Outbreaks of science myth-busting can be a bit of a puzzlement.
The science behind a popular headline-maker might be a tottering house of cards, but it can be impressively sturdy nevertheless. New studies might re-arrange it a little here or there, but it doesn't usually topple. Only those studies that reinforce the narrative seem to get attention.
Then a new study lands. It might not even shift the balance of knowledge mightily. Yet, something about it in particular gets attention. And the issue tips over into the myth-busting zone.
It happened recently with resveratrol. An epidemiological study had that kind of outsized impact, thanks to stories at the Washington Post and National Geographic.
Resveratrol is a polyphenol - an antioxidant found in many plants. There's a lot of it in grape skin. The substance provides fuel for the claim that red wine is good for you.
It's a particularly interesting case. Resveratrol has been the subject of a lot of serious study since the 1990s - and simultaneously on the Quackwatch list most of that time.
The jury's still not in on hypotheses about resveratrol - more on that later. Nevertheless, for the supplement industry, it's long been "the story of the little ingredient that could - and did."
And boy, did it: "Media hype has helped to keep consumer sales of products containing resveratrol growing...[S]ales of resveratrol dietary supplements alone hit $30 million in 2008." Add to that a range of other ways people buy it - in water and other beverages, and, as befits something regarded as a fountain of youth, skin cream, too.
It's not just media hype behind resveratrol, though. Journalists don't simply spin this out of whole cloth. And they're not taking their cues from manufacturer promotion, either. Or at least, mostly they're not. The aura around resveratrol begins with scientific literature.
Science-based myths are woven out of strands of hypotheses that are still far from substantiated, but have logically appealing narratives.
While people know the drill about logical fallacies - like mistaking correlation for causation, or reading too much into animal studies, they still succumb.
People may know each one of these cards on its own is not directly probative, but once there's a tower of them, it seems to become circumstantial evidence that's convincing.
The premature conviction can spread quickly. "Scientific papers are written to be utterly convincing; over the centuries their special language and style has been developed to make them read convincingly," writes sociologist Harry Collins.
There's a tendency, too, once lots of people are researching in an area, to have a "where there's smoke, there must be fire" response. If there are so many studies, and so many people are talking about "known benefits" there must be something to it, right?
Not necessarily. Poorly conducted studies and the common problems of data misinterpretation can give the impression that a lot is known for sure, when it's not. As Matt Briggs put it, false positives "can behave like rumors: easy to release, but hard to dispel." And according to John Ioannidis, lots of teams working in the same field might increase the chances of false positives if people race each other.
Once there's a lot of research activity, you can't get a handle on the sum of knowledge without a research project to systematically review what's known, sifting through to sort out wheat from chaff.
Here's a look at what recent systematic reviews tell us about resveratrol:
- Longevity: there are few animal or plant species in which resveratrol has been shown to extend life, and no solid proof was found that it can do so in humans (Hector 2012);
- High blood pressure: the evidence from the first trials is not showing a major impact from resveratrol supplements (Liu 2013);
- Diabetes: resveratrol supplements may have an impact on glucose and insulin measures in people with diabetes (although it's too soon to know about its impact on health outcomes), but not in people without diabetes (Liu 2014).
There are two more systematic reviews that don't appear to have as much methodological rigor as these three. They both conclude there isn't good evidence of benefit in people, too. One group looked for any benefit, including cancer prevention (Vang 2011) and another researcher looked at reducing blood lipids (fat) (Sahebkar 2013).
There's plenty of reason to keep researching hypotheses about resveratrol supplementation. But there is no strong case yet for a benefit - and the adverse effects that have been found in humans and animals after taking resveratrol supplements still need to be studied further, too.
What about the red wine hypothesis? The idea that red wine is healthy alcohol may be the biggest impact of the resveratrol hype. That's a discussion for a future post, with signs that there is more harm than we've realized from even light drinking. If it has encouraged us to drink more, then some of us could end up with more than a hangover from this episode.
More from Cecil and Harvard's resveratrol studies' "strictly guarded mouse lab" (!) ... A post on why results from animal studies usually fail to translate to humans from me at Statistically Funny.
If you're interested in Harry Collins' wonderful book, Are We All Scientific Experts Now?, here's a review by William Allen at LSE's The Impact Blog, and another by Athene Donald in Times Higher Education. Maggie Koerth-Baker interviewed Collins at Boing Boing a few years ago.
If you're interested in more about the science of analyzing bodies of research, here are some of the related Absolutely Maybe posts: 5 Key Things to Know About Meta-Analysis and Biomedical Research: Believe It Or Not?
The cartoons are my own (Creative Commons License): more at Statistically Funny.
The image for data on publications mentioning resveratrol comes from the results per year graph at PubMed.
* The thoughts Hilda Bastian expresses here at Absolutely Maybe are personal, and do not necessarily reflect the views of the National Institutes of Health or the U.S. Department of Health and Human Services.