Email this page to a friendShare on facebookShare on twitterBookmark & SharePrinter-friendly versionSubscribe to RSS
Some nutrition and diet studies may overstate results
Thursday, October 31, 2013
By Shereen Jegtvig
NEW YORK (Reuters Health) - Doctors, policymakers and everyday people may make decisions or give advice based on the results of published nutrition studies. But a new analysis shows researchers sometimes overstate the results of those reports.
Researchers from the University of Alabama at Birmingham looked at papers published about nutrition and obesity in leading medical and public health journals. They tracked how often authors overreached in the summary of their findings.
"We found that about one of 11 studies have some kind of issue that we identified that was degrading the fidelity of research reporting," Dr. Nir Menachemi said.
"In the article we call it an overreaching statement. That's probably the most fair way to characterize these infractions," Menachemi, who led the analysis, told Reuters Health.
His team's findings were published in the American Journal of Preventive Medicine.
Not much is known about how often findings are distorted by scientists or the media.
Results from so-called observational studies - which can't prove cause-and-effect - are often used to make potentially inappropriate recommendations without better data, the researchers said. One study showed half of press releases on more rigorous trials also overstated their results.
Menachemi and his colleagues wondered if researchers might be more likely to overstate the results of studies on politically and socially charged topics. They decided to focus on nutrition and obesity reports.
The researchers looked for articles published in eight leading journals in either 2001 or 2011 to examine changes in reporting over time. They found 937 papers, 377 from 2001 and 560 from 2011.
Close to nine percent of those had findings that were overstated in the study summary, called an abstract. Studies from 2011 were more likely to overreach than 2001 papers.
The overreaching statements included inappropriately describing a correlation as a cause-and-effect relationship and generalizing a study's claims to large groups of people when the study population was quite different.
Although those overstatements may be unintentional, they can distort what doctors, policymakers and the general public know about nutrition, the researchers said.
Unfunded studies had more problems than funded studies, regardless of what type of group paid for the study, they found.
"Whether it was a private foundation or government entity or for-profit organization, funded studies generally had in our study, less overstatements of results by authors," Menachemi said.
If researchers are overstating their results, the media is probably overstating them too, he added.
"I think it's overwhelming for the average person out there who is bombarded with so much information and frequently he or she doesn't always have the scientific tools available to be able to determine how much of it do I really need to take in and change my practices and how much of this is just interesting information," he said.
Overstating study results can cause problems for professionals as well as the general public, researchers said.
"I think the study was very enlightening, for sure, and frankly I think the authors have brought attention to an issue that needs to be addressed both within the scientific community as well as within the media," dietician Joy Dubost told Reuters Health. Dubost is a spokesperson for the Academy of Nutrition and Dietetics and wasn't involved in the new study.
"We all have a role to play in ensuring the consumers are getting that evidence-based science to really create trust in their minds," she said.
Dubost offered suggestions for people who read stories about nutrition research. "Be leery of anything that sounds like a quick fix, or results that are just based on animal studies, or if it sounds too good to be true, it probably is," she said.
Source: http://bit.ly/18DtbtE American Journal of Preventive Medicine, November 2013.
(c) Copyright Thomson Reuters 2013. Check for restrictions at: http://about.reuters.com/fulllegal.asp