This story says many of the right things about a new study indicating that daily low-dose aspirin appears to lower the risk of death from several common types of cancer. It quotes a researcher who said that the findings should not be viewed as reason for everyone to go out and starting taking aspirin. Unfortunately though, the story never explains to readers why the results might not provide a strong enough foundation for such a recommendation, so the call for restraint lacks sufficient heft when compared with enthusiastic description of benefits which were based entirely on relative risk comparisons. The potential for harm was mentioned only the last line of the story and never quantified. This is particularly worrisome considering that the prevention benefit is not realized for many years. Some of the other stories we reviewed gave more weight to the limitations of the data, included more independent voices and a more thorough discussion of potential harms.
Recommendations that affect the public health should ideally be based on the gold standard of evidence: randomized controlled trials. And while the data discussed in this story come from studies which appear to meet this standard, it is important to recognize that these trials were originally designed to look at heart disease, not cancer. The cancer analysis was conducted after the fact and provides a more limited basis upon which to draw conclusions. The findings are more likely to be biased and provide less information about the harms of long-term aspirin use than a trial that was designed to look at cancer from the start.
To be sure, the study discussed in this story was well conducted and provides some of the best quality evidence we’re likely to see on this question for many years to come. And this topic is important in general because aspirin is one of the only agents we have that may be effective for prevention of cancer. Although not harmless, it is easy to take for most patients, and with the benefits for heart disease and stroke, is an important tool for disease prevention. Because of limitations in this study, however, we don’t have all the evidence we would wish for when recommending a health intervention to the entire population.
The cost of aspirin is not in question. Nevertheless, a broad recommendation to begin taking aspirin daily in middle age would likely be followed by at least hundreds of millions of people, thus the cumulative costs would be large. Also, such widespread use would increase the costs of treating bleeding ulcers and other adverse events caused by aspirin, though hoped-for savings in reduced cancer treatment would be an offestting factor to consider. Still, because the cost of aspirin on an individual basis is low and well-known we won’t insist on this point.
The use of relative risks throughout this story inflates the likely benefits of aspirin. For example, the story states that overall cancer death risk “plummeted by 21 percent” among those taking low-dose aspirin while the studies were still underway. But in absolute terms, the rate of cancer death was 2.3% (327 deaths out of 14035 people) in the aspirin group and 3% (347 out of 11,535) in the placebo group, for a reduction of 0.7%. This is statistically significant difference and may have important implications for the overall public health, but the benefit that this represents for an individual is certainly much more limited than what the 21% figure suggests — especially when one factors in the uncertain risk of adverse effects. To be fair, we had to search the original study pretty thoroughly and make our own calculations to determine this absolute risk. Nevertheless, we expect stories to go the extra mile to make statistics meaningful for the decision-making of individual readers.
Potential side effects of aspirin mentioned in the story — “bleeding and stroke” — are relegated to the last line of the story and are given nowhere near the prominence they deserve. For one thing, “bleeding” doesn’t convey the seriousness of potentially fatal gastrointestinal hemorrhage that can result from daily aspirin. Second, the story didn’t make any attempt to quantify these risks so that they can be balanced against benefits.
Not bad, but not quite good enough for a satisfactory. The story cautions readers that the findings can’t be considered definitive, and it even quotes a researcher who states flatly that “these results do not mean that all adults should immediately start taking aspirin.” However, such admonitions ring hollow without a more substantive explanation of why the findings may not be reliable. Importantly, although the data used in the analysis come from randomized controlled studies, the trials were originally designed to look at heart disease, not cancer. Researchers in the current study went back after the fact and tried to determine the effects of aspirin on cancer deaths many years after the studies were completed. The retrospective nature of the study means the findings are more likely to be biased than a study that was designed and initiated from the start to look at cancer outcomes.
To the story’s credit, it included the total number of patients and range of study durations, and noted the fact that the studies were not originally meant to study cancer. However, the discussion of limitations was not detailed or prominent enough to provide an appropriate counterweight to the enthusiastic description of benefits.
The story did not exaggerate the impact of cancer and it accurately described the possible relevance of this study to a large segment of the population.
The story includes comments from an independent cancer expert. That’s the most important factor in this case, so we will pass this story on this criterion. However, while the study itself did not receive outside funding, readers should have been told that some of the authors have had consulting relationships with pharmaceutical companies.
Some of the other news stories about this study pointed out that avoiding smoking and obesity are known to reduce cancer risk. This story should also have pointed out these options, which also do not have potential adverse effects.
The availability of aspirin is not in question. The story would have been better if, like some of the other stories we reviewed, it had pointed out that aspirin is approved for use as an anagesic and, for certain people, to reduce the risk of some cardiovascular events, but it is not currently recommended for cancer prevention.
The story did not try to oversell the novelty of using aspirin to prevent cancer. However, it should have provided more discussion of the extensive body or previous research on aspirin and cancer. Previous trials and observational studies have come to conflicting conclusions about the benefits of aspirin for cancer prevention, and this new study needs to be viewed in the context of this existing research.
The story acknowledges taking some quotes directly from a news release about the study, and a number of passages bear a resemblance to the text of the release. However, the story does also solicit feedback from an independent expert not affiliated with the study. Borderline, but enough for a satisfactory.