Reuters Health quoted Eric Jacobs from the American Cancer Society saying, “The lower risk of colorectal cancer risk seen among bisphosphonate users in this study is intriguing….However, these results should be interpreted with caution and require confirmation by additional studies.” From that point on, the story took a cautious, objective look at the study’s findings. The HealthDay story, by contrast, quoted Dr. Gad Rennert, the study’s lead author, sounding effusive: “These [new] findings are meaningful because they point to a possible protective effect of this class of drugs being relevant to prevention of many different cancers. … This is [similar] to the effect that we and others have shown for [cholesterol-lowering] statins.” From there on, Rennert is allowed to pump up the benefits of these drugs and to sweep aside any concern about their risks or broad application to a large population. The HealthDay story did hit some of our marks, too, but, overall, Reuters Health did a better job putting this study in the proper context. In addition, you can’t judge the stories by their headlines. The Reuters headline, “Women on bone drugs have less colon cancer” is not an accurate overview of the study’s finding. The HealthDay story’s headline gets in right, “Popular bone drugs linked to reduced colon cancer risk.”
As this story notes, colorectal cancer is the third leading cause of cancer deaths in the United States and affects roughly 1 out of every 20 people. If there is already a relatively inexpensive preventative measure on the market, this could prove beneficial to millions of people. Because cancer is such a complicated disease, though, reporters need to be extra careful in explaining a study’s findings and its limitations. Current colon cancer screening involves tests, like a colonoscopy, meant to identify pre-cursor lesions called polyps before they go on to cancer or cancers early in their course before they spread. The goal is to prevent future death due to the cancer. This story addresses a different way to prevent cancer – by using a treatment, in this case a medicine that is used to treat osteoporosis, to prevent the development of the precursor polyp and thus the future cancer. So instead of trying to remove things before they cause trouble, this form of prevention is designed to decrease the development of the precursor or cancer in the first place. While appealing conceptually, it isn’t really known whether this actually works. And even if it actually works, questions remain about whether adding it to current prevention efforts like colonoscopies result in fewer cancer deaths and at what cost. It is important to realize that no expert is currently saying that if you take this class of bone medicines, you can stop the colonoscopy or other screening test. The concept is that this is an addition to our current efforts. Taking all of that into considertaion, this story was written with an appropriately cautious tone.
Unlike the HealthDay story, this one mentions cost. It says that the drug in question, Fosamax, “costs around $10 per month in the U.S.”
Although, of course, the real cost effectiveness would be determining costs in a population that had routine current screening plus the medicine. Then one could determine the relative benefits in terms of decreased cancer and the attendant costs of treatment.
The story only presented the benefits in relative percentage terms, saying, “Even after considering other factors tied to the disease — like aspirin or statin use and eating lots of vegetables — their risk was 59 percent lower than that of women who hadn’t taken the drugs.” Again, readers deserve to know the raw numbers here. The story should have said that 138 women in the non-cancer group took biophosphonates and 97 in the cancer group took them, meaning that 41 women appear to have benefited from taking the drugs. If the study had found that 97 women had taken these drugs in both groups, the study would not have made headlines. Also, it might have been helpful to show that there were actually larger differences in some of the other factors that were studied. For example, 120 women in the group of cancer patients had a history of colorectal cancer in the family compared to 75 women in the control group.
Neither story quantified the potential harms or any of the side effects found in the study itself. But this story did say, “”The long-term effects are less well-known, however, and in rare cases it can cause bone death of the jaw, which would be important if healthy people were to take it.” It also was the only story to mention that an “earlier study from the UK had found no link between bisphosphonate and colon cancer. Indeed, it found a higher risk of throat cancer in patients on the medication.”
The HealthDay story started out with effusive quotes from the lead researcher. The Reuters story right away established that the findings were being met with criticism. The story says in the second sentence, “The finding has them excited about the prospect of using the drugs — called bisphosphonates — to help prevent cancer in healthy people, but other experts are less enthusiastic.” The story then goes on to do a better job explaining the particulars of the study design. It says, “The new work looked at 933 women with colon cancer, whose average age was just over 70. The researchers then found a comparison group of women without the disease, who matched the first group in age, ethnicity and clinics where they received treatment.” Unlike the HealthDay story, this story made note of a similar study that did not show the same level of benefit. It says, “Jacobs of the American Cancer Society added that one earlier study from the UK had found no link between bisphosphonate and colon cancer. Indeed, it found a higher risk of throat cancer in patients on the medication.” The big problem with both stories, though, is that they both fail to point out that this study did not look at 933 women who took the drugs and 933 who did not. Of those 1,866 women, only 97 of the cancer patients took the drugs, and only 138 of the women without cancer had taken them. So, yes, the researchers were able to find what appears to be a statistically significant risk reduction, but they also found differences in physical activity, body mass index, vegetable intake and the use of statins and aspirin. The study highlights the biophosphonate connection, but the story should have explained why a difference of 41 women without cancer taking these drugs out of a total of nearly 2,000 would be enough evidence to get really excited about.
Also, evaluating the quality of the evidence requires acknowledging this type of study can only assess relative risk. It cannot provide an assessment of the absolute benefit of using the medicine vs. not using it. This is a key limitation of the study design. This type of study can only identify if a relationship may exist between the medicine and colon cancer. It can’t say how big the effect is. That requires a different type of study. So evaluating the quality of the evidence involves acknowledging the relationship between the medicine and less cancer. This study cannot determine whether this relationship is “real” or truly measure the magnitude of the benefit. Here is where both stories fell flat by not discussing the huge limitations of trying to control for differences between women who took the medicine and those who didn’t. We know that people who engage in preventive behaviors such as taking a bisphosphonate may also engage in other healthy behaviors that may relate to lower risk of cancer. So it may not be the bisphosphonate but rather another factor not controlled for that distinguishes patients who took the medicine from those who didn’t. That is the real factor that influences the quality of the evidence. Clearly the Reuters story did a better job here, but it too missed this point.
Neither story engaged in disesae mongering, but the Reuters story did a better job of explaining the scope of the disease. It said, “According to the American Cancer Society, one in 19 men develops colorectal cancer at some point, and slightly fewer women do. The disease is the third leading cause of cancer deaths in the U.S.”
Both stories only quoted one outside source: Eric Jacobs of the American Cancer Society. But the Reuters Health story made better use of his comments and did not allow the lead author to go unchallenged with some of his more sweeping comments.
Both this story and the HealthDay story ended with Jacobs from the American Cancer Society helping put the findings in context. In this story, he made a more forceful statement, saying “Based on current evidence, bisphosphonates should not be used for prevention of colorectal cancer.” Then, he said the same thing that he said in the HealthDay story: “Fortunately, there are proven ways to help prevent colorectal cancer. In particular, all Americans, 50 or older, should get a screening test so that precancerous polyps can be detected and removed before they turn into cancer.”
This story is clear on the drugs’ availability. It says that the drug studied was “mainly the drug alendronate (Fosamax)”.
This story made it clear that the findings are part of a series of findings, some of them contradictory, about the side benefits of different drugs for cancer prevention.
The story didn’t rely on a news release.