Essence, Ebony & O: Breast Cancer Coverage in Black Magazines

Abstract

Black women 45 and older are less likely than women of other races to be diagnosed with breast cancer, but Black women are more likely to die of the disease and more likely to be diagnosed at younger ages. It is crucial that Black women have access to high-quality information about breast cancer. In this study, we examined the comprehensiveness and accuracy of breast cancer information in magazines popular among Black women (Essence, Ebony, and O, the Oprah Magazine). Independent coders compared magazine content to factual statements about breast cancer selected by breast cancer experts. The 55 articles about breast cancer published between 2002 and 2007 were much more likely to focus on family history than increasing age as a risk factor. Racial disparities were mentioned in about 40% of the articles, but, when mentioned, only about 50% of articles provided fully accurate information. In addition, only 40% of the articles that mentioned the then-accepted recommendation for women 40 and older to have regular mammograms were fully accurate. Breast cancer information provided to Black women in popular magazines may not offer readers the information they need to assess their breast cancer risk accurately; the study suggests the magazine content may not contain the information necessary to optimize health promotion and early detection strategies for breast cancer.

KEYTERMS:

• accuracy
• Black women
• breast cancer
• magazines

Acknowledgments

We acknowledge the National Cancer Institute, which provided the funding for the study described in the article (R03 CA097807). In addition, we gratefully acknowledge Dr. Judith Lightsey and Ms. Gigi Moore-Higgs, who served as our expert panel members.

Notes

*All circulation figures are from the Magazine Publishers of America's 2007 Average Total Paid & Verified Circulation for Top 100 ABC Magazines.

†African-American/Black Market Profile, n.d.

¹According to the American Cancer Society, the current recommendation is that women in their 20 s and 30 s should have a clinical breast exam every 3 years and then annually after they turn 40 (“Breast Cancer: Early Detection,” 2011). However, we did not look for an age distinction in articles referencing this recommendation.

*This difference approached statistical significance. χ² (2, N = 55) = 5.53, p = .06.

¹According to the American Cancer Society, the current recommendation is that women in their 20 s and 30 s should have a clinical breast exam every 3 years and then annually after they turn 40 (“Breast Cancer: Early Detection,” 2011). However, we did not look for an age distinction in articles referencing this recommendation.

*In some cases, Cohen's kappa levels appear artificially low because the PRAM program expects each article to have been coded by all four coders. In addition, when there was no variation in the codes used for a given variable by a particular coder pair—as was often the case when a coder pair found no mention of that variable in any of the articles they coded—PRAM sets the coefficient for that variable/coder pair to 0. Thus, for this analysis, percentage agreement may actually be the more accurate indicator of intercoder reliability.

For high-risk patients, doctors can offer chemoprophylaxis to reduce the likelihood that they will develop breast cancer; in addition, prophylactic surgery does reduce risk in high-risk patients.

All of these articles were published before the U.S. Preventive Services Task Force's November 2009 release of new recommendations about when women should begin having mammograms, so this key fact accurately represented medical consensus at the time the articles examined in the study were published. Thus, controversy surrounding the Task Force recommendations could not have influenced the coverage on this issue.

Intercoder agreement was calculated in relation to the entire dataset, not only the articles in the three magazines discussed in this article. We did not include in our dataset the dates when particular items were coded, and new data were simply added to the main dataset as they were completed. Each time we reviewed inter-coder agreement, we examined each variable separately to determine that each had adequate coder agreement (85% or better) and to assess individual coders' consistency with the rest of the coding team. However, we did not maintain records of the results of these periodic assessments.