Communication Assessment of the Most Frequented Breast Cancer Websites: Evaluation of Design and Theoretical Criteria

Abstract

The internet is a prominent source of health information for the public. This research evaluated both basic use and design tenets and the presence of theoretical components to motivate healthy breast cancer behavior for users of the most frequented breast cancer websites.

Methodology

Each website was evaluated with 2 sets of questions. The first attends to the basic use and design tenets of the website. The second coding form assessed the websites use of 3 behavior change theories.

Results

The majority of the websites fared well with regards to their use of basic tenet and design principles. Theoretical components were used sparingly on the majority of websites.

Discussion

The most frequented breast cancer websites are currently well equipped with basic use and design tenets. However, their lack of behavioral change components is likely to impede users’ motivations to protect themselves against breast cancer.

Evaluation of Design and Theoretical Criteria

Eighty percent of American internet users visit health websites and over 50% claim that the visit influenced one or more of their health-related behaviors (Fox, 2006). The internet, however, can act as a double-edged sword. It can deliver well founded wisdom or make public incorrect information posted by a single person. For this reason, researchers have often called the quality of health websites into question (Benigeri & Pluye, 2003). Despite criticism, health website consumerism is increasing (Satterlund, McCaul & Sandgren, 2003).

Interested organizations have taken advantage of the power available through the internet to publicize health-related messages about breast cancer. As a result, breast cancer websites have been studied in an effort to assess their quality (Hoffman-Goetz & Clarke, 2000; Satterlund et al., 2003). Research studies on this topic have discussed the frequencies of basic design tenets possessed by breast cancer websites (Hoffman-Goetz & Clarke, 2000). However, thus far, popular breast cancer websites have yet to be researched for their inclusion of behavioral change theory components.

Communication researchers have documented the impact of key strategies on the success of motivating individuals to modify health-related behaviors. Having well-designed websites is not enough to motivate users to make healthy decisions. For this reason, the main goal of this research was to examine the use of behavioral change theory components present on breast cancer websites which inform about the potential for users to take protective steps against breast cancer. As a secondary goal, this study also documented breast cancer website's use of design components. In order to do this, three theories were employed in this research; the transtheoretical model (Prochaska, 1984; DiClemente & Prochaska, 1982; Prochaska, Velicer, DiClemente & Fava, 1988), the theory of planned behavior (TPB) (Ajzen & Fishbein, 1980; Ajzen, 1985), and the extended parallel process model (EPPM) (Witte, 1992). Of particular interest here is preventative and environmental content related to breast cancer that is present on breast cancer websites. Performing preventative behaviors for breast cancer are important because, according to the American Cancer Society (2007b), only 5 to 10% of breast cancer cases are genetically caused. In order to conduct this research, a search was performed to identify the most commonly accessed breast cancer websites and then these websites were coded using a basic tenet design coding form and a behavioral motivation coding form.

The remainder of this paper will provide an introduction to breast cancer and the internet, discuss the evaluation of basic design tenets and theoretical aspects of health websites including those for breast cancer, explain the three behavior change theories involved in this study, and present the methods, results, and conclusions of this study.

Breast Cancer and the Internet

According to the American Cancer Society (2007a), it is estimated that over 180,000 women will be newly diagnosed with breast cancer in the United States in 2008 and more than 40,000 will die as a result of the disease. As with many other diseases, women with breast cancer and their families are turning to the internet for health information. One study found 41% of breast cancer patients actively use the internet to find information on their disease; another study similarly reported 44% of breast cancer patients use the internet (Satterlund et al., 2003; Fogel, Albert, Schnable, Ditkoff & Neugut, 2002). Although the number of women who use the internet specifically to seek breast cancer prevention information is unknown, studies have shown that 60% of all online health website visitors are searching for some kind of preventative measures to stay healthy (Cain, Sarasohn-Kahn & Wayne, 2000). Adding this information to the findings that cancer is one of the top diseases searched on the internet and the fact that women are twice as likely as men to visit health websites makes studying preventative breast cancer content on websites a valuable endeavor (Satterlund et al., 2003; Howard, Rainie & Jones, 2001).

How Health Websites Have Been Studied

Building evaluation tools for health websites is not a new idea. Dozens have been created with the goal of aiding patients and researchers in determining the quality of any given health website (Gagliardi & Jadad, 2002). Websites, in general, have been studied largely for two aspects. First they have been studied for their “hygiene” aspects (basic design) which cause dissatisfaction in users if they are absent, and “motivator” aspects (content) which cause satisfaction in users when present (Zhang & von Dran, 2000). Website design studies on health websites and more specifically on breast cancer websites have documented key issues (Ivory & Megraw, 2005; Turchin & Lehman, 2000; Pendergrass, Nosek & Holcomb, 2001; Hoffman-Goetz & Clarke, 2000). Hoffman-Goetz and Clarke (2000) for example, concluded that the majority of breast cancer websites put users at a disadvantage in assessing the trustworthiness of the website. A second study done on Swedish breast cancer websites also concluded that the websites needed to renovate their coverage and correctness of information before they could be appropriate education tools for both the general public and professionals alike (Nilsson-Ihrfelt et al., 2004). Breast cancer websites have also been examined for their popularity. Meric and colleagues (2002) determined that more popular websites differed from less popular sites based on the topics they discussed.

Two specific areas of website design have become of particular interest to researchers as the internet has developed: usability and interactivity. Usability refers to “the ease of getting the web site to do what the user intends it to do” (Palmer, 2002b, pp.153). It has been shown to influence user satisfaction, user traffic, whether or not users return to a site, error reduction, and accuracy issues (Flavián, Guinalíu & Gurrea, 2006; Palmer, 2002a; Lecerof & Paterno, 1998). The criteria used to evaluate usability on sites has varied across studies, including judgments on dimensions of useful information, personal involvement, page loading time, navigation efficiency, and more (Agarwal & Venkatesh, 2002). Some studies on website usability have focused specifically on health for target populations such as older populations (Mead, Lamson & Rogers, 2001).

Ghose and Dou (1998) suggest that interactivity can actually enhance the usability of a site. Interactivity is defined as the “extent to which users can participate in modifying the form and content of a mediated environment in real time” (Steuer, 1992, pp.84). Interactivity has been called the best facet of health communication for the purpose of promoting health issues because of the unique abilities it offers users (Rice, 2001). Numerous studies on health related websites have included interactivity as a variable of interest (Eysenbach, Powell, Kuss & Sa, 2002; Wang & Liu, 2007; Stout, Villegas & Kim, 2001). Many health issues have taken advantage of the interactivity available on the internet and have created health interventions via the internet (Winzelberg, Classen, Alpers, Roberts, Koopman, Adams et al., 2003; Stout et al., 2001; Etter, 2006). The presence of interactivity within a website has been shown to have several positive outcomes, such as increased attractiveness and quality, effects on user satisfaction, user attitude, efficacy of information delivery, and efficiency of information retrieval, and increased potential for consumer attentiveness to information (Ghose & Dou, 1998; Teo, Oh, Liu & Wei, 2003; Berthon, Pitt & Watson, 1996). Despite these widely noted positive effects, however, interactivity has been shown to produce some negative outcomes, such as serving as distractions or possibly being less persuasive (Gerdes, 1997; Bezjian-Avery, Calder & Iacobucci, 1998)

Health websites have also been coded for their content. Antismoking websites have been evaluated for their informational content, persuasive strategies and design elements to better gauge their ability to deter youth from smoking. In 2005, Lin and Hullman coded 100 antismoking websites and found that the websites differed based upon source. One area where this was true focused on persuasive appeals. Industry sites were found to have the weakest appeals whereas grassroots websites provided the strongest appeals. Pro-eating disorder websites have been assessed for their use of the EPPM's variables (Lapinski, 2006). Using the search engine Google, 19 pro-eating disorder websites were found and 400 pages of content were coded revealing all of the EPPM's variables to be present. However, most were not well represented except for response efficacy, which accounted for 48% of the total messages on the websites. In 2003, Evers and colleagues wanted to assess websites for their use of the transtheoretical model, the health belief model, the theory of planned behavior and social learning theory. Two hundred and seventy-three health websites were evaluated that dealt with a variety of health topics. Forty-two were found to have what the authors termed at least four of the “5A” screening criteria to motivate users to undertake healthy actions which were advise, assess, assist, anticipatory guidance, and arrange for a follow up. Thirty-seven of those sites were evaluated and were termed to have strength in motivating users because of their ‘5A’ criteria. However, only nine of the health websites evaluated stated they were based upon communication theories of behavior change.

These past results of health website studies have offered insight into this powerful health communication tool, yet the assessment of such theories to motivate users toward healthy breast cancer prevention actions has not been studied. The benefit of broadening such endeavors to include behavior change theories is more than an academic exercise. Results from this assessment have the ability to reduce the incidence and mortality of breast cancer by providing a feedback loop to the designers and sponsors of the websites. Three specific theories that offer explanatory value regarding the effectiveness of website content are overviewed in the next section.

Three Contributing Theories

A variety of health behavior change theories were examined for possible use in this study. Three theories stood out due to their concrete definitions of the variables they predicted to impact behavior change that could be useful in evaluating websites. All three theories have been used successfully and extensively in the realm of health, however there is a dearth of studies which use these theories to examine information sent via communication technologies such as the world wide web. Used together these theories may provide great insight in to this academically unknown void. They all have similarities, such as the fact that they all assess self-efficacy. However, they also branch out into different areas that they believe are integral in behavioral motivation that can be provided by messages. Table 1 provides a report of the variables used in this study and the theory from which they were derived.

Table 1.

Variable	The Extended Parallel Process Model	Theory of Planned Behavior	Transtheoretical Model
Consciousness Raising			x
Dramatic Relief			x
Environmental Re-evaluation			x
Social Liberation			x
Self-Reevaluation			x
Stimulus Control			x
Helping Relationships			x
Counter conditioning			x
Reinforcement Management, Attitudes		x	x
Self-Liberation, Perceived Behavioral Control	x	x	x
Subjective Norms		x
Response Efficacy	x
Severity	x
Susceptibility	x

The theory of planned behavior was formed in 1985 by Ajzen based off of an earlier theory of reasoned action (TRA) (Ajzen & Fishbein, 1980). The TPB consists of three variables which together predict the likelihood a person will perform a specific behavior; attitudes, subjective norms, and perceived behavioral control. Attitudes are measured by a person's personal attitude toward performing a recommended action, and subjective norms are measured by significant others’ personal attitudes toward performing a recommended action. These two variables are the basis for the TRA and are believed to affect a person's behavioral intentions directly which in turn affects a person's actual behavior. Unlike the TRA, however, the TPB also takes into account perceived behavioral control which is a person's belief or nonbelief that they are able to perform a given action. Perceived behavioral control is not only credited with impacting behavior indirectly through intentions, but also having the ability to impact behavior directly on its own.

The variables of the TPB are useful in assessing attitudes, subjective norms, and perceived behavioral control as they impact behavioral intentions. However, the TPB does not account for threat, which is the central focus in fear appeals due to its ability to motivate people toward behavior change (Stephenson & Witte, 2001). The extended parallel process model was employed within this study due to its inclusion of threat.

The extended parallel process model is designed to measure the perceived threats and efficacy a person receives from a message. Threat is measured by perceived severity which “is an individual's belief about the seriousness of the threat” and perceived susceptibility, which “is an individual's beliefs about his or her chances of experiencing the threat,” (Witte, 1992, p. 339). Efficacy is measured by perceived self-efficacy, which “refers to an individual's beliefs in his or her ability to perform the recommended response,” and perceived response efficacy, which “refers to an individual's beliefs as to whether a response effectively prevents the threat” (Witte, 1992, p. 339). If a person perceives no threat, the EPPM states that a person will do nothing. However, if a threat is perceived, efficacy then plays the deciding factor in which of the two phases a person will take part in; fear control or danger control. Fear control takes place when a person perceives high threat but low efficacy and is therefore expected to avoid protective behaviors since they do not feel able to complete them. Danger control, on the other hand, occurs when a person perceives high threat and higher efficacy and is therefore motivated to take action to protect themselves from the threat by performing the suggested behavior.

The EPPM is valuable in assessing severity, susceptibility, response efficacy and self-efficacy, but it lacks a focus on other variables that could serve to motivate breast cancer website users through the different stages of breast cancer protection awareness, such as dramatic relief, consciousness raising, and environmental reinforcement. Therefore, the transtheoretical model was included in this study.

The transtheoretical model is also often also called the stages of change model (Prochaska, 1984; DiClemente & Prochaska, 1982; Prochaska et al., 1988). Its purpose is to demonstrate the stages and processes involved in behavior change. This frame gives researchers the ability to tailor messages for audiences in particular stages. The transtheoretical model has two main parts known as the stages of change and the 10 processes of change. The stages of change explain the five stages people may experience when making lifestyle changes such as changing their diet or quitting a bad habit. The first stage is called precontemplation, when the person has no intention of changing. The second stage is contemplation, when the person begins to have the intention of changing. The third stage is preparation, when the person has decided they are going to change and readies themselves for the change. The fourth stage is action, when the person actually changes. The fifth stage is maintenance, where the person continues living their changed lifestyle. The transtheoretical model was originally linear, but has been created as a spiral model, because the majority of people move backwards as well as forward through these stages.

The second part to the transtheoretical model is the 10 processes of change which were created to describe techniques that aid a person's motivation to change. The first five are termed experimental processes and are used in the early stages of change including, consciousness rising, dramatic relief, environmental reevaluation, social liberation, and self-reevaluation. The last five are termed behavioral processes and are valuable in the later stages of change including, stimulus control, helping relationships, counter-conditioning, reinforcement management, and self-liberation.

Summary of Theories

Employing key elements from three complementary theories at once allows researchers access to a plethora of information on how current breast cancer websites are equipped with the use of fundamental concepts from prominent theories to motivate users to take protective steps against breast cancer and how theses websites can be improved in the future with further use of theoretical underpinnings. The above described variables within these three theories can be measured in different ways. Conventionally, they are measured through directly asking an audience to self-report their attitudes or perceptions (Godin, Valois, Lepage & Desharnais, 1992; O'Connor, Carbonari & DiClemente, 1996; McMahan, Witte & Meyer, 1998). At other times these theories are used as a guideline for message creation and evaluation (Roberto, Meyer, Johnson & Atkin, 2000; Witte, Cameron, Lapinksi & Nzyuko, 1998; Sheeran & Silverman, 2003; Campbell, DeVellis, Strecher, Ammerman, DeVellis & Sandler, 1994, Woods, Mutrie & Scott, 2002; Reger et al., 2002). This research measured these variables in the second way explained. This second method of use for these theories has been transferred to the internet through measuring variables via a content analysis of messages present on web-sites (Lapinski, 2006; Evers, Prochaska, Prochaska, Driskes, Cummins & Velicer, 2003; Cheh, Ribisl & Wildemuth, 2003; Doshi, Patrick, Sallis & Calfas, 2003). Hence, this research used these theories to examine the possible impacts that breast cancer websites could have on their audiences under the assumption that the use of strategies from theories would lead to potential behavior change. This innovative thinking may even lead to a new metatheory specifically tailored for use on newer technologies such as the Internet. In addition, the presence of basic design tenets on breast cancer websites has only been examined superficially. Therefore, the aims of this study were twofold: to assess breast cancer websites’ use of basic design tenets and behavior change theory components.

Research Questions

Based upon the foregoing arguments, two research questions were formed.

RQ1: How do the most frequented breast cancer websites employ basic use and design tenets?

RQ2: How many of the most frequented breast cancer websites include certain behavioral motivation techniques as measured by the TPB, the EPPM, and the transtheoretical model?

Methodology

This study was conducted in three stages. First, formal searches were performed to identify the most commonly frequented breast cancer websites. A coding scheme was then applied to assess basic tenets and design qualities followed by application of a coding scheme to assess use of underpinnings from the three theoretical models.

Search Strategy

The most frequently used search engines (Google, Yahoo and MSN) according to the Nielsen/NetRatings (Bausch & Han, 2006) were used to find the most frequented breast cancer websites. Next, the keywords “breast cancer” were entered into the search boxes of Google, Yahoo and MSN. The first 100 resulting websites were saved from all three. Websites that discuss environment related issues about breast cancer are of key importance to this study due to funding from the U.S. National Cancer Institute and the U.S. National Institute of Environmental Health Sciences for projects that explore environmental factors that may be related to breast cancer. For this reason a separate search was also done on all three search engines using the keywords “breast cancer” and “environment.” These 600 resulting links were then scanned thoroughly for overlap and such sites were deleted. After the exclusion criteria and overlap, only 157 of the 600 websites were chosen to analyze for this study (see table 2 in appendix for inclusion/exclusion criteria).

Next, in order to find a concrete checklist by which websites could be evaluated, the World Wide Web and online electric journal services were searched. This study employed questions from Anderson et al.'s (1999) website evaluation checklist, McLachlan's (2002) “WWW Cyberguide Ratings for Website design” and “WWW Cyberguide Ratings for Content Evaluation.” All three of these tools are available to view online (see references). In addition, questions were developed specifically for this project to address unique inquires such as the presence of interactive components.

After being evaluated for basic use and design tenets, each website was searched and evaluated for its use of theoretical components. The coding schemes for theoretical components on the websites were created by employing key concepts from the three behavior change theories chosen for use.

Due to the extreme breadth of some of the websites, only content that could be reached in three clicks or less was included here as this has been found to be the extent to which most website users search a website (Dhillon, Coss & Hackney, 2001). This process excluded “Enter Site” pages. However, fourth links that could be seen from a third link, and that were self-explanatory, were included (i.e., “How to do a breast self exam.”)

Coding Scheme

Before beginning to code the websites, intercoder reliability for each of the categories of the evaluation sheets were established. To do this, two research assistants separately evaluated 16 websites. Reliability was then determined by the Cohen's Kappa which ranged from 73% to 100% on all of the variables. After satisfactory reliability was established, the two research assistants coded the remainder of the websites from September of 2007 until November 2007. Upon completion, the data were entered into statistics software (SPSS) to run descriptive statistics.

Results

Frequencies of variables provided interesting insight into the current condition of the most frequented breast cancer websites regarding both their use of behavioral change theory components. Basic use and design tenets are also reported for the reader's knowledge. Information provided by such frequencies allows assessment of breast cancer websites’ ability to serve their audiences. See tables 3 and 4 in the appendix for the coding schemes used, the results of each variable, and the Cohen's Kappa obtained for each variable.

Basic Use and Design Tenets

Several basic traits of the most frequented breast cancer websites were revealed through this research. The majority of the websites were run by nonprofit organizations (72%).Websites that had a main focus on breast cancer (45%) were nearly as common as those that had a main focus on some other noncancer topic (43%). Sixty-six percent of the websites were appealing to a general audience, whereas 34% were appealing to a specific audience. Fourteen percent of websites appealed to multiple ethnicities, 17% appealed to multiple age groups, and 1% appealed to those with disabilities. Additionally, no specific placement category for navigation bars accounted for more than 27% of all websites. The total number of graphics on a website ranged from 1 to 340, but 75% of the websites had less than 49 graphics on pages within one click of its navigation bar.

This research also revealed that many of the basic use and design tenets were employed by the most frequented157 breast cancer websites. One hundred percent of the websites listed their source, had homepages that downloaded within three seconds, worked both on the Mozilla “Firefox” browser and on Internet Explorer, used color well to enhance their website, and had a left-aligned layout. Similarly, 80% or more of the websites were attractive, used graphics appropriately (they did not contain irrelevant graphics), removed their out of date information, had properly working interactive elements (if any), presented their information in a stable and clear manner, provided contact information, contained what seemed to be accurate facts, had fair and unbiased information, did not have any errors or obvious omissions, had a navigation bar that was easy to use, and had the page title appear on the top window bar. Findings also indicated that 61% of websites had information regarding risk factors for breast cancer.

However, some basic use and design tenets were not represented on many of the most frequented breast cancer websites. Only one website posted a hit rate. Over 56% of websites did not include a date of last revision, contact information on every page, the webmaster's name, copyright restrictions, a help section, a site map, or consistent navigation throughout the website. Interactive elements were not frequently found on websites either. Fourteen percent had surveys, 8% had chats, 4% had games, 17% had message boards, and 23% had some other kind of interactive element. Additionally, if a website did contain irrelevant graphics, they accounted for more than 10 percent of the website's total graphics on 89% of such websites indicating heavy inappropriate graphic use.

Also, websites generally did not include basic health information, such as what breast cancer is, what scientific words related to breast cancer meant, and benefits and risks associated with treatment. Only 31% of websites had information on environmental risk factors. Information on treatment for breast cancer was the most widespread health information and was found on 47% of the websites. In the realm of diversity, around 30% of websites were available in more than one language (all were available in English). Additionally, around 23% of websites contained material that was readable by someone with less than a ninth-grade education according to the Flesch-Kincaid measurement.

By simply stating the findings regarding use of basic design tenets this research will leave up the quality of these websites via design to the reader. What is key to this study is determining if the content on websites is presented in a way that will have the potential to motivate protective breast cancer behaviors. The next section presents the results in the incorporation of motivational designs as prescribed by the TPB, the EPPM, and the transtheoretical model.

Behavior Change Theories

The majority of websites did not fare well when evaluated for their use of behavioral change theory components or other motivational variables. Only one category was represented by over 80% of websites (use of statistics) and only six categories were included in over 50% of the websites (creating awareness about early detection or prevention, creating awareness for types of treatment, providing negative consequences for taking steps in a direction against healthy breast cancer behavior, providing positive consequences for taking steps in a direction toward healthy breast cancer behavior, informing users of the serious negative consequences of breast cancer, and informing users of breast cancer research findings). If there was an author present on the website, factors that build credibility and trust were also present, such as using credible language (81%) and remaining consistent in their views (95%). The remainder of the findings for these variables will be reported under the theory from which they were drawn. Also, note that this study included acts of advocacy as equal to a breast cancer protective behavior; this is because many websites believe advocacy can protect people from breast cancer through aiding research and therefore finding cures and causes for the disease.

Theory of Planned Behavior

The TPB assesses how attitudes, subjective norms, and perceived behavior control can lead to behavior change. Attitudes were measured by determining how many websites either provided positive consequences for taking steps toward healthy breast cancer habits (56%) or provided negative consequences for taking steps away from healthy breast cancer habits (53%). One site did both of these by telling users a healthy lifestyle would decrease their chances of getting cancer and drinking alcohol would increase it. Also, 17% of websites which provided positive consequences did so for the act of giving money to their organization. One site tells users that they will be making a difference in the fight against breast cancer by donating money or buying their products.

Subjective norms were evaluated by asking if the website conveyed that signifi-cant others think the user should engage in healthy preventative breast cancer behaviors. Less than 3% of websites did this. One website which provided health information for children did this by telling children to tell their mothers to get mammograms. Perceived behavioral control was present in 17% of the websites as measured by the question “Does the website have a section to create in users the belief that one can change to healthy breast cancer protection behaviors?” However, 7% of these messages involved the giving of money. For example, one site describes becoming a member (through donation) as quick and easy. Analysis of perceived behavior control for TPB also asks if the website conveys to the user that they have the ability to overcome obstacles to act out protective breast cancer behaviors. Only 6% of websites contained such a message. One site did this by letting users know where to get free mammograms, therefore solving the problem of health care costs. Due to low use of the concepts of attitudes, subjective norms, and perceived behavioral control on breast cancer websites, TPB would not predict that users would be motivated to change their behavior to include healthy breast cancer preventative action by the vast majority of the most frequented websites.

The Extended Parallel Process Model

The EPPM states that the presence of high perceived threat and high perceived efficacy together will motivate people toward behavior change. Perceived behavioral control is also apart of the EPPM which was already analyzed by the TPB. However, perceived behavioral control is called self-efficacy in the EPPM and does not include helping people overcome obstacles in their way of performing a desired behavior. The EPPM also deals with response efficacy which was assessed by asking if a website gave information on the effectiveness of certain behaviors to avert the threat of cancer. One website lets users know that physical activity reduces a person's chances of getting breast cancer specifically by 39%. Under the EPPM, web-sites’ ability to make the user feel vulnerable to getting breast cancer was evaluated to assess the concept of susceptibility. Seventy-two percent of websites created this effect, either by giving information on the number of women that get breast cancer per year or by telling users they had a one in eight chance of getting breast cancer. The EPPM also assesses the severity of a threat. Nearly 60% of websites gave information on the serious negative side effects attached to breast cancer. Often times, websites did this by listing death rates of women who get breast cancer. According to the EPPM, breast cancer websites may actually encourage people to avoid healthy preventative breast cancer behaviors because their components of threat are higher than their use of efficacy components. This leads to fear control and maladaptive behavior.

The Transtheoretical Model's Processes of Change

The transtheoretical model includes 10 processes of change which motivate people toward behavior change. Forty percent or more of the websites used some form of consciousness raising by raising awareness of the definition of breast cancer, ways to prevent or detect breast cancer, the serious outcomes involved with breast cancer, the fact that males can get breast cancer, or treatment options for breast cancer. Less than 36% of websites had any type of dramatic relief by either using emotional appeals, or telling breast cancer stories in first or third person. For instance, many websites made use of testimonials of people who have gone through breast cancer, such as one that has a “stories of inspiration” section. Less than 3% of websites provided at least one type of environmental reevaluation tool which consisted of explaining how one's personal habits toward breast cancer affect their social environment or explaining how one can be a positive or negative role model to others regarding breast cancer. For example, one website told users that by joining their organization they would not only be helping the fight against breast cancer themselves, but they would also be empowering others around them to do so.

Social liberation was demonstrated in 22% of websites that informed users about social opportunities they could take part in to act out health preventative breast cancer behaviors, and 38% of websites increased alternatives for people to act out healthy preventative breast cancer behaviors. However, 27 % of the websites that increased alternatives did so for money. For instance, one website tells users they can help beat breast cancer by donating in several different ways, such as purchasing products they make, giving their vehicles or stocks, or by using their credit card online. Websites demonstrated the variable of self-reevaluation only through showing healthy role models for breast cancer protection, which 22% of websites did. For example, one site provided examples of healthy role models by having a spotlight section for volunteers who were helping their cause to fight breast cancer. Websites included the variable of stimulus control through adding prompts for healthy breast cancer protection habits (8%). One site did this by telling women to perform breast self-exams on the same day each month to help them to remember.

Helping relationships were evaluated by checking if a website had personal or general social support for users. Personal support was provided by 48% of websites, and general support was provided by 22%. One site provided personal support by having a hotline for users to call so they can talk to someone about breast cancer issues, and they also provided general support with their informational brochures meant to aid women experiencing breast cancer concerns.

Counterconditioning was done by 14% of websites which gave users substitutes for problem behaviors. One site does this by advising users to substitute refined grains for whole grains in their diet to decrease their risk of breast cancer through maintaining a healthy lifestyle. Reinforcement management was already discussed under the section for TPB as this concept is very similar to TPB's concept of attitudes. Self-liberation was also discussed under the TBP as perceived behavioral control, although like the EPPM, the transtheoretical model's idea of self-liberation does not include helping people overcome obstacles in their way of performing a desired behavior. It does, however, include urging users to commit to the belief they can perform an action and urging users to take action on that belief. None of the 157 websites urged users to commit to the belief that they could perform an action but 16% of websites were found to urge users to take action on their belief. This tactic was used 8% of the time on websites for the purpose of urging user to donate. Again, low use of these theoretical components indicates that the majority websites would not be likely to motivate people into changing their behaviors regarding breast cancer prevention based on the theoretical model's ten processes of change.

Discussion

This study applied specific attributes of three behavior change theories to the most frequented 157 breast cancer websites and also reported on their basic design tenets. Studying this has revealed both weaknesses and strengths of the websites. These findings are paramount because if the guidelines analyzed by this research are not met by breast cancer websites, such websites are most likely not accomplishing the mission of encouraging women to enact protective breast cancer behaviors– a mission which many breast cancer organizations hold as their main purpose.

The three theories used for this study, the TPB, the EPPM, and the transtheoretical model, were useful in demonstrating the lack of strategic behavior change motivators present on the most frequented breast cancer websites. The only theoretical components present on the majority of websites were consciousness raising, negative attitudes for taking steps away from preventative breast cancer actions, positive attitudes for taking steps toward preventative breast cancer actions, and threats. If the website listed an author, which occurred in 53.5% of the websites, it displayed characteristics that warranted trust and credibility. A specific issue for alarm is the fact that the majority of the websites had both components of threat based on the EPPM: susceptibility and severity. However, a low percentage of web-sites had either type of efficacy the EPPM describes: self-efficacy and response efficacy. According to the EPPM, high threat coupled with low efficacy would create the need for website users to undergo fear control and would therefore cause them to respond in maladaptive ways. This might entail avoiding breast cancer prevention behaviors. For the transtheoretical model and TPB, the absence of their components on the majority of websites merely indicates the majority of websites are not likely to encourage users to change their behaviors regarding breast cancer prevention based upon the assumptions of these two theories. Also, at times, when these elements actually are used, it is for the purpose of receiving donations.

These results indicated that changes are needed for the majority of breast cancer websites in order to motivate them toward healthy preventative breast cancer behaviors. Breast cancer websites can be improved by employing strategies from the three theories to encourage behavior change, especially if they use threats on their website. Additionally, when targeting a general audience, breast cancer websites should also have sections that cater to individual needs based on ethnicity, literacy, language, age, or disability.

In summary, although it is laudable that the most frequented breast cancer websites are incorporating most necessary basic design tenets, many changes still need to be made in terms of these websites use of behavioral change theories. More attention and effort should be put into message strategy design so that these websites may research people both cognitively and emotionally.

This study had limitations. The internet is a dynamic medium that is always changing. Hence, some of the information found in this study may already be out of date. Navigation challenges were also encountered in this study. Only information available within the first three clicks of a website's homepage was included in this study. Pertinent information could have gone undiscovered because of this choice as some of the websites were massive. Additionally, the websites included in this study were chosen based on hit popularity. This strategy might have missed some high quality websites that, for whatever reason, do not have high hit rates.

These findings open doors of interest for future research. Specifically, the relationship between the presence of threat on breast cancer websites without the accompanied presence of efficacy should be studied with regards to its effect on users. Is this indeed causing a boomerang affect as the EPPM would predict? Another area of interest would be the presence of messages on websites promoting prevention versus those that promote early detection on the most frequented breast cancer websites. Finally, the concept of behavior change theory components being used to increase the likelihood that users will donate should be further investigated.

Figure 1.

The theory of planned behavior model from Ajzen and Madden, 1986, p. 458.

Figure 2.

The extended parallel process model from Witte, 1992, p. 338.

Figure 3.

A spiral model of stages of change from Prochaska, DiClemente, and Norcross, 1992, pp. 1104.

Biography

Pamela Whitten, Ph.D., is an associate professor in the Department of Telecommunications at Michigan State University. She is also a Senior Research Fellow for Michigan State's Institute of Healthcare Studies. In her current position, Dr. Whitten is responsible for conducting technology and health-related research as well as teaching graduate and undergraduate telecommunications courses. Dr. Whitten's research focuses on the use of technology in health care with a specific interest in telehealth and its impact on the delivery of health care services and education. Dr. Whitten's Ph.D. is in organizational communication from the University of Kansas; her M.A. is in communication from the University of Kentucky, and her BS is in management from Tulane University.

Address – 287 Communication Arts & Sciences Building Michigan State University, East Lansing, MI 48824, USA

Sandi Smith, Ph.D., is Director of the Health and Risk Communication Center and Professor in the Department of Communication at Michigan State University where she teaches courses in interpersonal communication, relational communication, communication theory, and persuasion. Her research interests parallel these course topics. In specific, she is focusing her research on health communication, the portrayal of interpersonal relationships in the media, and the impact of memorable messages received from important others. Sandi received her Ph.D. and her M.A. in Communication from the University of Southern California. Her B.S. is in Psychology from the University of California at Irvine.

Address – 572 Communication Arts and Sciences Building Michigan State University, East Lansing, MI 48824, USA

Samantha Munday, B.A., is an graduate student in the Health Communication Masters Program at Michigan State University.

Address – 438 Communication Arts and Sciences Building Michigan State University, East Lansing, MI 48824, USA

Carolyn LaPlante is an undergraduate in the Communication Department at Michigan State University

Address - 572 Communication Arts and Sciences Building Michigan State University, East Lansing, MI 48824, USA

Appendix

Table 2.

Inclusion/Exclusion criteria for websites

Inclusion Criteria	Exclusion Criteria
Definition of a website taken from Answers.com which uses the American Heritage Dictionary “A set of interconnected web pages, usually including a homepage, generally located on the same server, and prepared and maintained as a collection of information by a person, group, or organization.”	The URL leads to a website composed only of links to other websites, it contains no original information regarding breast cancer
The website URL must be found on one of six lists used for this research.	The URL leads to only research or news articles
At least one of the articles regarding breast cancer is original to the site (i.e.- it cannot be barrowed from another source)	Due to required logins the average person would not be able to access the majority of the website. If in one attempt researchers failed to log in and reach information on the website, it was not used.
The URL leads to an actual website which is working.	The URL leads to a website where the main intent is to sell a product
	The basic URL is already on our list (However: if the basic URL leads to a home page of its own with a navigation bar it will be included in the list as a new website)
	The URL does not lead to more than one article (i.e. – it cannot be a free floating htm or PDF with no link to an attached website)
	The URL leads to a website for a breast cancer conference or grant which has no actual information on the disease of breast cancer.

Table 3.

Basic Tenets N = 157

Variable	% No	% Yes	Cohen's Kappa
Is there a hit rate available on the website?	99.4	0.06	94.7
Does the website list a source?	0	100	100
Dose the website target a general audience?	33.8	66.2	84.2
Does the website target a specific audience?	66.2	33.8	82.4
Does the website's homepage download within 3 seconds?	0	100	100
Does the website work on both Mozilla Foxfire and Internet Explorer web browsers?	0	100	100
Is the website's homepage attractive?	4.5	95.5	89.47
Does the website list a purpose?	26.8	73.2	94.7
Does the website's homepage have a copyright and date?	42	58	100
Does the website have a last revised date?	70.1	29.9	94.7
Have all out of date materials been removed from the website?	15.9	84.1	98.47
Is the website updated often?	34.4	65.6	94.7
Is information on the website stable throughout?	5.1	94.9	100
Is a working link provided for a contact person or address for further information?	9.6	90.4	94.7
Are there working links to other websites that relate to users needs/purposes?	20.4	79.6	84.2
Is the content of most (over half) of the working linked sites relevant to breast cancer issues?	6.4	73.2	89.47
Is contact information available on every page?	72.6	27.4	94.7
Is there a single individual author/webmaster presented for the website?	84.1	15.9	94.7
Are individual authors listed for specific pages/articles within the website?	46.5	53.5	84.2
Is information presented clearly on the website (in a manner that is easy to follow)?	2.5	97.5	94.7
Are copyright restrictions clearly indicated on the website?	60.5	39.5	94.7
Do facts on the website seem to be accurate?	1.3	98.7	100
Does the website appear to be fair and unbiased?	2.5	97.5	94.7
Are there any errors in the website's content?	96.8	3.2	100
Are there any omissions on the website's content?	98.7	1.3	89.47
If the website has errors are they significant?	1.3	98.7	100
If the website has omissions are they significant?	1.3	98.7	89.47
Is any biased information presented as fact?	98.1	1.9	89.47
Is it possible that the website may have a conflict of interest?	96.2	3.8	94.7
Do all graphic elements include informational content (implies a traceable, significant, logical connection to the website)?	17.9	82.1	100
Does the website's use of color enhance the website?	0	100	94.7
Are any of the website's noncontent bearing graphics animated?	5.1	12.8	100
Does the website have any flashing, scrolling, or otherwise visually distracting graphic or text displays?	71.3	28.7	89.47
Do most of the website's graphics that are links have equivalent text links?	41.4	33.8	84.2
Is the most important information on the website located near the top of the homepage? (Users don't have to scroll down to see it)	24.8	75.2	89.47
Is the website in left layout?	0	100	84.2
Does the website contain any pop up menus?	38.2	61.8	84.2
Does the website's interactive elements work?	1.3	41.4	94.7
Does the website's multimedia elements work?	2.4	33.1	89.47
Is navigation of the website easy to use?	1.9	98.1	100
Does the website have a navigation bar?	2.5	97.5
Does the website have a search engine?	32.5	67.5	100
Does the website have a help section?	84.7	15.3	94.7
If the website has a help section is it easy to use?	3.8	12.1	94.7
Does the page title appear on the web browser's top window bar?	3.2	96.8	89.47
Does the page location appear on the web browser's top window bar?	27.4	72.6	89.47
Does each page of the website have a link back to the home page?	47.8	52.2	84.2
Does the website impair a user's ability to use their back button?	96.8	3.2	89.47
Does the website's navigation system remain constant throughout the website?	56.7	41.4	78.9
Does the website have a site map?	58.6	41.4	100
Does the website have any dead links?	63.7	36.3	84.2
Does the website have any sections that cost money to view or use?	99.4	0.6	100
If the website has a section that costs money to view or use is that cost reasonable?	0.6	0	100

Variable	% Under 10%	% Over 10%	% N/A	Cohen's Kappa
If the website does have dead links what percent of the total number of links on the website are dead?	31.8	4.5	63.7	84.2

Variable	%Author's name	%Name with credentials	%Name with credentials and contact info	%Does not apply no author	%N/A	Cohen's Kappa
If the website has a single author/webmaster, information about the author includes:	4.5	1.9	2.5	75.8	15.3	84.2

Variable	% 0–49	% 50–100	% 101–150	% 151–200	% 201–250	% 251–300	% 301–350	Cohen's Kappa
How many total graphics are there on the website within one click from the homepage's navigation bar?	75.2	16.6	6.4	1.3	0	0	0.6	93.75
N = 156

Variable	%Ten percent or below	%Above ten percent	%N/A	Cohen's Kappa
What percent of the websites total graphics are irrelevant?	1.9	15.9	81.25	94.7

Variable	%0	%1 to 3	%4 to 6	%more than 6	Cohen's Kappa
How many priority 1 errors does Webexact indicate the website has?**	23.6	75.8	0.6	0	94.7

Variable	%0	%1 to 4	%5 to 8	%more than 8	Cohen's Kappa
How many priority 2 errors does Webexact indicate the website has?**	2.5	66.9	30.6	0	94.7

Variable	%0	%1 to 5	%6 to 10	%more than 10	Cohen's Kappa
How many priority 3 errors does Webexact indicate the website has?**	0	100	0	0	94.7

Variable	%Most are	%Some are	%Few are	%N/A	Cohen's Kappa
If some of the website's interactive elements are not working, what portion does work?	0	0.6	0	99.4	100

Variable	%Most are	%Some are	%Few are	%N/A	Cohen's Kappa
If some of the website's media elements are not working, what portion does work?	0	0	0.6	99.4	100

Variable	% Top of the screen	% Bottom of the Screen	% Top and bottom of the screen	% Left of the screen	% Top and left side of the screen	% Top, left and bottom of the screen	% Bottom and left of the screen	% other	Cohen's Kappa
Where is the navigation bar located on the website's homepage?	13.4	5.7	26.8	17.2	6.4	10.8	6.4	13.4	84.2

Variable	%.org	%.gov	%.edu	% .com	% From another country (ie, ca, uk, au)	% Other
What does the website's address end in?	54.8	7	5.7	26.1	3.8	2.4

Variable	% Cancer in general	% Breast cancer	% Environment	% Female Specific Organizations	%Other	Cohen's Kappa
Is the website's overall purpose centered around cancer in general, breast cancer, the environment, females, or does it have some other purpose?	15.9	47.8	5.7	3.8	42.7	100

Variable	% Profit	% Nonprofit	Cohen's Kappa
Source	23.6	76.4	94.7

^**Webexact uses W3C (The World Wide Web Consortium) guidelines to check for priority errors. W3C is an international organization which has developed ways to evaluate websites to ensure they are accessible to the entire public regardless of disabilities. Priority one errors are identified as ones that must be fixed in order for the a website to meet W3C guidelines, priority 2 errors should be fixed, and priority 3 errors may be fixed. For the complete information on priorities visit Web Content Accessibility Guidelines. 1.0 W3C Recommendations online at http://www.w3.org/TR/WCAG10/#priorities.

Table 4.

Theoretical Variables N=157

Variable	% No	% Yes	Cohen's Kappa
Does the website have a section for creating breast cancer awareness by providing a definition?	56.1	43.9	94.7
Does the website have a section for creating awareness about the outcomes associated with breast cancer?	58.6	41.4	84.2
Does the website have a section for creating awareness about prevention or early detection of breast cancer?	22.3	77.7	89.47
Does the website have a section for creating awareness about treatments for breast cancer?	42	58	100
Does the website have a section for creating awareness about male breast cancer?	56.1	43.9	94.7
Does the website have a section to move people emotionally?	77.7	23.3	73
In the section meant to move people emotionally does it also reduce this emotion by giving users appropriate action to take?	82.8	17.2	73
Does the website have a section for first person story telling about breast cancer?	64.3	35.7	78.9
Does the website have a section for third person story telling about breast cancer?	77.1	22.9	87
Does the website have a section to explain how the presence of personal habits toward preventing breast cancer affects their social environment?	98.7	1.3	100
Does the website have a section to explain how the absence of habits toward preventing breast cancer affects their social environment?	100	0	100
Does the website have a section to create awareness that one can serve as a positive role model?	97.5	2.5	89.47
Does the website have a section to create awareness that one can serve as a negative role model?	100		100
Does the website have a section to increase social opportunities to act out healthy breast cancer behaviors?	77.7	22.3	89.47
Does the website have a section to increase alternatives to act out healthy breast cancer behaviors?	62.4	37.6	78.9
Does the websites list of alternatives to act out healthy preventative breast cancer behaviors involve the giving of money?	13.4	26.8	100
Does the websites list of alternatives to act out healthy preventative breast cancer behaviors only involve the giving of money?	10.8	14.6	81.25
Does the website have a section to urge users to evaluate their self image with healthly preventative breast cancer habits?	100	0	100
Does the website have a section to urge users to evaluate their self-image without healthy breast cancer habits?	100	0	100
Does the website have a section to show role models for healthy breast cancer behaviors?	78.3	21.7	78.9
Does the website have a section to remove cues for unhealthy breast cancer habits?	100	0	89.47
Does the website have a section to add prompts for healthy breast cancer habits?	91.7	8.3	89.47
Does the website have a section to promote the learning of healthier behaviors that can substitute for problem behaviors?	86	14	84.2
Does the website have a section for providing negative consequences for taking steps in a direction against healthy breast cancer actions?	47.8	52.2	78.9
Does the website have a section for providing positive consequences for taking steps in a direction toward healthy breast cancer actions?	43.9	56.1	89.47
Are users given positive consequences for donating money?	38.9	16.6	87.5
Are users given positive consequences only for donating money?	13.4	3.2	93.75
Does the website have a section to create in users the belief that one can change to healthy behaviors?	82.8	17.2	84.2
Does the website have a section to urge users to commit their belief that one can change to healthy breast cancer behaviors?	100	0	100
Does the website have a section to urge users to act on their belief that one can change to healthy breast cancer behaviors?	84.1	15.9	89.47
Does the section which creates the belief in users that they can change to healthy breast cancer behaviors involve the giving of money?	9.6	7	93.75
Does the section which urges users to act on their belief that one can change to healthy breast cancer behaviors involve the giving of money?	8.3	7.6	93.75
Does the section which urges users to act on their belief that one can change to healthy breast cancer behaviors only involve the giving of money?	1.9	5.1	100
Does the website have a section to convey that important significant other(s) think the user should engage in healthy breast cancer behaviors?	97.5	2.5	94.7
Does the website have a section to convey to users that they have the ability to overcome obstacles in order to perform healthy breast cancer behaviors?	93.6	6.4	89.47
Does the website create awareness that diagnosis of breast cancer is likely for users by giving risk statistics?	28	72	84.2
Does the website create awareness that there are serious negative consequences for those who are diagnosed with breast cancer?	41.4	58.6	94.7
Does the website have a section that gives information on the effectiveness of certain behaviors to avert the threat of breast cancer?	75.2	24.8	84.2
Does the website have a section for the purpose of providing social therapy occurring in communication networks, operating for the purpose of helping people to cope with incongruities?	52.2	47.8	78.9
Does the website have a section for the purpose of providing general social therapy occurring in communication networks operating for the purpose of helping people to cope with incongruities?	77.7	22.3	94.7
Does the website have a section to inform users about anything that increases a person's chances of getting breast cancer?	39.5	60.5	89.47
Does the website have a section to inform users about anything in the environment that increases a person's chances of getting breast cancer?	69.4	30.6	94.7
Does the website have a section to inform users about breast cancer diagnosis?	70.7	29.3	89.47
Does the website have a section to inform user about breast cancer scares?	68.8	31.2	78.9
Does the website have a section to inform user about breast cancer treatment?	53.5	46.5	89.47
Does the website have a section to inform users about the monetary costs associated with breast cancer?	89.8	10.2	84.2
Does the website have a section to inform users about the benefits of breast cancer treatment?	72.6	27.4	89.47
Does the website have a section to inform users about the risks of breast cancer treatment?	72	28	89.47
Does the website have a section to define breast cancer terms?	70.7	29.3	84.2
Does the website's source have a title that warrants credibility?*	28.6	71.4	94.7
Does the website's source use language that warrants credibility?*	19	81	89.47
Does the website's source have an education that warrants credibility?*	26.2	73.8	94.7
Does the website's source have professional training that warrants credibility?*	23.8	76.2	94.7
Does the source appear to argue against their own interest?*	100	0	100
Does the source maintain a consistent view that warrants trust?*	3.6	96.4	89.47
Does the website's source take a surprising stand that warrants trust?*	100	0	100
Does the website have chats?	92.4	7.6	93
Does the website have surveys?	86	14	93
Does the website have games?	96.2	3.8	93
Does the website have message boards?	83.4	16.6	84.2
Does the website have any other additional interactive elements?	77.1	22.9	81
Is the website useable for users who read more than one language?	70.1	29.9	93
Is the website useable for users from various literacy levels?	77.1	22.9	97
Does the website contain certain devices specifically for users with disabilities?	99.4	0.6	100
Does the website have a section which applies to users from a variety of age groups?	83.4	16.6	81
Does the website have a section which applies to users of more than one ethnicity?	86	14	81
Does the website have information about the conclusions of credible studies that have been done in the field of breast cancer research?	30.6	69.4	93
Does the website make use of statistics?	15.3	84.7	87
Does the website have a section with information regarding topics of interest to the users which it targets?	100	0	100
Does the website combat any misconceptions its target audience may have about breast cancer?	80.9	19.1	93
Does the website employ breast cancer advocates from their target community for assistance in addressing topics of interest?	59.2	40.8	87
Does the website give users specific steps to complete in order to perform healthy preventative breast cancer behaviors?	71.3	28.7	81
Does the website ask for donations?	48.4	51.6	93.75

^*N = 84, only completed if the website had an individual author to assess credibility for.