Fact-Checking Cancer Information on Social Media in Japan: Retrospective Study Using Twitter
Bibliographic record
Abstract
BACKGROUND: The widespread use of social media has made it easier for patients to access cancer information. However, a large amount of misinformation and harmful information that could negatively impact patients' decision-making is also disseminated on social media platforms. OBJECTIVE: We aimed to determine the actual amount of misinformation and harmful information as well as trends in the dissemination of cancer-related information on Twitter, a representative social media platform. Our findings can support decision-making among Japanese patients with cancer. METHODS: Using the Twitter app programming interface, we extracted tweets containing the term "cancer" in Japanese that were posted between August and September of 2022. The eligibility criteria were the cancer-related tweets with the following information: (1) reference to the occurrence or prognosis of cancer, (2) recommendation or nonrecommendation of actions, (3) reference to the course of cancer treatment or adverse events, (4) results of cancer research, and (5) other cancer-related knowledge and information. Finally, we selected the top 100 tweets with the highest number of "likes." For each tweet, 2 independent reviewers evaluated whether the information was factual or misinformation, and whether it was harmful or safe with the reasons for the decisions on the misinformation and harmful tweets. Additionally, we examined the frequency of information dissemination using the number of retweets for the top 100 tweets and investigated trends in the dissemination of information. RESULTS: The extracted tweets totaled 69,875. Of the top 100 cancer-related tweets with the most "likes" that met the eligibility criteria, 44 (44%) contained misinformation, 31 (31%) contained harmful information, and 30 (30%) contained both misinformation and harmful information. Misinformation was described as Unproven (29/94, 40.4%), Disproven (19/94, 20.2%), Inappropriate application (4/94, 4.3%), Strength of evidence mischaracterized (14/94, 14.9%), Misleading (18/94, 18%), and Other misinformation (1/94, 1.1%). Harmful action was described as Harmful action (9/59, 15.2%), Harmful inaction (43/59, 72.9%), Harmful interactions (3/59, 5.1%), Economic harm (3/59, 5.1%), and Other harmful information (1/59, 1.7%). Harmful information was liked more often than safe information (median 95, IQR 43-1919 vs 75.0 IQR 43-10,747; P=.04). The median number of retweets for the leading 100 tweets was 13.5 (IQR 0-2197). Misinformation was retweeted significantly more often than factual information (median 29.0, IQR 0-502 vs 7.5, IQR 0-2197; P=.01); harmful information was also retweeted significantly more often than safe information (median 35.0, IQR 0-502 vs 8.0, IQR 0-2197; P=.002). CONCLUSIONS: It is evident that there is a prevalence of misinformation and harmful information related to cancer on Twitter in Japan and it is crucial to increase health literacy and awareness regarding this issue. Furthermore, we believe that it is important for government agencies and health care professionals to continue providing accurate medical information to support patients and their families in making informed decisions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.004 |
| Science and technology studies | 0.002 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".