Risk Messages About HPV, Cervical Cancer, and the HPV Vaccine Gardasil: A Content Analysis of Canadian and U.S. National Newspaper Articles

Abstract

The human papillomavirus vaccine (Gardasil) is a significant advancement in women's health. We compared the reporting of fear-inducing messages about human papillomavirus, cervical cancer, and the human papillomavirus vaccine in Canadian and U.S. national newspapers between January 2006 and December 2007. Significant differences between countries were found in the number of articles containing fear messages about human papillomavirus, cervical cancer, and the human papillomavirus vaccine. Educational level of readability was higher than recommended for the public, and the emotional tone of the articles became progressively negative over time. Our findings suggest that public discussion of some elements of the human papillomavirus vaccine message that could cause alarm or worry for women may need to be addressed within political and cultural contexts.

KEYWORDS:

• human papillomavirus vaccine
• newspapers
• risk communication
• public policy
• fright factors

INTRODUCTION

Human papillomavirus (HPV) is one of the most frequently occurring sexually transmitted infections (STIs) in Canada, the U.S., and globally, with a peak prevalence among adolescents and young adults (Health Canada, 2007a; Centre for Disease Control [CDC], 2007; World Health Organization, 2008). While no definitive treatment is available for the virus itself, many immunologically competent individuals are able to clear the infection without treatment (Health Canada, 2007a). However, certain types of persistent HPV infections are the leading cause of cervical dysplasia (precancerous changes) and cervical cancer. Over 100 HPV types have been identified to date, and more than 40 of those infect mucosal surfaces (CDC, 2007). Some HPV types are classified as “high risk” because they can trigger the development of cancer, whereas others are classified as “low risk” because they rarely cause cancerous lesions but can lead to other abnormal growths (National Cancer Institute [NCI], 2007). Two high risk types (types 16 and 18) are responsible for about 70% of cervical cancer cases, and two low risk types (types 6 and 11) cause 90% of genital warts. All cases of cervical cancer can be traced back to one of the cancer-causing HPV types (CDC, 2007). Other variants of the virus can lead to oropharyngeal, anal, vaginal, and penile cancers (NCI, 2007). Although HPV infection occurs among both males and females, it is a significant etiologic factor in chronic disease burden among women and especially poor, underserved, and marginalized women (Kaplan-Myrth & Dollin, 2007).

The consequences of HPV exposure make a possible vaccine a welcome advancement in the field of women's health. The first prophylactic vaccine introduced in North America was Gardasil, a quadrivalent vaccine directed against HPV types 6, 11, 16, and 18, developed by Merck & Co. Not unexpectedly, the introduction of Gardasil brought much media and public attention, but it has not been without controversy. Although the vaccine was introduced within a short time span in Canada (July, 2006) and the U.S. (June, 2006), different courses of action occurred in the implementation of the vaccine. In the U.S., controversies surrounded a swift executive order by Texas Governor Rick Perry in February 2007 requiring girls entering middle-school (11 and 12 years old) to be vaccinated against HPV (Haber, Malow, & Zimet, 2007). This legislative action occurred amid reports of Merck & Co. contributing funds towards Governor Perry's re-election campaign. In Canada, concerns over lobbying were more subdued. The cost effectiveness of allocating public funds for school-based immunization programs for girls in Grade 8 starting in September 2007 in various provinces was debated due to the lack of long-term data from clinical trials (Lippman et al., 2007). In both countries, the involvement of many stakeholders, each with a distinct objective, and the sensitive nature of vaccination programs for a STI being targeted to young girls, contributed to the mixed reactions from the public and health officials alike. In both countries, HPV vaccination campaigns were deemed failures due to the preoccupation with mandates and legislation rather than the vaccine itself and “sensationalist” media reporting (Dyer, 2007; Haber, Malow, & Zimet, 2007).

Mass media has a large role in communicating about and amplifying the risks of the HPV vaccine as members of the public often obtain information about risk from the media (Bennett, 1999). More specifically, the media was cited as a source of information about the HPV vaccine (Gardasil), and vaccine uptake was associated with media use (including the use of print media) (Hughes et al., 2009). While not intended to be health education material, newspaper articles are a leading source of publicly available health information (MacDonald & Hoffman-Goetz, 2002). Unfortunately, media sources do not always provide balanced and complete information about vaccines (Danovaro-Holliday, Wood, & LeBaron, 2002; Goodyear-Smith et al., 2007). For example, negative media coverage of the pertussis vaccine and the MMR vaccine led to a decrease in vaccine confidence and less uptake due to skewed and inaccurate portrayals of perceived vaccine risk by anti-vaccine movements and consumer groups (Goodyear-Smith et al., 2007; Gangarosa et al., 1998). This may have contributed to the resurgence of whooping cough and measles in some areas where negative media coverage was highly prevalent (Goodyear-Smith et al., 2007; Gangarosa et al., 1998). Focusing on factors that raise concern, anxiety, or fear about the implementation of the HPV vaccine could contribute to both the failure to present accurate and balanced risk information and the lack of vaccine uptake by the concerned or frightened public.

The public reacts differently to varying situations that present a potential risk. In an effort to categorize why some risks bring about more fear than others, Bennett (1999) compiled a diagnostic checklist of “fright factors” that defined aspects of a (health) risk that are worrying and unacceptable to the public (Table 1). These fright factors were derived from a summary of psychometric research aiming to answer questions pertaining to risk perceptions and why certain risks can bring about alarm and worry that are unrelated to any scientifically estimated degree of risk.

Although other attempts have been made to understand how public and individual risk perceptions are formed (e.g., Langford, 2002; Sandman, 2001), Bennett's identification of key items is inclusive and applicable to broad public health-related risks. It also provides a diagnostic checklist that can be systematically applied to different health topics (e.g., fright factors have been cited by the Government of Canada in addressing environmental and workplace health concerns (Health Canada, 2007b)). Reactions to differing courses of action, methods of health care delivery (public versus private health care), and social/cultural perspectives in Canada and the U.S. could contribute to varying fright factors being more prominent in each country. While previous content analyses of health news stories have explored the accuracy and inclusiveness of reporting on STIs (including HPV) (Calloway et al., 2006; Anhang et al., 2004; Davidson & Wallack, 2004), none have examined how attributes of fear or “fright” about HPV, cervical cancer, and the HPV vaccine are presented in mass print media.

TABLE 1 Bennett's Fright Factor Diagnostic Checklist for HPV, Cervical Cancer, and HPV Vaccine News Stories

Fright factor	Application to HPV, cervical cancer, or HPV vaccine
Threaten a form of death, illness, or injury arousing dread	HPV: etiologic agent behind genital warts, cervical cancer, other cancersCervical Cancer: descriptions of cervical cancer progression and treatment, morbidity/mortality rates
Inescapable by taking personal precautions	HPV: very high prevalence, confidence in protective measuresCervical Cancer: sensitivity/specificity of screening methods
Arise from unfamiliar or novel source	HPV: lack of public awareness about prevalence and link to cervical cancerCervical Cancer: lack of public knowledge
Result from man-made sources	HPV Vaccine: consequences of policies, rushed approval, reports on side effects
Cause hidden or irreversible damage	HPV: may cause cervical cancer years later, some individuals unaware of infectionCervical Cancer: slow progression of disease
Pose particular danger to small children	HPV Vaccine: uncertainty associated with target group (young girls); effect of current policy on future health
Inequitably distributed	Cervical Cancer: higher risk with lower SES/under-served populationsHPV Vaccine: cost concerns, only girls vaccinated
Damage identifiable victims	Cervical Cancer: experiences of victimsHPV Vaccine: experiences of adverse events
Poorly understood by science	HPV: no definitive treatment for virusHPV Vaccine: lack of long-term data on efficacy, safety
Subject to contradictory statements from responsible sources	Cervical Cancer: high versus low riskHPV Vaccine: highly effective and warranted for mass use versus needing more research and not yet warranted for mass use
Involuntary	Not applicable

Therefore, the purpose of this study was to describe the presentation and portrayal of risk messages by comparing the type and frequency of fright factors about HPV, cervical cancer, and the HPV vaccine in Canadian and U.S. national newspaper articles published shortly before and after the HPV vaccine was approved and implemented into policy. We hypothesized that the types of fright factors in Canadian and U.S. national newspapers would not differ due to the universal impact of health beliefs on vaccine uptake and similar issues and concerns about HPV, cervical cancer, and the HPV vaccine prior to the introduction of the vaccine in both countries (Brewer & Fazekas, 2007; Zimet, Mays, & Fortenberry, 2000). In addition, we hypothesized that the number or volume of fright factors cited would differ between the two countries due to differing courses of action (such as the introduction of mandates in the U.S.), health care systems, and social/cultural perspectives.

METHODS

Sample Selection and Data Sources

We evaluated news coverage of the HPV vaccine (Gardasil) published between January 2006 and December 2007, inclusive. We used the online databases, LexisNexis Academic and Factiva, to obtain news stories about the HPV vaccine. To focus the study on highly visible news stories that were more likely to reach a large audience, we searched two national newspapers in Canada—The Globe and Mail (average weekly circulation for the years 2006 and 2007: 333,475 and 335,013) and The National Post (average weekly circulation for 2006 and 2007: 214,118 and 240,030)—and in the United States we searched two national newspapers—The Wall Street Journal (average weekly circulation for 2006 and 2007: 2,011,882 and 2,043,235) and USA Today (average weekly circulation for 2006 and 2007: 2,293,137 and 2,269,509)—to find articles about the HPV vaccine. We also included The Toronto Star as part of the Canadian news coverage. Although this newspaper is not considered national in distribution, the circulation numbers (average weekly circulation for 2006 and 2007: 447,145 and 452,502) indicate an audience reach greater than either of the two national papers. The New York Times (average weekly circulation for 2006 and 2007: 1,037,828 and 1,086,797) was included as a national newspaper and a newspaper of record in media outlets for the U.S. (Audit Bureau of Circulations [ABC], 2008). Information about circulation rates for each of the newspapers was based on average circulation rates from Monday to Friday and was obtained from the Audit Bureau of Circulations (ABC, 2008).

Because the HPV vaccine (Gardasil) is directed against HPV, an important etiologic factor in the development of cervical cancer, in our search of news coverage of the HPV vaccine we included articles that also mentioned HPV and cervical cancer linked to the vaccine. The search of the two databases was conducted using the search string “(HPV OR human papillomavirus) AND cervical cancer AND (Gardasil OR vaccine)” in the title or text, with a date range from 01/01/2006 through 12/31/2007, and limited to the chosen newspaper outlets. This time period was chosen because it included news stories published shortly before or after approval of Gardasil by the U.S. FDA (June 8, 2006) and Health Canada (July 18, 2006) and the release of recommendations from federal advisory boards (on February 15, 2007 by the National Advisory Committee on Immunization at the Public Health Agency of Canada and March 12, 2007 by the Advisory Committee on Immunization Practices at the CDC). This was also a time frame when public policies about the vaccine were being discussed or implemented, making the personal relevance of the information in the articles an immediate one to the reader who may be a parent concerned about their daughter or an individual contemplating vaccination.

The search of LexisNexis and Factiva resulted in 167 total news articles (80 Canadian and 87 U.S. articles) from the six media outlets. After the full text was received, articles were reviewed and those not appropriate to the study were excluded; excluded articles were those mentioning cervical cancer or HPV but without coverage about the vaccine. When duplicate news stories were published in different newspapers within each country (Canada or U.S.), only the first newspaper to publish it was counted in the total tally (i.e., coded only once). Articles were included if all three keywords (HPV, cervical cancer, Gardasil or HPV vaccine) appeared at least once in the article.

Data Coding Process

To develop the coding instrument, a directed content analysis approach was used. Directed content analysis is based on an a priori theory or framework that allows for interpretation of the text in a systematic way in which the text can be coded to assess emergent themes (Hsieh & Shannon, 2005). Initial coding was based on Bennett's diagnostic checklist of fright factors that affect public perception of risk and inform health risk communication (Bennett, 1999). Throughout the coding process, the researchers applied this typology of fright factors to the news stories on HPV, cervical cancer, and the HPV vaccine (Table 1). Articles were read in an iterative fashion and new or emergent themes (as relevant to fright factors) were coded as they appeared in the news article text within a category identified as “Other.” These emergent themes included: pharmaceutical lobbying, promiscuity and changes in sexuality, and the effect of the vaccine on screening practices. Based on this coding, data were recorded for both the total number (volume) of fright factors and the number of articles that contained each fright factor. To ensure reliability of the coding, approximately 15% (n = 25) of the articles were read and coded independently by both researchers, and the results were compared. No disagreements or discrepancies in coding by the researchers occurred. The primary author then coded the remaining articles.

This study focused on the volume (number) and type of fright factors in news stories about the HPV vaccine in Canadian and U.S. high-circulating newspapers. However, individual perception of risk involves more than the identification of a fright factor (Sjöberg, 2002); emotional tone (valence) of the article influences how people interpret information (Grabe & Kamhawi, 2006). Tone was classified as positive or negative using the analytical approach of Taylor and Sorenson (2002). Positive tone (positive valence) is illustrated by the appearance of words such as “confidence,” “triumph,” “tremendous,” and “100% effective” in the article text. Negative tone (negative valence) was captured by words such as “scary,” “difficult,” “unwelcome,” and “controversy.” The total number of positive and negative valence words in each news article was determined and a tally was recorded. Articles were designated as neutral if there were equal numbers of positive and negative words or if there were no words coded as either positive or negative in the news article.

Additional Variables Examined

Other coded variables included the month and year of publication and educational level to which the article is directed for readability. Readability was assessed because it can influence the degree to which the public understands the information and the public's ability to make an informed decision (Lunin, 1987). Readability of each news article was assessed using a “Simple Measure of Gobbledygook” (SMOG) analysis online calculator, which assigned a reading grade level (RGL) score based on the number of sentences and polysyllabic words in the article (Words Count, 2007). This score represents the level of attained education needed to understand a text (McLaughlin, 1969).

Data Analysis

Articles were coded manually, and text and codes entered in NVivo software (NVivo Software; QSR International Pty Ltd., Cambridge, MA, Version 7, 2006). This software enables the user to arrange coding categories into free nodes or hierarchical tree nodes. For this analysis, the tree nodes were “HPV,” “Cervical Cancer,” and “HPV Vaccine,” with the fright factors as the sub-nodes. The number of times each sub-node was coded in the article text for each country was obtained.

Descriptive statistics were generated to give general summary information about the HPV vaccine articles using SAS (SAS Software, Version 9.1, 2003) (SAS Institute Inc., Cary, NC). Citations were recorded and tallied to assess the volume of fright factor reporting. Simple analyses were conducted using cross tabulations and two-by-two chi-square tests to explore associations between country and the number of articles citing each fright factor. Fright factors with no cases (no articles) or expected cell counts less than five were not included in the chi-square analysis for statistical validity of the test (Sirkin, 2006). Simple regression analysis was used to determine whether the emotional tone of the Canadian and U.S. articles changed over time (the model was based on article valence plotted over time using a designation of 1 for positive articles, 0 for neutral articles, and −1 for negative articles). Analysis of variance using a general linear model was used to test differences in the rate of change in valence between the two countries. Unpaired t tests were applied to the readability level (SMOG analysis) as a function of country where the articles were published. All values presented are means with 95% confidence intervals, and p < 0.05 was accepted as statistically significant.

RESULTS

Coverage by Country

A total of 79 news articles were identified on the HPV vaccine from the three Canadian and 85 from the three U.S. newspaper outlets for the period January 1, 2006 through December 31, 2007. The distribution of articles in the Canadian newspapers was 39% in The Globe and Mail, 27% in The National Post, and 34% in The Toronto Star. U.S. national newspapers had almost two times the volume coverage in The New York Times, 45%, compared with USA Today, 25%, whereas The Wall Street Journal had a volume of coverage falling between these two of 31%. The overall difference in volume of news articles between the two countries was not statistically significant.

Coverage by Month and Year

Spikes in the number of articles reporting about HPV, cervical cancer, and the HPV vaccine between the two countries occurred in June 2006 (5% of Canadian newspaper articles; 14% of U.S. newspaper articles), February 2007 (4% of Canadian articles; 25% of U.S. articles), and August–September 2007 (48% of Canadian articles; 5% of U.S. articles) (Figure 1). These spikes likely correspond to FDA and Health Canada approvals (June 8, 2006 and July 18, 2006, respectively), attempts to introduce vaccination mandates in Texas in February 2007, and the beginning of the school year which included the implementation of universal immunization programs of girls in school in Canada.

Coverage by Fright Factors

The greatest number of fright factor citations was in reference to HPV. A total of 432 fright factor citations (Canada: 210 [49%]; U.S.: 222 [51%]) were recorded for HPV. This was followed by 309 (Canada: 185 [60%]; U.S.: 124 [40%]) citations in reference to the HPV vaccine and 215 (Canada: 116 [54%]; U.S.: 99 [46%]) citations for cervical cancer.

HPV

For coverage in both countries, the fright factor cited most often (by volume and number of articles) about HPV was that of threatening death, illness, or injury (Total: 256; Canada: 121 [47%]; U.S.: 135 [53%]); this was followed by descriptions of HPV as being inescapable or pervasive in the population (Total: 101; Canada: 49 [49%]; U.S.: 52 [51%]). The number of articles with the fright factors of threatening death/illness, inescapable, and hidden or irreversible damage differed significantly between Canada and the U.S. (Table 2); for each of these fright factors, a greater number of articles appeared in Canadian compared to U.S. newspapers. Examples of HPV fright factors taken from the Canadian and U.S. media reports are illustrated by the following quotes:

FIGURE 1 Total number of articles on HPV, cervical cancer, and the HPV vaccine published per month in Canadian and U.S. national newspapers.

HPV fright factor: Threaten death, illness, or injury

CANADA: “… the four strains of HPV targeted by the vaccine cause an estimated 36,000 new cases of female genital warts each year in Canada, resulting in more than 85,000 doctor's visits for diagnosis and treatment.” The Globe and Mail (December, 2006)

CANADA: “These virus types are responsible for most cases of cervical cancer, almost all cases of genitals warts, many cases of vaginal and vulvar cancer, as well as pre-cancerous lesions or abnormal cells …” The National Post (August, 2007)

U.S.: “Persistent infection with certain HPV types leads to cervical cancer.” USA Today (April, 2006)

U.S.: “Instead, both vaccines aim to prevent patients from becoming infected with two types of HPV—types 16 and 18—that are thought to cause 70% of cervical-cancer cases world-wide.” The Wall Street Journal (August, 2007)

TABLE 2 Number (Percent) of Articles Containing Fright Factors in Canadian and U.S. National Newspapers ^a

Fright factor	Canada # (%)	U.S. # (%)	Chi-square	p value
HPV
Threaten death/illness	57 (72)	38 (45)	12.66	0.0004
Inescapable	23 (29)	11 (13)	6.52	0.01
Hidden and irreversible damage	14 (18)	4 (5)	7.10	0.01
Cervical cancer
Threaten death/illness/injury	51 (65)	30 (35)	14.03	0.0002
Damage identifiable victims	10 (13)	4 (5)	3.32	n.s.
Inequitably distributed	8 (10)	14 (16)	1.42	n.s.
HPV vaccine
Result from man-made sources	18 (23)	15 (18)	0.67	n.s.
Inequitably distributed	35 (44)	17 (20)	11.17	0.0008
Poorly understood by science	33 (42)	19 (22)	7.13	0.008
Subject to contradictory statements	12 (15)	2 (2)	8.64	0.003
Other themes
Pharmaceutical lobbying	13 (16)	18 (21)	0.60	n.s.
Promiscuity and changes in sexuality	17 (22)	15 (18)	0.39	n.s.
Effect on screening	12 (15)	5 (6)	3.82	0.05

Cervical cancer

Similar to the findings for HPV, news coverage in both countries cited cervical cancer as threatening death, illness, or injury more than any other fright factor resulting in a total of 141 citations (Canada: 76 [54%]; U.S.: 65 [46%]). Canadian newspapers appeared to have a greater frequency of citations under identifiable victims (Total: 20; Canada: 15 [75%]; U.S.: 5 [25%]). Data from U.S. newspapers suggested a greater emphasis on the inequitable distribution of cervical cancer compared to Canadian newspapers (Total: 27; Canada: 8 [30%]; U.S.: 19 [70%]). A significant difference was observed between Canada and the U.S. in the number of articles reporting on the fright factor of threatening death/illness, with more Canadian articles citing this fright factor compared with U.S. articles (Table 2). Representative examples of fright factors identified about cervical cancer are given below:

Cervical cancer fright factor: Hidden and irreversible

CANADA: “The Pap test didn't catch it, and when it doesn't, cervical cancer is incredibly hard to detect until in most instances it is too late …” The Toronto Star (August, 2007)

CANADA: “Some diseases, such as cancer, cannot be successfully treated once they progress beyond a certain stage.” The Globe and Mail (March, 2007)

U.S.: “Pap testing is effective only because it is done often; a lesion can take 10 years to turn into a cancer …” The New York Times (January, 2007)

U.S.: “Cervical cancer grows slowly, and it is long-lasting HPV infection that leads to the disease.” USA Today (November, 2007)

HPV vaccine

The fright factors which were most salient in differing by volume were “poorly understood by science” (Total: 96; Canada: 63 [66%]; U.S.: 33 [34%]) and “contradictory statements” (Total: 27; Canada: 19 [70%], U.S.: 8 [30%]) where a greater proportion of citations were from Canadian newspaper sources for each of the two fright factors. Analysis of the number of articles for each fright factor indicated significant country differences for those factors of “poorly understood by science” and “contradictory statements” (Table 2). For both of these fright factors, more articles were from Canadian newspapers than from U.S. newspapers. Examples of quotes of fright factors about the HPV vaccine in the newspaper outlets in Canada and the U.S. are shown below:

HPV vaccine fright factor: Poorly understood by science

CANADA: “However, it's not known how long its protection will last or if women will have to receive booster shots later in life.” The Globe and Mail (June, 2006)

CANADA: “Lead author, McGill University epidemiologist Abby Lippman, warned that the long-term effects of the Gardasil vaccine are not known.” The National Post (November, 2007)

U.S.: “Public health officials want to vaccinate girls early, before they become sexually active, even though it is not known how long the immunity will last.” The New York Times (July, 2006)

U.S.: “However, an editorial comment accompanying the Lancet article noted that while the data are encouraging, their interpretation has limitations given the short follow-up period of the trial, which averaged 15 months.” The Wall Street Journal (June, 2007)

Emergent themes

A total of 55 citations were found for pharmaceutical lobbying (Canada: 21 [38%]; U.S.: 34 [62%]), 48 citations were found for references to promiscuity and changes in sexuality (Canada: 29 [60%]; U.S.: 19 [40%]), and 21 citations were found for references to vaccine effects on screening and safe sex (Canada: 15 [71%]; U.S.: 6 [29%]). The reporting of screening practices by article differed significantly between Canada and the U.S., with more articles from Canadian sources reporting on this theme than U.S. sources (Table 2). Examples of quotes of emergent themes about the HPV vaccine in news media outlets in Canada and the U.S. are shown below:

Emergent theme: Pharmaceutical lobbying

CANADA: “Some also whisper that it stems from too a cozy relationship with the pharmaceutical company that makes the vaccine.” The Globe and Mail (March, 2007)

CANADA: “Some critics questioned Health Canada's speedy adoption of Gardasil, saying it followed a massive lobbying campaign by its makers, Merck-Frost.” The National Post (November, 2007)

U.S.: “Prompted in part by a vigorous lobbying campaign by Merck, which stands to earn billions of dollars if the vaccine is required, legislators … proposed mandating vaccination against HPV for girls as young as 11.” USA Today (February, 2007)

U.S.: “Merck aggressively lobbied state legislatures to make vaccination a school requirement for 11- and 12-year-old girls …” The Wall Street Journal (May, 2007)

Other Variables Affecting Risk Presentation and Understanding

Valence of article

No statistically significant difference was observed between Canadian and U.S. newspaper coverage of the HPV vaccine with respect to the frequency of positively or negatively valenced news articles (Canada: 29% positive, 28% negative, 43% neutral; U.S.: 26% positive, 32% negative, 42% neutral). Regression analysis indicated that articles in the Canadian (r = −0.49, df = 77, p < 0.0001) and U.S. (r = −0.26, df = 83, p < 0.05) newspapers had a more negative tone over time, albeit this association was weaker for the U.S. (Figure 2). No statistically significant difference was found in the rate of change in valence between Canada and the U.S., indicating that one country did not become more negative over time than the other.

Readability level

The RGL as determined by SMOG for articles on the HPV vaccine published in Canadian and U.S. newspapers was considerably higher than recommended for the general public. Materials written at or below the 5th-grade reading level were categorized as excellent, those written at the 6th- to 8th-grade reading level were categorized as adequate, and those written at a 9th-grade reading level or higher as unsuitable (Doak, Doak, & Root, 1996). Articles published in Canadian newspapers were written at a lower average RGL of 14.6 (95% CI = 14.4–14.9) compared to those published in U.S. newspapers (15.2 RGL, 95% CI = 14.8–15.5). The difference in RGL between news articles on HPV, cervical cancer, and the HPV vaccine published in the U.S. and those published in the Canadian newspaper outlets was significant (t = −2.39, p < 0.05).

FIGURE 2 Change in valence over time in the discussion of the HPV vaccine in Canadian and U.S. national newspaper articles.

DISCUSSION

Our directed content analysis of Canadian and U.S. national newspapers revealed two important and novel findings with respect to inclusion of fright factors in news coverage of the HPV vaccine. First, whereas the types of fright factors identified between the two countries were similar, significant differences by country were found for the number of articles citing fright factors about HPV, cervical cancer, and the HPV vaccine as well as differences in volume of some fright factors reported for cervical cancer and the HPV vaccine. Second, differences in volume between countries were found for emergent themes. Factors that influence the differences in reporting between the two countries may be due to differences in social/cultural normative perspectives and health care funding and delivery.

Similarities in fright factor volume were found between Canadian and U.S. national newspapers in reporting about HPV. Many articles reported on prevalence rates and the sequelae of HPV infection (coded as threatening death, illness, or injury or as being inescapable when high rates were emphasized). Reporting on statistical information of this nature was mirrored in a content analysis of newspaper articles covering STIs and illustrates the type of information journalists deem as newsworthy (Davidson & Wallack, 2004). For example, journalists were more likely to report on STIs if supplied with new or striking information on the statistics or consequences of infection. This, combined with findings that HPV news coverage omits accurate cancer risk information and that most women with HPV do not progress to cervical cancer (Anhang et al., 2004), may serve to heighten or bias risk perceptions about HPV and its role in the development of cervical cancer.

Our results indicated that the two countries differed in the volume of fear messages about cervical cancer. Possible differences in the frequency of some fright factors over others may be linked to distinct cultural and social orientations and the impact of health care funding and delivery. Canadian newspapers constituted a higher percentage of citations for “identifiable victims” than did the American newspapers. The reason for this difference is not clear from our results, but may be due to differing societal and health values and normative frames. Canada has a public health care system, which more reflects a “solidaristic” perspective; this perspective emphasizes that collective needs supercede individual needs and society secures the necessities of life for the individual. In contrast, an “individualistic” orientation is one in which individuals are expected to obtain their own needs without (or only with limited) collective help (Kahan & Braman, 2006). Emphasis on the impact of cervical cancer on the individual in Canada could be indicative of this collective sense of responsibility to meet health care needs for all individuals and an inclination to personalize the commitment to the collective good. In contrast, the greater proportion of citations on the inequitable distribution of cervical cancer in articles from U.S. newspapers could be indicative of the lack of collectivism and the burden of disease on uninsured Americans. Socioeconomic differences in cervical cancer mortality exist in both countries, but the disparity was found to decrease in Canada while it endured in the U.S. (Ng et al., 2004). While definitive reasons for these disparities are not clear, some of the disparity may be attributed to differences in access to preventive health care services.

The number of articles containing fright factors about the HPV vaccine differed between Canada and the U.S. Differences in reporting can further be seen based on differing overarching perspectives in the Gardasil debate. Canadian newspapers had more articles discussing Gardasil as being poorly understood by science and contradictory statements about the vaccine than U.S. newspapers. These two fright factors can be linked in that contradictory statements about Gardasil were often in relation to its scientific merit. Canadian newspapers reflected debate about the merits of Gardasil as a vaccine and questioned the balance of benefits to risks; U.S. newspapers reflected debate about the suitability of mandates from social and rights perspectives (reflected by the higher number of citations under fear of man-made sources in U.S. newspapers). This difference in the discussion of the HPV vaccine may be related to social orientation differences between Canada and the U.S.: “individualists” tend to be dismissive of risk claims, while “egalitarians” tend to be receptive towards them (Kahan & Braman, 2006).

Differences in the volume of emergent themes between Canada and the U.S. suggest variation in the focus placed on certain political and social issues. As time progressed after FDA approval, the concern of pharmaceutical industry lobbying of legislators was prominent in U.S. newspapers and eventually shifted the focus away from the benefits of the HPV vaccine. The consequences of this political action of pharmaceutical lobbying on the dissemination and discussion of the HPV vaccine were reiterated in research and opinion papers on the topic (Haber, Malow, & Zimet, 2007; Wynia, 2007). Concern over changes in sexuality and promiscuity tended to be more prominent in Canadian newspapers due to responses to school-based vaccination programs from Catholic school boards in Ontario. These examples indicate the focus on certain risks or fears may reflect social concerns in the public arena in each country.

Assessing the progressively negative emotional tone of article coverage of the HPV vaccine in both Canada and the U.S. suggests that valence could impede or influence understanding and potentially uptake of the vaccine. Earlier articles described the vaccine as “100% effective” which contrasted with later articles that reported accounts of adverse events and concerns of long-term safety. Detailed discussion or criticism of vaccine policies and methods of implementation were largely absent in earlier news coverage, but dominated later newspaper coverage. News coverage on the Rotavirus vaccine from 1987 to 1999 indicated a similar trend of “early idealization—sudden condemnation” in reporting, and indicated that changes in reporting coincided with either scientific information or public health action (Danovaro-Holliday, Wood, & LeBaron, 2002). Considering the impact of the HPV vaccine on women's health, the use of negative tone is particularly concerning. Gender differences in processing negative news may lead to avoidance or to less attention by women to negatively valenced stories than to more positively valenced stories (Grabe & Kamhawi, 2006). Messages that contain a fear component are only beneficial if followed by a strong efficacy message (Witte & Allen, 2000). The propagation of fear through fright factors about HPV and cervical cancer coupled with negatively valenced reporting about the HPV vaccine may contribute to denial, defensive avoidance, or reactance towards the fear. This may result in an individual's inability to process the information effectively and lack of confidence in the HPV vaccine to alleviate the fear brought about by HPV and cervical cancer.

The readability of articles in newspapers from both countries was clearly at a very high level, was well beyond that which could be easily understood by the general public, and could be categorized as inadequate (higher than grade 8) (Doak, Doak, & Root, 1996). National literacy levels in Canada and the U.S. are measured across multiple domains including prose literacy, which is defined as the knowledge and skills required to comprehend and use information from text (ALL, 2005); this competency would include use and comprehension of news stories. Level 3 represents the minimum level of ability needed to cope with demands of everyday life and is the level of skill needed for secondary school completion. In 2003, over 80% of Canadians aged 16 to 65 were estimated to be at Level 3 or lower; 2003 estimates in the U.S. indicate approximately 87% of Americans were at or below Level 3 (ALL, 2005). Average readability levels of news stories in Canadian and U.S. newspaper articles evaluated in our study suggest the need for post-secondary education by the reader in order to fully comprehend the text. High readability scores affect the degree to which the public can understand health information, and ultimately influence their ability to make an informed choice (Lunin, 1987). This may be particularly important during a time when policies could have a direct affect on segments of the population who may not have fully grasped all aspects of the issue but are forced to either make a choice or resort to avoidance. Combined with women's tendency to avoid negative news, lack of understanding due to high readability may further impede women's abilities and willingness to make an informed choice about the HPV vaccine.

This study had several limitations. First, the results and conclusions inferred from this content analysis are limited to a select number of high circulating U.S. and Canadian newspapers and may not reflect newspaper coverage at state, provincial, municipal, and local levels. Nevertheless, given the widespread reach and influence of these national newspapers and the public health significance of the topic, national news media are likely to mirror the overall coverage of the issue in local news media (Demers et al., 1989). Moreover, other forms of non-print media, such as television and radio, were not analyzed. A second limitation was that the data set was limited to the specified time period (January 2006–December 2007) and selected media outlets. It is possible that coverage after December 2007 would have fewer fright factors or less negative valence. This could be due to less emphasis being placed on HPV and cervical cancer in news stories. Limiting the study to the selected time period and newspaper outlets may have contributed to a smaller sample size and an inability to detect significant differences (for example, in country specific citations about cervical cancer). A third limitation was that we categorized themes using Bennett's checklist and typology of fright factors and this resulted in a content analysis specific to these categories. Alternate typologies or methods of analysis may have resulted in different categories and results. A fourth limitation was that the results obtained through this study were not intended to be reflective of past or current risk perceptions towards HPV, cervical cancer, and the HPV vaccine, and may not reflect current reporting on the topic. A fifth limitation was that newspaper readers typically have higher income and education levels than the average population (Newspaper Association of America, 2006); thus, the newspaper audience may exclude lower income women and their daughters who may be most at risk of cervical cancer. Finally, media reporting of risk messages about the HPV vaccine does not capture actual uptake of Gardasil by women; clearly, other studies would be needed to measure the behavioral impact on women of the risk messages in the print media.

PUBLIC HEALTH IMPLICATIONS AND CONCLUSIONS

Our study suggests that some factors that elicit fear about the HPV vaccine and cervical cancer were common to both countries, while the frequencies of others may have varied based on the cultural and political environments. Bilateral (shared or more universal) factors include the impact of health beliefs about vaccines and the dread or fear associated with cancer and STIs, whereas culturally specific factors could include health care funding and delivery and societal norms that influence reactions to policy. Our findings also demonstrate that communicating health risk information in the mass media is not separate from social and political contexts and that health communication efforts may be overshadowed by negative media coverage. By identifying the facets of a particular health issue that are prone to cause alarm and worry, and by acknowledging the role media outlets play in heightening that reaction, public health practitioners and policy makers can steer communication towards attenuating reactions to fear and misperceptions. This can ensure that women have the necessary information and resources to make an informed choice about the HPV vaccine based on clear and unbiased risk communication that empowers them to take charge of their health.

Notes

Research was supported by a grant from the Social Sciences and Humanities Research Council of Canada.

^aFright factors with no (0) citations or expected cell counts less than 5 are not shown.