Scientific research is often difficult to understand and interpret, even for those of us who do it frequently. Translating research into sound bites that fit the news media’s tight space constraints is even more difficult. Throw in a catchy phrase that fails to capture that complexity, and you have the mix for an interesting situation.
Such was the case recently with an article and a news report that mentioned “bad luck” and cancer. It was published on January 2, 2015, in the widely read and well-regarded journal Science.
The authors of the article — Bert Vogelstein and Cristian Tomasetti — are a formidable team of scientists. Vogelstein, a professor of oncology and pathology at Johns Hopkins University in Baltimore, needs no introduction to many of us, especially those familiar with his legacy of outstanding research in understanding the propagation of cancer at the genetic level. Tomasetti, a researcher and computational mathematician at Johns Hopkins, is clearly well versed in how to apply advanced mathematics to understanding some of the questions that have vexed scientists for many years.
When you read their report carefully, you understand the question they were trying to answer, a question that has interested cancer experts for a long time: Why do some tissues produce so many cancers (such as colon, lung, breast), while cancers of others are much more rare (such as brain and bone)? And why do some tissues seemingly with the same infrastructure — such as the stomach, small intestine, and colon — have vastly different frequencies of pre-cancers and cancers? Why do those same tissues have different relative frequencies of cancer in different species (more small intestine cancers in mice, compared with more colon cancers in man)?
To answer these questions, the researchers found information on 31 tissue types where the number of stem cells — the basic cells that give rise to each organ and tissue in our bodies — had been determined. They found that much of the difference in frequency of cancer among organs was related to the number of times the stem cells divided in a particular organ. More divisions meant more cancers — and vice versa. They concluded that about 65 percent of the time, the difference in frequency of cancers in specific tissues was related to the number of times that a particular tissue’s stem cells divided. Errors can occur during those divisions that our bodies’ repair mechanisms can’t fix. This accumulation of errors can lead to problems in a stem cell’s DNA, which can give rise to cancer. To explain this more clearly, the authors called this cause of the majority of cancers in this group “bad luck.”
There were some missing pieces of data in the research, especially the lack of information on two of the most frequent cancer types: breast and prostate. But the association was — at least from a statistical point of view — very intriguing.
Vogelstein and Tomasetti then took the next logical step. They pointed out that some cancer rates are determined by problems occurring when cells divide. In these situations, the authors concluded that this process was responsible for most of the cancer burden related to those tissues.
However, there are other cases where environmental and genetic influences impact the frequency of cancer in a specific tissue (think lung cancer in smokers versus non-smokers and breast cancer in BRCA positive women). To emphasize this point they showed a significant difference in cancer rates where such factors may or may not be present. For example, pancreatic islet cell cancer — which isn’t known to be related to factors such as smoking — was at one end of the spectrum, while colon cancer in families who have an inherited syndrome — which results in many colon polyps and a greater risk of colon cancer — was at the other.
Conveniently, the model showed a much lower rate of lung cancer for non-smokers, with a higher rate for smokers. But there were some inconsistencies for cancers where there are known risk factors, such as melanoma, which fell slightly more into the “bad luck” camp, despite its strong association with sun exposure. Basal cell skin cancer on the other hand — also related to sun exposure — did fall well into the environmental side of the equation.
Another example was esophageal cancer, which was like melanoma, more of the “bad luck” variety, despite its well-known associations with alcohol, tobacco, and obesity. Some critics latched on to this fact to discredit the findings, ignoring the fact that as with much thought-provoking research, Vogelstein and Tomasetti’s work doesn’t explain everything we have observed. But not being perfect doesn’t make the theory completely wrong.
A theory based on computational research that fits a consistent mathematical pattern and suggests an answer to a question makes sense. It certainly contains some thoughts worth pondering, despite the absence of data on some key cancers and some inconsistencies noted above.
Then there was the zinger: For whatever reason, the authors used the words “bad luck” in the abstract for the paper, which imparted a sense that most cancers are in fact out of our control. It wouldn’t be hard to predict that the press would seize upon the phrase “bad luck” when it came to explaining the report to the public. What good reporter could ignore such a catchy phrase?
Almost none of them did — even in the news section of the same issue of Science. The fundamentals of the research got lost in translation. Whatever the reporters’ original intentions, “bad luck” ruled the headlines. Worse, it translated easily into headlines that 2 out of 3 of the cancers in the United States are out of our control. The real story was that the research explained about two-thirds of the differences in cancer rates between various organs — before considering any other influences, like tobacco. It did not say this explained how we develop two-thirds of all cancers.
So, those headlines were incorrect. They tried to convey a complicated scientific concept in a few words, completely losing the message of the research. It’s not the first time the public has been led down the wrong path, and it won’t be the last. But it emphasizes the point that we need to be careful about what the media reports, how it is reported, and more importantly, not to rely on headlines and short summaries as scientific fact.
Forget about tobacco, forget about diet and exercise, forget about excessive sun or tanning bed exposure: Your fate is sealed, and there is little you can do to change it. To make matters worse, in their report the authors had opined that preventive strategies should include screening for infrequent cancers — never mind that such screening has not been proven to be effective.
The uproar in the advocacy community was immediate. There was fear that the phrase “bad luck” would catch the attention of policymakers, whose attention spans are already swamped by the deluge of information and the difficult choices about resources (read: money) they would assign to research and programs on cancer prevention.
Lines were quickly drawn in the sand. Efforts were mounted to immediately and loudly counter the “bad luck” interpretation. Criticisms of the research became polemics on how wrong the authors were about the interpretation of their work and the impact their interpretation would have on public health. Years of messaging could become irrelevant.
The authors scrambled to clarify their intent, but damage had been done. Science jumped back into the fray in a follow-up story outlining what had happened, including comments about their own miscues. The lessons we can learn from this episode are many:
Don’t Assume Headlines are the Entire Message
Communicating science is difficult. Translating science into understandable language is even more difficult. Getting the correct message — along with the subtleties involved in research — into a headline or short article is a herculean task. Understand that there is more to the story.
Stick to the Science
Perhaps the most prominent lesson is that we should stick to the science in our reports and not use catchy phrases to summarize complicated concepts. Catchy phrases tempt too many to abuse.
Words Get Bent
Understand that when it comes to communicating the findings of complex science, our words may become distorted even when communicated by well-meaning journalists. We live in a sound-bite world in which ideas having depth and context are condensed at times to less than 120 characters. “Bad luck” fits nicely into 120 characters. Mathematical calculations, which define stochastic errors in cell division, do not.
Leave Editorial Comments to Editorialists
Our research and our words carry meanings we may not have intended. We need to write what the science shows. Editorializing beyond the science should be left to the editorialists. I suspect that the authors of this paper had no intent to disrupt our current infrastructure of cancer screening and prevention, but others clearly and quickly saw such a threat.
Let’s Talk About News Releases
And one final observation, perhaps a plea: News releases could do a better job of explaining the science and its relevance to our current knowledge base. Too often they become a vehicle for other intentions, which may not be clearly obvious to the typical person or reporter.
The debate over the implications and conclusions of this research will undoubtedly continue for some time. And that is as it should be in the scientific process. But this paper may be best remembered for the firestorm it engendered. That is something that was out of the authors’ control and probably not their intention.
Truly, this was an unfortunate example of some very “bad luck.”