Screening the media for breast cancer news stories published between June 2019 and June 2020 resulted in 9,811 hits. Of these, 1,594 news stories had at least 1,000 total shares.
Source of news stories
Most of the stories in our sample were published in digital media (76.73%), whereas 23.27% originated from traditional media. However, the most shared news story (Table 2) was published in Fox News, a traditional media outlet. Among the twenty stories with the most shares, thirteen were published by traditional media outlets: four times by Fox News and once by nine other media entities, such as The Epoch Times, Metro, CNN and NBC News. The most shared story in a digital media is authored by the blog The Breast Cancer Site.
Table 2 Top 20 most popular news stories related to breast cancer (measured by total shares in social networking sites), its credibility and content type, between June 2019 and June 2020
Credibility and type of rumours
Regarding RQ1, of all news stories selected for coding, 69.7% have not been classified according to credibility. This is due to the fact these news items do not address science, risk factors, prevention, treatment or other aspects, which can be assessed for scientific accuracy. Among the news classified according to credibility (n = 483), 17.25% are ‘verified’ and 13.05% are ‘rumours’.
When we examine the most common types of rumours, we see ‘false connection/context’ represent 62.7% of the total, ‘misleading content’ are 34.9% of the total, and totally false content, that is, ‘fabricated content’ category, represents 2.4% of the total.
In consideration of the number of shares in relation to the credibility of the content (Fig. 1), we see the content classified as “rumours” tends to be more shared than those scientifically correct, both in traditional and digital media. Although less frequent in our sample, “rumours” totalled 5,755,192 shares, whereas “verified” stories were tallied at 1,747,352 total shares (3.29 times less).
Mean of total shares in relation to the credibility of the content, separated by traditional media and digital media
We can observe a very strong statistical connection (Cramer’s value = 0.313) between categories “credibility” and “content type”. News stories regarding “treatment” are 37.9% “rumours” and 62.1% (1.6 times more) “verified”. “Real life stories” are 58.3% (1.4 times more) “rumours” and 41.7% “verified”. About “risk factors”, 56.3% are “rumours” and 43.8% are “verified”; in “new technology”, 53.8% (nearly 1.5 times more) are “rumours” and 46.3% are “verified”. Stories classified as “educational” are 13.7% “rumours” and 86.3% (6.3 times more) “verified”. Finally, the dimensions “solidarity” and “complaint” are both 100% “verified” in our sample.
There is also a very strong association (Cramer’s value = 0.431) between categories “type of rumour” and “content type”. In the dimension “risk factors”, we observe that 79.5% of the rumours were classified as “false connection/context”, and 20.5% were deemed “misleading.” In the “treatment” dimension, 29.3% of the rumours are “misleading” and 70.7% (2.5 times more) have “false connection/context.”
Regarding RQ2, when we examine the distribution of content type categories in our sample (Fig. 2), we see most stories are classified as ‘real-life story’ or ‘solidarity’ (67.69%). These stories have no scientific content, since they are focused on narrating the life of an individual or family members with cancer, publicizing actions to raise money for cancer hospitals or requesting donations of any kind for patients in need, to mention a few examples.
Percentage of news stories in each content type dimension
During the period studied, we also noticed a recurrence of the same news among the stories with the most total shares (Table 2). These stories multiply on different web pages, often with the same title and text, or few variations.
Among the most shared stories, we see how the trial vaccination against breast cancer of a patient in Florida was highlighted (the story was repeated four times in the Top Twenty). The death from COVID-19 of a mom of 6 who survived breast cancer was also noteworthy, being repeated four times.
We also noticed a highlight in news that addresses celebrities with breast cancer, including the North-American actress Shannen Doherty, the British-Australian singer Olivia Newton-John, and the North-American television broadcaster Robin Roberts.
Mentions prevention and early detection/screening
To answer RQ3, most analysed news stories do not address ways of preventing or early detecting breast cancer (Table 3). In our sample, 5.08% of the stories comment on prevention and 19.7% mention early detection. There is an extraordinarily strong statistical connection (Cramer’s value = 0.435; Fisher’s exact test, p < 0.001) between content type and prevention; and between content type and early detection (Cramer’s value = 0.355; Chi-square test: p < 0.001).
Table 3 Percentage of news stories that address prevention and early detection, according to each content type dimension
Stories whose theme are “opinion”, “educational” and “risk factors” have the highest proportion of references in prevention. In relation to early detection, we see this characteristic in stories regarding “new technology”, “opinion” and “complaint”. We did not observe statistical connection between prevention versus credibility and between early detection versus credibility.
Breast cancer awareness month
Regarding RQ4, we compared the news published in October (known as the “Breast Cancer Awareness Month” or “Pink October” in several countries around the world ) with the other months. There are several variations in relation to the topics covered and in relation to the credibility of the most shared content (Chi-square test: p < 0.001; Cramer’s value = 0.300, extraordinarily strong connection).
There is a significant increase in news stories classified as “solidarity” in October (28.4% versus 9.3% in other months). On the other hand, there was a decrease in content that addresses “risk factors” (3.6% versus 13.1% in other months), “real-life stories” (47.9% versus 54.5% in other months) and “new technology” (1.0% versus 7.0% in other months). We can observe a slight increase in educational content (7.8% versus 5.3%).
When we compare the credibility of the news shared in October with the other months of the year, we see there is a statistically significant difference in the distribution of the types of rumours (Fisher’s exact test: p = 0.030; Cramer’s value = 0.172, strong connection). There is an increase in rumours classified as “false connection/context” (81.3% in October versus 59.3% in other months), whereas it is possible to note a decrease in “misleading content” (15.6% in October versus 38.4% in other months). There is no noteworthy difference in relation to “fabricated content” (3.1% in October versus 2.3% in other months).
We found a moderate connection (Chi-square test: p = 0.003; Cramer’s value = 0.138) between news stories’ credibility in October and in other months. Overall, in October there is an increase in news stories classified as “verified” (69.8% in October versus 53.3% in other months).