New research from North Carolina State University shows that most online news stories about cancer contain language that likely contributes to public uncertainty about the disease - a significant finding, given that at least one-third of Americans seek health information online.
"Previous studies show that more than 100 million Americans seek health information online, and that their findings affect their health decisions," says Dr. Kami Kosenko, an assistant professor of communication at NC State and co-author of a paper describing the study. "But, while people facing uncertainty about cancer issues are likely to seek out additional information, we've found that there are features of the information they're seeking that may actually exacerbate the uncertainty."
"We found that nearly two-thirds of cancer news articles contain at least some uncertain terms - words or phrases that reflect probability or ambiguity rather than certainty," says Dr. Ryan Hurley, a senior lecturer of communication at NC State and lead author of the study. The researchers evaluated more than 800 news articles on cancer issues, ranging from prevention to diagnosis to treatment. The articles were found on Google News, Yahoo! News, CNN.com and MSNBC.com.
Specifically, the researchers found that uncertain terms were used most often in reference to cancer treatment. "If you are trying to find clarity about cancer treatment options, reading news articles online may actually confuse the issue further," Hurley says. For example, one news article said, "There is no evidence that adding chemotherapy right away helps, and it may even worsen patients' chances." Hurley explains that this sentence creates uncertainty for readers because it indicates a lack of information (no evidence) as well as ambiguity about treatment efficacy (may even worsen).
To measure the use of uncertain terms, the researchers developed a scheme that captures five specific "message features" that are theoretically related to uncertainty. These features are conflicting information, complex information, ambiguous information, having too much information and having too little information. The researchers assessed the cancer news articles to determine the extent to which each included one or more of the uncertainty message features.
"To this point, no one has developed a means of systematically identifying and quantifying uncertain terms," Hurley says. "We believe the scheme we've created could be applied to identify uncertain terms in any text, from news articles to advertisements."
The researchers plan to use the scheme in forthcoming research efforts, including the design of experiments that can help us understand how uncertainty in messages influences people and affects behavior.
AI models redefine TIL scoring in breast cancer but face challenges in real-world validation