Exploratory Research Focusing on Oral Cancer Prevention: Challenges of Dealing With Informational and Cognitive Barriers

Information Seeking and Avoiding as Key Factors for Effective Health Promotion

Active health information seeking is known to have a strong influence on health-related behavioral intentions and decisions and is determined by psychosocial, sociodemographic, socioeconomic, and environmental factors (Johnson & Case, 2013; Lambert & Loiselle, 2007). Cancer-related issues cause psychological distress and uncertainty which is particularly relevant in the view of the fact that most people do not possess comprehensive knowledge on this topic (Hesse, 2009). Information acquisition can help to reduce uncertainties, remove ambiguities, improve preventive behaviors, and support decision making on early detection and treatment options (uncertainty management; see Brashers, 2001). For these purposes, individuals can choose from many different sources such as mass media, websites, family members, and friends as well as health professionals (Kelly et al., 2010).

Research shows that people are more susceptible to information if they feel involved in the issue, if they are able to understand the information given, if they perceive it relevant, and if they expect to derive benefits from it (e.g., Rimal & Juon, 2010; Rimal & Real, 2003; Turner, Rimal, Morrison, & Kim, 2006). These studies also provide evidence that people have a higher motivation to deal with health information if they perceive risks for themselves and if they experience self-efficacy for themselves.

However, people do not always welcome health information and also engage in information avoidance, which is a common behavior in everyday life (Narayan, Case, & Edwards, 2011). For example, Emanuel et al. (2015) showed for a representative U.S.-American sample that 39% of the population preferred to avoid information about personal cancer risks.

The uncertainty management theory (UMT; Brashers, Goldsmith, & Hsieh, 2002) suggests that the type of health-related information behavior (seeking or avoiding) is closely linked to the individual’s perception of uncertainty which is defined by the perceived discrepancy between one’s current level and the desired level of uncertainty. In a state of uncertainty, information seeking can reduce uncertainty that is distressing, whereas information avoidance can maintain perceived uncertainty that allows for hope and optimism (Barbour, Rintamaki, Ramsey, & Brashers, 2012; Brashers et al., 2002; Brashers et al., 2000). Thus, information behavior can both be used to reduce, to maintain, and to raise uncertainty (Barbour et al., 2012). Uncertainty is not necessarily a negative and uncomfortable emotional state but can even be desirable in certain situations (Barbour et al., 2012; Brashers, 2001; Brashers et al., 2002). Distancing oneself from potentially existing worries and perceptions of uncertainty can help to reduce cognitive dissonance in everyday life (Festinger, 1957). When the individuals’ uncertainty appraisal suggests that more information about the topic or risk might cause discomfort and raise cognitive discrepancies they might not be able to cope with, they tend to decide to maintain a certain degree of information insufficiency and thus avoid further information on the risk. Thus, it might also be useful to distance oneself from undesired messages and to purposely avoid information (van’t Riet & Ruiter, 2013).

Furthermore, the strategy of uncertainty management depends on the individual’s skills to gather information. To successfully find information can be complicated by the nature of information or the complexity of information-seeking behaviors (Brashers, 2001). Accordingly, self-efficacy assessments influence health-related information behaviors. Especially, under the condition of low self-efficacy, the disposition toward information avoidance becomes more likely (Barbour et al., 2012; see also Goodall & Reed, 2013). Taken together, cancer information is multifaceted, complex, emotionally distressing, and often very disturbing. This can increase the tendency to not actively seek or even avoid information to reduce complexity and to cope with these uncertainties.

According to UMT (Brashers, 2001), our study assumes that both information seeking and avoiding are a rational strategy for coping with health-related issues. Against this background, we aimed to explore the underlying cognitive patterns of people’s information seeking or avoiding behaviors, taking into consideration the issue-specific problem awareness, people’s beliefs, and their openness toward information and preventive advice regarding oral cancer.

Participants and Procedure

To comprehensively deal with the purpose of the study, we decided to follow a triangulation approach. We involved three different subgroups in our sample that provided different perspectives on the object of research. The first group consisted of people at risk who are predominantly male, aged 60 years and above, with low socioeconomic status (SES), living in urban areas with an excessive alcohol and/or tobacco consumption. This is the main risk group for oral cancer (Hertrampf et al., 2012).

To comprehend problem awareness and risk perception, informational needs and communication behaviors with special regard to symptom detection, we included persons with diagnosed oral cancer and skin cancer in the face being in aftercare treatment as a second group of participants. The majority of patients are men at the age of 60 years and above (Andersen et al., 2008; Conway et al., 2008). Their retrospective assessments were supposed to help adjust the campaign strategy to the effect that people at high risk who might have already experienced some symptoms that should be clarified might become susceptible to the campaign. For this subsample, it has to be considered that they are probably not suitable to uncover the typical patterns of information behaviors as their self-perceptions and experiences as patients will have overlaid their memories.

As a third subgroup of our sampling strategy, we additionally recruited relatives of oral cancer patients because the individuals’ health behaviors and health-related decisions are also significantly influenced by the social environment (Sallis, Owen, & Fisher, 2008). This is particularly important in the context of oral cancer as the people at risk are predominantly males at an older age who are known to be supported by their wives and family members in regard to health-related information and medical decision making.

In October and November 2011, 28 persons were recruited to take part in the study: 18 subjects at risk (SR); 7 subjects affected from oral cancer or with skin cancer in the face (SA); 3 relatives of subjects affected (RSA). A description of the sample is given in Table 1. People diagnosed with oral cancer and skin cancer in the face and their relatives were recruited in a German Clinic of Oral and Maxillofacial Surgery in the setting of their regular aftercare by the medical staff. People at risk were recruited via different local social service providers, an alcoholics support group, and by additional snowball sampling. Before any appointments for interviews were made, every participant received written information on the project and expressed his or her written consent. The data were strictly handled in an anonymous form only. The project was approved by the Ethical Committee of the University Hospital of Kiel, Germany (AZ: D 462/11).

OPEN IN VIEWER

Table 1.

Sample Description.

	SR	SA	RSA	Total
Male	n = 12	n = 5	n = 1	n = 18
Female	n = 6	n = 2	n = 2	n = 10
Total	n = 18	n = 7	n = 3	n = 28

Note. The abbreviations classify the type of target group: SR = subject at risk; SA = subject affected from oral cancer; RSA = relative of subject affected. They were also used to anonymize the quotations.

Overall, 19 interviews (14 single interviews, four interviews in pairs, and one interview in a group setting) were conducted. This mixture of different interview settings resulted from the different situational parameters of access to the field and the participants’ preferences. Our primary objective was to offer rather natural and comfortable interview settings and to lower the barriers of participating as far as possible. Originally, only individual interviews were planned as method of data collection. In four cases, patients recruited in the clinic were accompanied by their spouses and preferred to be interviewed in pairs. Recruitment of participants with very low SES and excessive alcohol and tobacco consumption who fit our sampling criteria turned out to be particularly challenging. We only had the opportunity to talk to six members from a support group for former alcoholics. They requested to realize the interview setting as a group session. We are aware of the fact that these different types of interviews have different methodological implications. For example, participants in the group setting may have been influenced by group dynamics, and the group characteristics may have affected the results. But our research objective was only to explore the different patterns of information seeking and dealing with information related to oral cancer by single persons. Analogous to the single interviews, the interviews in pairs and in the group setting focused on the collection of information from individuals and not on interpersonal group dynamics and interactions between participants, not on uncovering the role of relationships or situational mechanisms of social interactions. So, we did not motivate participants in the pair or group setting to interact with each other, and we addressed each respondent with our questions separately. This means, we approved this methodological lack of conclusiveness caused by different interview settings in favor of field work pragmatics, the participants’ preferences, and comfort.

In the first step of computer-assisted qualitative content analysis, every interview was analyzed separately and the results were considered regarding differences which may stem from the different interview techniques. As data analysis did not indicate interpretable differences between the participants’ statements in different interview settings, we did not split the results according to the interview technique. To provide transparency, references to the participants’ statements are marked with the type of interview (“I” = individual interview; “P” = interview in pairs; “G” = group setting).

Persons with diagnosed oral cancer were interviewed after their regular aftercare in the Clinic of Oral and Maxillofacial Surgery in a conference room of the clinic. The other interviews were conducted on the participants’ request in their homes or in the rooms of the local social service providers. The interviews took between 35 and 90 min. At the beginning of each interview, the participants were informed about the study, and they agreed to give written consent for their participation.

Interview Guide

A flexible research approach was required to determine the individuals’ barriers and triggers to dealing with oral cancer and to identify motivational and emotional patterns, underlying feelings of uncertainty, and the openness toward information about the topic. All interviews were conducted face-to-face and tape-recorded. The same semistructured interview guide was used in all interviews. The questions comprised the participants’ level of knowledge and their cancer-related decisional strategies, information and communication needs, patterns, and barriers. The participants were asked to describe their perceptions and experiences regarding cancer in general as well as oral cancer in particular and to make recommendations for suitable ways of information and message strategies in oral cancer prevention. The key topics were addressed as open questions by the interviewer. Supplementary questions were used to deepen an issue and to maintain a natural flow of conversation. The facilitator was a female communication researcher and member of the project team, being experienced in qualitative interview techniques.

Analysis

The goal of qualitative analysis is the enhancement and widening of our perspectives on social phenomena. Thus, it is important to identify the different attitudes, opinions, and behaviors regardless of how many people expressed them. The focus is on the individual, and we aim at identification and understanding—not at quantification—of individual patterns of mind-sets. Thus, instead of frequencies, we concentrate on the identification of different phenomena.

Each of the interviews was transcribed verbatim. Numbers were used as pseudonyms for the individual participants. A qualitative content analysis based on the principles of coding step by step according to Mayring (2000) was conducted for the whole text material by one empirical researcher who also realized the interviews, experienced in computer-assisted qualitative data analysis. First, the material is divided into content analytical units that are subsequently coded with categories being in the center of the whole analytic process. The coding scheme was developed by combining inductive and deductive strategies of developing and applying categories by means of coding (Mayring, 2000): The major topics and dimensions of analysis (e.g., the individuals’ health and risk status, health consciousness, health, risk, and prevention behavior; informational needs, health information–seeking behavior, evaluation of information obtained; cancer-related experiences, cognitions, and emotions) were derived from the interview guide and served as main categories of the first draft of the coding scheme (deductive category application). This basic structure was differentiated and modified during the course of qualitative data analysis and helped to identify and to organize more specific codes developed from the participants’ single statements (inductive category development). In this process, the large amount of material was reduced to those units of meaning being particularly relevant for the scope of research. Going through several stages of iterative category development and assessment, modifications and revisions of codes (feedback loops) allowed discovering reasonable structures in the data and theoretical saturation. The coding was carried out by Atlas.ti (version 6.2; Friese, 2012). The data set contains every interview-transcript (primary documents), the units of analysis (quotations), and every category (codes) that was used and links these codes to the original quotations of the study participants in the transcripts. This allows others to comprehend and re-enact the whole process of qualitative data analysis. A second researcher of the team comprehensively reconstructed the analysis of coding, identification, and interpretation of patterns described in the next section.

One problem we want to mention is that the quotations used to prove and to illustrate our findings and interpretations were in German and had to be translated for this article. Of course, every translation is always also an interpretation of text, and our post hoc translations carry the risk of being also shaped by our theoretical consideration and interpretations. Therefore, one person with and two people without any relatedness to the study and to communication science translated the quotations for us, and we compared the translations.

Seeking Information About Oral Cancer

Overall, the participants of our study appeared to be rather restrained and of low proactivity regarding information and communication seeking. They predominantly described themselves as taking a passive role in communication matters. Even if they were affected from oral cancer or were aware of their risk status, they did not consequently want to deal with the topic in more detail to avoid raising fears and uncertainty. For example, a patient’s husband stated that he was not interested in gathering more information because this would have only stirred him up: “And you’ve got the feeling that you’re in good hands so that I don’t have to inform myself through all kinds of media. Because that’s just upsetting and makes me feel unsure” (RSA, male, 70–80 years old, P).

Patterns of Self-Distancing

The next step of the qualitative analysis was to identify the main dimensions characterizing the different determinants of interest and openness toward oral cancer–related information and communication which provide the main explanations and dimensions of self-distancing. These were the components of patterns of self-distancing which were identified in the subsequent step of analysis. The following categories indicating personal, situational, and contextual factors were extracted from the interviews as main reasons for the rather passive self-conceptions accompanied by inactive information and communication behaviors:

These factors have a close match to the determinants of health-related information seeking and avoiding behaviors outlined in the literature (Johnson & Case, 2013; Lambert & Loiselle, 2007). The character of our data allowed us to gain deeper insight into the underlying cognitive mechanisms. Based on our interview data, we reinterpreted these determinants and their interrelations against the individual backgrounds of our participants’ complex and dynamic situation-dependent contexts of life. Four multidimensional patterns of self-distancing from the problem of oral cancer and of managing uncertainties and fears associated to this topic were identified.

Limitations and Future Research

This qualitative research approach aimed at looking beyond the surface and existing scientific knowledge about risk awareness and knowledge about oral cancer and information seeking and avoiding behaviors among people at risk (Hertrampf et al., 2012; Horowitz et al., 2002; Patton et al., 2004; Scott et al., 2008; Scott et al., 2011). This methodological approach is in line with studies focusing on early detection and oral cancer patients’ attitudes and behaviors (Scott et al., 2009; Scott, Grunfeld, Main, & McGurk, 2006; Scott et al., 2011) and underlines that qualitative methods of data collection and analysis are indispensable for exploring in depth the personal and cultural constructions of cancer and risk perceptions, informational needs, and decision-making processes dealing with life-threatening and emotionally charged health issues such as cancer (Hay & Craddock Lee, 2009; Sandelowski, 2004). Of course, using an interview guide restricts openness of the interview in favor of comparability of answers, which is always a decision based on weighting pros and cons of semistructured interview techniques.

The strategies of self-distancing were derived as distinctive motivational patterns, but individuals are not necessarily limited to one of these perspectives. Our results rely on participants who were able and willing to articulate their feelings and thoughts and who retrospectively reported their experiences and behaviors. Some of our interview partners came from educationally disadvantaged backgrounds, were physically and mentally impaired and socially disadvantaged (e.g., homeless, disorientated). This had an effect on their ability of verbalization, on their capacities to articulate their feelings and thoughts and to describe their experiences and behaviors. Here, our situational and problem-oriented way of data collection reaches its limits. Instead, an ethnographic approach would have provided deeper insights.

Moreover, there might be a methodological bias due to the fact that different interview techniques—single interviews, interviews in pairs, one interview in a group setting—were combined which might have affected the findings. Possible effects of the group setting were not taken into account or could be controlled during data collection and analysis. We used the same interview guide, treated the group members during the interviews as individual interview partners, and analyzed each participant’s data as individuals. The decision for this methodological approach was based on the aim of comparability. This is accompanied by the fact that the potential of the group situation has not been exhausted. In general, the aim of comparability and the use of an interview guide can be critically evaluated because it decreased openness and subject-centeredness.

Based on the identification of the target group for oral cancer prevention, this study was conducted as the second step to guide the development and implementation of an oral cancer prevention campaign strategy, particularly, addressing people at risk in Northern Germany. We identified four distinctive patterns of information (non)motivation which should be regarded in the progress of the formative analysis and later on in the campaign planning process. Based on our qualitative data and explorative analysis, we can develop hypotheses on the interrelations between cognitive mechanisms and different strategies of health information avoidance. A focus on the avoidance of information still seems to contrast with most information-campaign research (Case et al., 2005). Further steps of formative evaluation providing evidence on the prevalence and empirical relevance of these cognitive patterns and their interrelations to different types of information avoidance by quantitative research might help to guide oral cancer campaign development. Against the background of the theoretical considerations outlined in this article and our exploration of different self-protective behaviors and uncertainty management strategies by self-distancing from the topic and by avoiding information, we can first conclude that information avoidance cannot be treated as the opposite pole to information seeking, that is, avoiding is not equivalent to not seeking. This requires to develop theoretical models which specify information seeking and avoiding behaviors separately and not only by using one outcome variable covering information seeking as a continuum with extensive seeking and avoiding as two conditions of the same phenomenon. Second, our data indicate that information avoiding is not a unidimensional construct which is only rarely mentioned in the literature to date (e.g., Case et al., 2005). Instead, there are different types of avoiding which need to be differentiated and explained differently.