Study details
For the study, reported by Johnson et al., two National Comprehensive Cancer Network panel members were selected as content experts for each of the four cancers and were tasked with reviewing the primary medical claims in each article. The experts then completed a set of ratings to arrive at the proportion of misinformation and potential for harm in each article.
Of the 200 articles, 41.5% were from nontraditional news (digital only), 37.5% were from traditional news sources (online versions of print and/or broadcast media), 17% were from medical journals, 3% were from a crowdfunding site, and 1% were from personal blogs.
This expert review concluded that nearly one-third of the articles contained misinformation, as noted above. The misinformation was described as misleading (title not supported by text or statistics/data do not support conclusion, 28.8%), strength of the evidence mischaracterized (weak evidence portrayed as strong or vice versa, 27.7%) and unproven therapies (not studied or insufficient evidence, 26.7%).
Notably, the median number of engagements, such as likes on Twitter, for articles with misinformation was greater than that of factual articles (median, 2,300 vs. 1,600; P = .05).
In total, 30.5% of all 200 articles contained harmful information. This was described as harmful inaction (could lead to delay or not seeking medical attention for treatable/curable condition, 31.0%), economic harm (out-of-pocket financial costs associated with treatment/travel, 27.7%), harmful action (potentially toxic effects of the suggested test/treatment, 17.0%), and harmful interactions (known/unknown medical interactions with curative therapies, 16.2%).
The median number of engagements for articles with harmful information was statistically significantly greater than that of articles with correct information (median, 2,300 vs. 1,500; P = .007).
A limitation of the study is that it included only the most popular English language cancer articles.
This study was funded in part by the Huntsman Cancer Institute. Dr. Johnson, Dr. Lycette, and Dr. Southwell have disclosed no relevant financial relationships. Some study authors have ties to the pharmaceutical industry.
A version of this article first appeared on Medscape.com.