Inaccurate, harmful cancer information prevalent on social media
When his wife was diagnosed with cancer, Skyler B. Johnson, MD, did what many concerned spouses do. He consulted Google.
A medical student at the time, Johnson had one advantage many panicked Googlers lack —he could recognize bogus information when he saw it.
Unfortunately, it was everywhere.
“I came away from that experience with quite a bit of empathy for patients,” Johnson, a physician-scientist at Huntsman Cancer Institute and assistant professor of radiation oncology at University of Utah, told Healio. “I encounter a lot of patients through my oncology practice who believe misinformation they read online and through social media, and it has led to delays in their diagnosis and treatment. It has led to refusal of proven cancer treatments and, ultimately, their early deaths.”
Now, Johnson — whose wife is doing well — is shining a light on the pervasiveness of the problem.
He and his colleagues at Huntsman conducted a study, published in July in Journal of the National Cancer Institute, that showed one-third of the most popular cancer treatment articles on social media include inaccurate information. Moreover, most of the misinformation identified had the potential to harm patients by advocating approaches that could have a negative effect on treatment quality and survival.
“If you talk to most oncologists, I think they will tell you that this is one of the more upsetting encounters we have with patients — especially for people with curable cancers,” Johnson said.
Misinformation gets attention
For the study, members of Johnson’s team reviewed social media posts and identified 50 of the most popular articles on each of the four most common cancers (breast, colorectal, prostate and lung).
For all 200 articles, they reported the extent of misinformation and potential for harm and calculated their association with the number of social media engagements using a two-sample Wilcoxon rank-sum test.
The researchers found that of the 200 articles, 37.5% (n = 75) were from traditional news sources (online versions of print/broadcast media); 41.5% (n = 83) were from non-traditional news sources (digital only); 1% (n = 2) were from personal blogs; 3% (n = 6) were from crowdfunding sites; and 17% (n = 34) were from medical journals.
Review of the articles by content experts, including two National Comprehensive Cancer Network panel members for each cancer type, showed 32.5% (Cohen’s kappa coefficient = 0.63; 95% CI, 0.5-0.77) included misinformation, most often characterized as misleading (title not supported by text, statistics/data that does not support conclusion, 28.8%), having misrepresented strength of evidence (weak evidence described as strong or vice versa, 27.7%) or involving unproven therapies (not studied, insufficient evidence, 26.7%).
Overall, 30.5% (Cohen’s kappa coefficient = 0.66; 95% CI, 0.52-0.8) of articles included information the researchers deemed harmful. They categorized these as harmful inaction (could lead to delay or refusal of medical attention for treatable/curable conditions, 31%) economic harm (out-of-pocket financial costs associated with treatment/travel, 27.7%), harmful action (potentially toxic effects of the suggested test/treatment, 17%), and harmful interactions (known/unknown medical interactions with the curative therapies, 16.2%). More than three-quarters of articles containing misinformation (76.9%; n = 50 of 65) included harmful information.
The researchers also found that articles containing misinformation generated a higher number of total engagements (median, 2,300 vs. 1,600; P = .05) than articles with evidence-based data.
Johnson emphasized that addressing and correcting social media misinformation is more important now than ever, given the “infodemic” of inaccurate content around COVID-19 and its available vaccines.
“I think what makes this study important is that we know that patients are increasingly using the internet as a source of health information,” he said. “It’s absolutely front and center.”
Much of the misinformation on social media advocated forgoing conventional cancer treatments in favor of “alternative” treatments such as cannabis, which was especially popular during the span of the study, Johnson said.
Additionally, much of the content warned that evidence-based treatments such as chemotherapy are ineffective or harmful, and that conventional medicine should not be trusted.
These social media posts may hold appeal because they align with individuals’ preexisting, deeply held false beliefs, Johnson said.
“There is an interest that surrounds feeling like you are entitled to this unique and rare information you won’t find in traditional medicine,” he said, “but it’s often not accurate, and that is the larger concern.”
Social media algorithms may contain “information silos” that amplify this type of false information for users who might be drawn to it, he added.
“We didn’t evaluate that directly, but what we do know is that the articles that contained misinformation and harmful information were shared more frequently and commented on more frequently than articles that were factual and safe,” he said.
A previous study by Johnson and colleagues evaluated the correlation between use of alternative medicine vs. conventional cancer treatment and survival.
“We found that patients who select unproven cancer treatments have between two-and-a-half to almost six times the risk [for] death,” he said.
Reducing the effect of harmful social media misinformation will require several components, one of which is patient education, Johnson said.
“One of the goals of future research is to identify individuals at risk for encountering and believing misinformation,” he said. “We can then try to create interventions that might help modify their behavior.”
Johnson said he also hopes to identify predictors of misinformation and alert patients to this potential harm.
“If we can identify the predictors of misinformation, we can help educate patients,” he said. “Ultimately, this type of information would be helpful for companies like Facebook and could help inform their future in terms of health policy and managing this information.”
Clinicians also may benefit from being more aware of social media misinformation.
“We know that our patients are going to encounter this misinformation and, as providers, we need to know how to help them navigate it,” he said. “There are a lot of arms to this type of research, but it’s all aimed at improving survival. It’s very unfortunate that anyone could be misled such that they would compromise the quality of their cancer care and, ultimately, compromise the length of their life.”
Although addressing social media misinformation at the policy level is one important component of eradicating it, Johnson emphasized that clinicians also can combat misinformation at the patient level by simply acknowledging it, understanding it and confronting it directly.
“As a medical community, we can’t put our heads in the sand and pretend that our patients won’t encounter cancer misinformation or assume that they’ll ignore it. We need to address the issue head on by working with our patients, informing them of the types of misinformation they will encounter, and by giving them accurate sources of information that they can seek out,” Johnson said. “Most importantly, we need to empathize with our patients, establish trust, and work with them as they make challenging and often life-altering decisions.”
For more information:
Skyler B. Johnson, MD, can be reached at Huntsman Cancer Institute, 1950 Circle of Hope, Salt Lake City, UT 84112; email: email@example.com.