Online comment sections may influence readers' opinions on health issues

In this study led by Holly Witteman, an assistant professor in the Faculty of Medicine at Université Laval, nearly 1,700 participants were asked to carefully read a mock news article on home birth. The mock article was a composite of real news articles from various U.S. publications. "We took paragraphs from each source, including quotes from health care professionals who were for or against home birth in order to create a balanced news item," explained professor Witteman, who is also a researcher at Centre de recherche du CHU de Québec-Université Laval.

The article was followed by ten social media comments that had been posted on the original articles. To reproduce the one-sided opinions often found on social media, the researchers varied the comments viewed by the study participants. Some participants viewed comments that were unanimously in favour of home birth, while others read only negative comments. To assess the effect of personal narratives, the researchers also randomly assigned study participants to read either comments that contained personal stories about home birth or comments that had no such personal accounts. Another group of participants viewed comments representing a breadth of views, and a control group read the article without any comments.

After reading the articles and comments, participants were asked to state their opinion of home birth on a scale of zero to one hundred (extremely negative to extremely positive). Although all the participants read the same article, their opinion on the subject was influenced by the nature of the comments. Participants who viewed balanced comments and those who read the article without comments expressed an average opinion of 52, while the average opinion for the negative comments group was 39 and the average opinion for the positive comments group was 63. Comments with personal stories increased the divide.

"However, this doesn't mean we should shut down comment sections or attempt to suppress personal stories," says Witteman. "If sites fail to host such discussions, they are likely to simply happen elsewhere. Although the quality of comments is sometimes debatable, social media is a valuable tool that allows people to share and find information on subjects related to their health. That kind of engagement is arguably a good thing. What's more, sharing information can prove particularly useful when there is no consensus on the topic in the scientific community or if a person's choice comes down to their values or personal preferences."

Professor Witteman believes that this study reveals, above all else, the potential danger of polarized discussions. "Organizations that seek to communicate health information and to support discussions of that information may wish to ensure they adequately represent different viewpoints so that readers can form their own opinions," says Witteman. "In some cases, it may even be reasonable to allow for expressions of unconventional ideas, accompanied by respectful responses clarifying the facts. That's why we recommended in our paper that health websites with social media features have a budget for a social media community manager and for content experts like healthcare professionals or medical librarians who can explain the nuances of evidence and answer questions raised by readers."