TL;DR
Get health and wellness essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Researchers evaluating online nutrition information for people diagnosed with breast cancer found that 12.5% of 151 claims were supported by guidelines, systematic reviews or established scientific facts. The study, based on searches conducted in early 2025, found the largest proportions of unsupported or contradicted claims on TikTok and X; it did not test whether patients followed the advice or suffered harm.
A study of 151 online nutrition claims aimed at people diagnosed with breast cancer found that just 12.5% were supported by reliable evidence or established scientific facts, according to research published in Patient Education and Counseling. The findings highlight how patients seeking dietary guidance online may encounter claims that are weakly supported, unsupported or contradicted by existing research.
The researchers gathered material from eight online platforms between January and April 2025: Google, Facebook, Instagram, Pinterest, YouTube, YouTube Shorts, TikTok, X and ChatGPT-4. They used five non-academic search phrases, including “breast cancer diet” and “cancer nutrition,” and reviewed selected search results and chatbot responses.
A multidisciplinary expert panel assessed claims against clinical guidance from organizations including the American Society of Clinical Oncology, the European Society for Clinical Nutrition and Metabolism, and the Clinical Oncology Society of Australia, as well as peer-reviewed systematic reviews. The researchers classified claims as supported, contradicted or backed by little evidence. They reported that 63% had no evidence, 21% had weak evidence and 4% were contradicted by established literature. Another 8.6% were supported by guidelines or systematic reviews, while 4% reflected established scientific facts.
Platform results varied. TikTok had the highest reported proportion of claims described as lacking support or contradicted, at 76.3%, followed by X at 72.9%. Google had 33.3% and ChatGPT 30.0% in the study’s supported-information category. The ChatGPT result came from only five responses, and the researchers said answers were similar and general across the five prompts. These figures describe the material selected for this study, not every post or answer on each platform.
Why Patients May Encounter Conflicting Advice
People affected by breast cancer may search for food advice to manage treatment side effects, support general well-being or learn about recurrence risk. Online information is readily available, but the study suggests that popular search results and social posts may not consistently reflect clinical guidance. This can make it harder for patients to distinguish established recommendations from personal opinions or claims with little research behind them.
The findings do not establish that online advice caused harm, or that patients acted on it. They instead identify a gap between the evidence base and the claims present in the selected results. For patients making diet decisions during or after treatment, the distinction matters: a claim’s popularity or confident presentation does not show that it is clinically supported. The study points to the value of checking nutrition questions with oncology care teams or qualified dietitians, rather than treating a post or chatbot response as individualized medical guidance.
As an affiliate, we earn on qualifying purchases.
How Researchers Checked Online Claims
The study reflects an information environment in which people diagnosed with cancer can consult clinicians and dietitians, but also use search engines, social media and conversational AI. Its authors noted that earlier research had found unsupported health claims online, while the reliability of advice on newer social platforms and AI tools had been less well examined.
To make the search resemble typical patient behavior, the team used everyday phrases rather than academic search terms. It reviewed the top 10 results for each phrase on social platforms and Google’s first page, alongside five ChatGPT-4 responses. Non-AI claims were also grouped by creator: health professionals, personal accounts such as influencers and advocates, and groups or organizations including foundations and commercial websites.
Health professionals’ claims had a somewhat higher share of evidence-supported information—29.2%—than claims from personal accounts, at 23.5%, or groups and organizations, at 24.1%. The study reported that these differences were not statistically significant. Health professionals contributed about one-third of the non-AI content collected. Frequently recurring subjects included cruciferous vegetables, alcohol, plant-based diets and claims that sugar “feeds” cancer; how often a topic appeared does not establish whether a related claim is true.
What the Study Cannot Establish
The results are limited to the material collected using the study’s search terms, platforms and sampling approach in January-April 2025. Online results change over time, and the percentages should not be read as a full audit of all breast cancer diet content. The small number of ChatGPT responses also limits conclusions about AI-generated advice more broadly.
The researchers did not assess whether patients followed the claims, what decisions they made or whether the advice caused beneficial or harmful outcomes. They also cautioned that a claim classified as lacking evidence is not automatically false; it means the review found no strong supporting evidence under its criteria. The study therefore measures the evidence quality of selected claims, not the effects of particular diets on cancer outcomes.
How Patients Can Check Diet Claims
The report does not describe a planned follow-up study or a forthcoming change to clinical guidance. The immediate practical step for readers is to take specific diet claims to their oncology team or a registered dietitian with cancer-care experience, who can discuss whether guidance applies to an individual’s treatment and circumstances.
For researchers and online platforms, the findings identify areas for further examination: whether results differ across search terms and over time, how AI tools respond to a wider range of questions, and whether patients rely on the claims they encounter. Until those questions are studied, the report supports a limited conclusion: many claims in the sampled results lacked strong evidence, while their real-world effects remain unknown.
Key Questions
What did the study find about online breast cancer diet advice?
Among 151 claims in the selected results, 12.5% were supported by reliable evidence or established scientific facts. The researchers classified 63% as having no evidence, 21% as having weak evidence and 4% as contradicted by established literature.
In the study’s platform analysis, TikTok was 76.3% and X was 72.9% for claims lacking supporting evidence or contradicted by research. The figures apply to the material sampled, not every post on those platforms.
Did the study show that online advice harms patients?
No. The researchers did not assess whether people followed the advice or whether it affected their health. They evaluated the evidence behind selected claims, not patient outcomes.
Does “no evidence” mean a diet claim is false?
Not necessarily. The study says claims without supporting evidence are not automatically false; the classification means the review did not find strong support under its assessment criteria.
How many ChatGPT answers did researchers assess?
The study assessed five ChatGPT-4 responses, one for each search phrase. That small sample limits conclusions about AI-generated nutrition advice overall.
Source: rss
Columbus Day / Indigenous Peoples' Day Picks
long weekend sales
As an affiliate, we earn on qualifying purchases.
