Cancer Misinformation Online and the Evidence on Choosing Alternative Medicine

In one US review of the most-shared cancer articles on social media, about a third contained misinformation, and articles with misinformation or harmful content drew more engagement than factual or safe ones^[1]. In a separate US registry study, people with non-metastatic cancers who were recorded as choosing alternative medicine instead of conventional treatment had two and a half times the hazard of death over the follow-up period^[2].

Almost everyone affected by cancer goes looking for information, often late at night and often on a phone. They are looking for hope, for control, for something that makes sense of a frightening time. When the most widely shared material is also the least reliable, the problem lies with an information system that rewards attention over accuracy, not with the person scrolling. This study sets out what research has measured about how common cancer misinformation is online, what is known about outcomes for people who chose alternative medicine in place of conventional treatment, and what the evidence says helps.

About a third of popular cancer articles online contained misinformation, and it travelled further

The clearest cancer-specific measure comes from a study published in the Journal of the National Cancer Institute. Researchers collected the 50 most popular social media articles for each of the four most common cancers, 200 articles in all, and two cancer experts rated each one. Of the 200, 32.5% contained misinformation and 30.5% contained information the experts judged potentially harmful. The two overlapped heavily: 76.9% of the articles with misinformation also contained harmful information^[1].

The more troubling finding was about reach. Articles containing misinformation or harmful information attracted more social media engagement than factual or safe ones. The difference was statistically significant for harmful content (P = .007) and borderline for misinformation (P = .05)^[1]. The authors measured engagement, not who read the articles or what they did next, so the study cannot say how many decisions were changed. It does show that, in this sample, popularity was no guide to accuracy.

Video shows a similar pattern. A study of the first 150 YouTube videos on prostate cancer screening and treatment found that 115 (77%) contained potentially misinformative or biased content in the video or its comments. The videos had up to 1.3 million views each, an average of 45,223, and there was a significant negative correlation between scientific quality and viewer engagement^[3]. The comment sections under some videos carried advertising and peer-to-peer medical advice.

Cancer is not unusual here. A systematic review of 69 studies of health misinformation on social media, published in English before March 2019 and covering a range of platforms, found it was most common in studies of smoking products and drugs, reaching 87% of posts in some studies, and very common for vaccines (43%). Studies on diseases (noncommunicable diseases and pandemics) reported moderate misinformation rates of around 40%, especially in the case of cancer. Prevalence was highest on Twitter^[4]. The World Health Organization describes the wider problem, in the context of disease outbreaks, as an "infodemic": too much information, including false or misleading information^[5].

People recorded as choosing alternative medicine alone had two and a half times the hazard of death

The most-cited study of outcomes is by Johnson and colleagues, published in the Journal of the National Cancer Institute in 2018. It is worth being precise about what it did and did not do^[2].

• Design: a retrospective observational cohort study. Nobody was randomised; the researchers looked back at existing records.
• Data: the US National Cancer Database, covering people diagnosed between 2004 and 2013.
• Population: 281 people with non-metastatic breast, prostate, lung or colorectal cancer whose records showed treatment coded as "other-unproven: cancer treatments administered by nonmedical personnel", and no conventional cancer treatment.
• Comparator: 280 of them were matched 2:1 with 560 people who had conventional treatment.

Over the follow-up period, the alternative-medicine group had a hazard ratio for death of 2.50 (95% confidence interval 1.88 to 3.27) compared with the matched conventional-treatment group. Put plainly, at any point during follow-up, the rate of death was about two and a half times higher. Five-year survival was 54.7% in the alternative-medicine group (95% CI 47.5% to 61.3%) and 78.3% in the conventional-treatment group^[2].

The gap differed by cancer. The hazard ratio was 5.68 for breast cancer, 4.57 for colorectal cancer and 2.17 for lung cancer. For prostate cancer it was 1.68, with a confidence interval of 0.68 to 4.17 that crosses 1, so the study could not show a statistically significant difference for that cancer^[2].

The people in the alternative-medicine group were not who stereotypes might suggest. They were more likely to have breast or lung cancer, higher socioeconomic status, stage II or III disease and few other health conditions, and to live in the Intermountain West or Pacific regions of the US^[2]. The authors are clear about the limits: the study is observational, conventional treatment given elsewhere may have been missed, there may be unmeasured confounders or selection bias, and the database holds no information on which alternative therapies were used. The 281 people were also a small and specific group, so the hazard ratio describes them, not everyone who uses alternative approaches. For how hazard ratios and five-year survival should be read, see what survival statistics mean.

The higher risk appears to run through declined treatment, not through using extra therapies

A follow-up study by the same group, published in JAMA Oncology in 2018, asked a different question: what about people who used complementary medicine alongside at least some conventional treatment? It drew on the same US National Cancer Database, a cohort of 1,901,815 people with cancer diagnosed between 2004 and 2013, and compared 258 people recorded as using complementary medicine with 1,032 matched controls^[6].

The complementary-medicine group was far more likely to decline some part of the recommended conventional treatment. Among them, 53.0% were recorded as refusing radiotherapy (compared with 2.3% of controls), 34.1% chemotherapy (3.2%), 33.7% hormone therapy (2.8%) and 7.0% surgery (0.1%)^[6].

Before accounting for this, complementary-medicine use was linked to a higher risk of death, with a hazard ratio of 2.08 (95% CI 1.50 to 2.90). Once treatment refusal and delay were added to the model, there was no significant association between complementary medicine and survival (hazard ratio 1.39, 95% CI 0.83 to 2.33). The authors concluded that the higher mortality was mediated by refusal of conventional treatment^[6].

This matters for how the subject is discussed. The evidence does not show that using complementary approaches alongside full conventional treatment is itself what is linked to worse outcomes, although the adjusted confidence interval (0.83 to 2.33) is wide, so a smaller association cannot be ruled out. The association the researchers found runs through conventional treatment being declined or delayed. Like the earlier study, this one is observational, and registry records cannot show the reasons behind anyone's decisions. Those reasons are personal, and the evidence gives no grounds for judging them.

Corrections do work, especially when experts make them

The encouraging finding is that misinformation can be countered. A meta-analysis of 24 studies with 6,086 participants tested attempts to correct health misinformation on social media. On average, corrections reduced belief in misinformation with an effect size of d = 0.40 (95% CI 0.25 to 0.55), a small-to-moderate effect by the usual convention^[7]. Corrections were more effective when made by experts rather than non-experts, when people were personally involved with the health topic, and when the misinformation had come from news organisations rather than from peers^[7]. The authors could not rule out publication bias, and the abstract does not say how long the effect lasted.

People affected by cancer are using social media to make real decisions. In a US national online survey of 262 people with cancer and their caregivers, carried out in November and December 2021, 65% said they were likely or somewhat likely to use social media to make at least one kind of decision, such as about lifestyle, screening, vaccination, treatment, testing or choosing a provider^[8]. Among those who were, 42% said they were uncomfortable talking to their provider about information they had found on social media, and people without a college education were less comfortable than those with one. The authors call for a supportive environment that gives people dependable ways to verify what they find^[8].

A 2024 review in CA: A Cancer Journal for Clinicians reached a similar view. It argued that cancer experts should counter misinformation in real time and put out evidence-based content where people already are, that health professionals should give "information prescriptions" to patients and families and support health literacy, and that the response needs engagement from media platforms as well as from scientists and clinicians^[9].

Corrections also have limits. A meta-analysis of 52 studies with 6,878 participants, across topics and not limited to health, found large effects both for debunking and for the persistence of misinformation after it had been debunked. Persistence was stronger, and debunking weaker, when people had generated their own reasons for believing the original claim. More detailed debunking messages were associated with a larger debunking effect and, unexpectedly, with stronger persistence of the misinformation as well^[10].

What the evidence asks of us

The research points away from blaming people who are misled and toward the systems around them. Taken together, the studies suggest five priorities.

• Treat misinformation as a health-system problem. WHO describes infodemic management as the systematic use of risk- and evidence-based analysis and approaches, built on listening to community concerns, explaining risk and expert advice, building resilience to misinformation and engaging communities^[5]. Cancer plans and cancer charities can apply the same approach to cancer information, not only to outbreaks.
• Put trusted expertise where people look. Expert corrections worked better than corrections by non-experts^[7], and reviewers argue that cancer experts should share evidence-based content and respond in real time^[9]. That needs funded time and support for clinicians, nurses and charity information teams, not goodwill alone.
• Make it easy to check. Many people affected by cancer use social media to make decisions, more than four in ten of those who do are uncomfortable raising it with their care team, and people without a college education feel least able to^[8]. Services can routinely invite questions about things people have read, without judgement, and offer "information prescriptions" pointing to reliable sources^[9].
• Ask platforms to stop rewarding harm. When misinformative and harmful content draws more engagement than accurate content^[1][3], the incentive problem lies with how content is ranked, and the response needs platforms at the table^[9].
• Fund better research outside the US. The outcome studies here come from one US database and a few hundred people. We did not find comparable registry studies of alternative medicine used in place of conventional treatment from other countries, so the global picture is largely unmeasured.

For the evidence on common fears about what causes cancer, see what does not cause cancer. For the full collection, see the Cancer research library.

Language: this piece says "people with cancer" and "people living with and after cancer", and describes the illness and its treatment plainly, without combat metaphors. Clinical and statistical terms such as "stage", "hazard ratio" and "five-year survival" are used in the sense the cited source defines them.

Why we're publishing this

People affected by cancer deserve information they can trust, and the charities, nurses and support groups who answer their questions deserve evidence they can point to. Misinformation is often discussed with a sneer. The research supports a different response: people turn to unreliable sources for understandable reasons, the most popular content is often the least accurate, and well-made corrections from trusted experts do change minds. We have set out what the studies measured, and only that, so that organisations can cite it with confidence.

How to read this data

The prevalence figures come from content analyses: researchers sampled articles or videos and experts rated them. Results depend on the sample, platform, date and definition of misinformation, so they cannot be added together or read as a trend. The outcome figures come from two observational studies of the US National Cancer Database, people diagnosed between 2004 and 2013. A hazard ratio compares the rate of death over time between two groups; it is not the same as a relative risk or an odds ratio, and an association in observational data is not proof of cause. Both studies relied on how treatment was coded in hospital records, which may be incomplete, and neither could see which therapies were used or why. The correction evidence comes from two meta-analyses whose abstracts give average effects across health topics, or across topics generally, not results for cancer specifically. Most of the evidence here is from the US; cancer-specific data from other countries are thin.

Use this data

Free to cite with attribution to Better Health For Everyone. Charities, campaigners and journalists are welcome to reuse these figures and charts. For the underlying figures as a spreadsheet, get in touch.

Important: This article summarises published research for general information. It is not medical advice and cannot be used to diagnose, treat or predict anything about an individual. If you have concerns about your own or someone else's health, please speak to a qualified health professional.

Sources

1. Johnson SB, et al. Cancer misinformation and harmful information on Facebook and other social media: a brief report. Journal of the National Cancer Institute, 2022 (online 2021). europepmc.org/article/MED/34291289
2. Johnson SB, Park HS, Gross CP, Yu JB. Use of alternative medicine for cancer and its impact on survival. Journal of the National Cancer Institute, 2018;110(1):121-124. academic.oup.com/jnci/article/110/1/121/4064136
3. Loeb S, et al. Dissemination of misinformative and biased information about prostate cancer on YouTube. European Urology, 2019. europepmc.org/article/MED/30502104
4. Suarez-Lledo V, Alvarez-Galvez J. Prevalence of health misinformation on social media: systematic review. Journal of Medical Internet Research, 2021. europepmc.org/article/MED/33470931
5. World Health Organization. Infodemic (health topic). WHO, accessed 2026. who.int/health-topics/infodemic
6. Johnson SB, Park HS, Gross CP, Yu JB. Complementary medicine, refusal of conventional cancer therapy, and survival among patients with curable cancers. JAMA Oncology, 2018;4(10):1375-1381. europepmc.org/article/MED/30027204
7. Walter N, et al. Evaluating the impact of attempts to correct health misinformation on social media: a meta-analysis. Health Communication, 2021;36(13):1776-1784. europepmc.org/article/MED/32762260
8. Fridman I, Bylund CL, Elston Lafata J. Trust of social media content and risk of making misinformed decisions: survey of people affected by cancer and their caregivers. PEC Innovation, 2024;5. europepmc.org/article/MED/39323933
9. Loeb S, Langford AT, Bragg MA, Sherman R, Chan JM. Cancer misinformation on social media. CA: A Cancer Journal for Clinicians, 2024;74(5):453-464. europepmc.org/article/MED/38896503
10. Chan MS, Jones CR, Hall Jamieson K, Albarracín D. Debunking: a meta-analysis of the psychological efficacy of messages countering misinformation. Psychological Science, 2017;28(11):1531-1546. europepmc.org/article/MED/28895452