cancer Institute:

  2. Stamen
    March 14, 2013 at 4:20 PM

    This article is apropos. I’m pleased to see you have a link to the death of Chavez by cancer. I’ve seen a few articles suggesting he was murdered. Supporting evidence is given by the fact that a number of South American leaders have developed cancer lately. The CIA is certainly not shy about extra judicial murder, but a cluster can indeed be random. I would imagine these leaders also have in common, advanced age, rich western diets, lack of exercise.

  3. March 14, 2013 at 6:38 PM

    Skeptics with good memories will remember a very famous scare in the 1980’s in which people believed that power lines near your house caused childhood cancers. It was a classic “moral panic” created when an epidemiologist did a flawed study of a cluster of childhood leukemia in Colorado. It got picked up by a New Yorker writer named Paul Brodeur, who then wrote a series of fear mongering articles and later books on all manner of alleged problems caused by electromagnetic fields. This was a very big deal for about a decade, even affected resale value of homes for a while.

    It was all pretty much nonsense, as the original study had some serious flaws (see the link above). But Brodeur’s writing still gets cited by folks who are pushing the modern versions of this – cell phones, WiFi and Smart Meters. I wrote about that connection in this Skeptic History article for

    As usual it pays to remember the past – cancer cluster studies can be very useful, but if they are done poorly they can definitely lead one astray.

  4. AmSci
    March 15, 2013 at 12:47 PM

    If only more people payed attention in probability and statistics. Maybe they just do a bad job teaching it.

  5. Mr. Shreck
    March 15, 2013 at 1:07 PM

    I definitely had a few math profs that made their subject harder than it needed to be, but I’d be reluctant to blame the teaching in most cases. Even when people theoretically understand statistics, their beliefs and fears about other inputs and tendency to assign causality come into play and make them unable to accept this is statistics at work. Theoretical statistics are harder to accept than a concrete example that reinforces a bias.

Comments are closed.