Beyond the "Angelina Effect": A Longitudinal Analysis of Celebrity Breast Cancer Disclosures' Impact on News Media and Public Breast Cancer Information Seeking Outcomes

Abstract

A long research tradition exists investigating the content of news coverage of celebrity breast cancer disclosures and, to a greater extent, the impact these personal health narratives have on public cancer-related outcomes. However, the bulk of this research focuses on specific, large-scale media events, such as Angelina Jolie’s 2013 BRCA disclosure. The attention to individual disclosures provide insight about the specific media event, but does not further knowledge about the larger phenomenon of celebrity cancer disclosures. To go beyond the Angelina effect, this dissertation addresses three overarching research questions: 1) What breast cancer-related messages are present in media coverage of celebrity breast cancer disclosures; 2) do these messages impact public cancer-related behavioral outcomes (i.e., online breast cancer information seeking); and 3) are there attributes of the celebrity that predict media and public outcomes? To address these questions, first, 110 individual celebrity breast cancer disclosures between 2005 and 2016 were systematically identified. Then, two longitudinal studies were conducted. To address the first question, Study 1 used computer assisted and hand coded procedures to assess the presence of episodic frames (defined as containing information specific to the celebrity and her experience with breast cancer) and thematic frames (defined as including population and subpopulation breast cancer information [e.g., prevalance, risk, survial rates]). In addition, the presence of seven content categories classified as misinformation (defined as information which is innaccurate, misleading, or oversimplified) in news coverage was assessed. Results demonstrated that 80% of the news articles were written with an episodic frame, and 20% were written with a thematic frame, indicating very little information beyond the celebrity’s own experience with breast cancer was conveyed to the public. However, misinformation was largely absent in the news coverage—only misinformation pertaining to early breast cancer detection and mastectomy decisions was present in 10% or more of the news coverage. Study 2 attempts to determine if news content impacts information seeking by using the framing outcomes from Study 1 to predict Google Trends search query outcomes. Due to the disparate rates in the presence of episodic and thematic frames this dissertation is unable to provide support linking content and online breast cancer information seeking outcomes. However, time series models suggest that media coverage of celebrity breast cancer disclosures in the aggregate have a distal impact on the public’s breast cancer information seeking outcomes. For example, some analyses suggested effects happened as late as 17 months after news coverage of the disclosure. Yet the nature of these trends may be a function of the data. Establishing if celebrity attributes can predict media and public outcomes was done through a moderation analysis of the results of Study 1 and Study 2. Specifically, the extent to which the presence of episodic and thematic and misinformation were present and statistically significant information seeking models were examined as a function of the celebrity’s age, career type, breast cancer-event type, and level of celebrity status (defined as the degree of fame the celebrity achieved at the time of disclosure). Eighty-seven percent of thematic frames present were in news coverage of celebrities at the highest levels of fame. Specific categories in the age, career-type, and level of celebrity status variables predicted the presence of misinformation. Some preliminary evidence suggests level of celebrity status may predict online breast cancer information seeking outcomes. The implications of the dissertation’s findings for health communication research, mass media effects research, and professional health communicators are discussed.