MSc Thesis
Profiling Cancer Patients: Using Cluster Analysis based
on Motives for Online Health Information-Seeking,
Sociodemographic, Health-related and Medical
Characteristics
Song L.L. Duimel (12513598) Graduate School of Communication
Persuasive Communication, University of Amsterdam, The Netherlands Independent Foundation Kanker.nl
Dr. A.J. Linn Date: January 31, 2020
Acknowledgement
First and foremost, I wish to thank all the respondents from Kanker.nl who were willing to complete the extensive online survey on cancer patients’ online motives to seek health information. Next, I like to send my gratitude to the independent foundation Kanker.nl for their approval and help in recruiting patients online.
I am very thankful for all the support, feedback, suggestions and constructive criticism that I received from my supervisor Dr. A. J. Linn to whom I would like to give special thanks. Not only do I appreciate her convincing reassurance and interest into my research, I am also very much grateful for her motivation, enthusiasm and knowledge which took this thesis to the next level. I could not have imagined having a better supervisor.
Besides my supervisor, I thank Prof. Dr. J. C. M van Weert and Prof. Dr. E. M. A. Smets for their beliefs in me. I feel very honored to be given the opportunity to join their research. Furthermore, I am grateful to Dr. E.S. Smit for helping me get through the data analysis, and my dearest friends who helped me during the last steps. In addition, to my family and friends in general, I am thankful for them showing interest and encouragement.
Last but not least, I would like to show special appreciation to my Dad for his support, encouragement, and feedback, but most of all his unconditional love throughout my life. Song Duimel,
Abstract
Background: Understanding which cancer patient profiles exist is important to improve health information provision. Health providers often find it difficult to tailor information to the information needs of patients. Because of these unmet information and support needs, patients are likely to turn to the Internet. To date, no prior research is known to identify patient profiles according to cancer patients’ motives for online health information-seeking. Objective: This study aims to identify cancer patient profiles based on their motives to seek online health information in order to assist health providers to tailor their patient referral to online health information.
Participants and methods: A total number of hundred and seventy-eight (ex-)patients and ten close relatives responded to the online survey. Participants’ sociodemographic, health-related characteristics, and motives for their online health information-seeking behavior (HISB) were examined. Hierarchical cluster analysis was used to identify the profiles. Results: Data analysis revealed four profiles that show differences in the motives to seek online health information: the “high-information seeker” (34.4%), the “high-information and support seeker” (23.9%), the “moderate-information seeker” (28.6%), and the “moderate support seeker” (14.1%). In addition to these differences, clusters differed on psychological distress (i.e., intrusive thinking), intolerance for uncertainty and eHealth literacy.
Conclusion: The results of this study reveal that heterogeneity of cancer patients’ motives to seek for online health information exists. Health providers could use the profiles as tailoring method to refer their patient to online health information and support or platforms relevant to them. Furthermore, triangulation of qualitative and quantitative research is needed to gain more understanding of cancer patients’ online health information-seeking behavior. Keywords: cancer, online health information-seeking, cluster analysis, patient profiles, motives
Table of contents
Introduction 5
Use of the Internet 8
Online Health Information-Seeking Behavior 10
The Social Determination Theory 11
Motives to seek online health information and support 12
Materials and Methods 17
Kanker.nl 17
Participants and Procedure 17
Motives for Cancer Patients’ OHISB 18
Measures of Predicting Factors 19
Statistical Analyses 22
Results 25
Sample Characteristics 25
Visiting Kanker.nl and motives 25
Results of the cluster analysis 27
Discussion 31
Main Conclusions 31
Practical Implications 32
Study Limitations and Recommendations 32
Conclusions 36
References 37
Appendices 53
Appendix A. Model with the motives for OHISB 53
Appendix B. Motive categories 54
Appendix C. Ward’s Linkage Dendrogram 56
Appendix D. Sociodemographic, health-related, medical characteristics 57
Appendix E. Frequency of self-reported visits Kanker.nl 58
Appendix F. Results of one-way ANOVA for four clusters 59
Appendix G. Results of Chi-squared tests for four clusters 60
Appendix H. Request form participants Kanker.nl 62
Appendix I. The information letter for participants 67
Introduction
Having cancer is for most people a life changing event (Kim, Shah, Namkoong, Mctavish, & Gustafson, 2013; McCaughan & McKenna, 2007). The diagnosis causes a large amount of uncertainty, anxiety and fear among patients and their close relatives. To deliver high quality of care, the provision of adequate information is essential (Bol, et al., 2018; Matsuyama, Kuhn, Molisani, & Wilson-Genderson, 2013). More specifically, patients need support to know and understand their disease (i.e., cognitive needs) and to feel acknowledged and understood (i.e., affective needs) (Bensing & Verhaak, 2004; Goerling, et al., 2019; Sanders, et al., 2020). When visiting a provider, studies have shown that cancer patients vary in the information needs they have (Bol, Linn, Smets, Verdam, & Van Weert, 2020; Ellis & Varner, 2018; Goerling, et al., 2019). These needs differ depending, among other things, on patients’ sociodemographic and health-related characteristics (Nölke, Mensing, Krämer, & Hornberg, 2015), time since diagnosis (Sheehy, et al., 2018), information satisfaction (Goerling, et al., 2019), and feelings of social support (Fogel, Albert, Schnabel, Ditkoff, & Neugut, 2002).
Previous research indicated that health providers find it difficult to tailor the
information to the various needs (Russell & Ward, 2011; Skalla, Bakitas, Furstenberg, Ahles, & Henderson, 2004; Zeguers, et al., 2012). As a consequence, many patients experience unmet needs (Faller, et al., 2016; Goerling, et al., 2019; Tustin, 2010). Because of these unfulfilled needs, they may use other sources such as the Internet to fulfill them (Fiksdal, et al., 2014; Goerling, et al., 2019; Sanders, et al., 2020; Tustin, 2010). The Internet is
particularly a convenient medium since it provides abundant information at any time (Fiksdal, et al., 2014; Shaw, et al., 2008). In other words, the Internet gives people the opportunity to be self-active in finding information and support relevant to them (Anderson,
Rainey, & Eysenbach, 2003; Anker, Reinhart, & Feeley, 2011; Caiata-Zufferey & Schulz, 2012).
Today, much research exists that describes how cancer patients seek online (Mattsson, Olsson, Johansson, & Carlsson, 2017; Nelissen, Van den Bulck, & Beullens, 2018) and which predictors are associated with patients’ online health information-seeking behavior (OHISB) (Ginossar, 2016; Li, Orrange, Kravitz, & Bell, 2014; Nölke, Mensing, Krämer, & Hornberg, 2015). Less research has been done into what actually drives (i.e., motives) cancer patients to seek online (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Sanders, et al., 2020). Study findings show that patients often try to find cancer-related information for various reasons, such as to understand their disease, to complement, validate and/or challenge the information given by their health provider (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Sanders, et al., 2020; Tustin, 2010), to connect to and/or seek social support from fellow cancer patients, to exchange and/or ask for experience (Lobchuk, McClement, Rigney, Copeland, & Bayrampour, 2015; Sanders, et al., 2020), or as a result from
dissatisfaction with patient-physician relationship or medical encounters (Moumjid, Gafni, Bremond, & Carrere, 2007; Neumann, et al., 2011; Tustin, 2010). However, to date, no prior research is known that explored the predictors of the different motives to seek online. As previous research in patient-provider communication showed that needs differ depending on patients’ sociodemographic, health-related and medical characteristics (Germeni & Schulz, 2014; Goerling, et al., 2019; Jo, Park, & Jung, 2019), it may be possible that different predictors are associated with these motives. Many patients report to use the Internet before encounter with their health provider (Eheman, et al., 2009; Linn, et al., 2019). Hence, health providers should be aware that patients may be receiving and requesting for information outside of their healthcare settings (Anker, Reinhart, & Feeley, 2011; Kemp, et al., 2018; Maddock, Lewis, Ahmad, & Sullivan, 2011). Since individuals differ substantially from each
other, a “one-size-fits-all” approach to refer patients to a general website might not be as relevant (Zeguers, et al., 2012). It is therefore of paramount importance to examine the existence of different cancer patient profiles according to patients’ motives and to examine which predictors belong to a sound profile. The results of this study can assist health
providers by using the different patient profiles as a tailoring method to refer their patients to information, support, or online platforms relevant to them (Li, Orrange, Kravitz, & Bell, 2014). This may not only enhance the provision of health information (Jo, Park, & Jung, 2019), but also improve and increase cancer patients’ participatory role (Ford, Schofield, & Hope, 2003).
The current study therefore aims (1) to examine the sociodemographic, health-related, medical characteristics and motives of cancer patients who seek health-related information and support online, and (2) to identify patient profiles according to these motives. Hereby the research question “To what extent can distinct patient profiles be identified, based on the
motives of cancer patients to seek online health information and support?” gives structure to
the study. The current study adds to the academic field in several ways. By adopting a cluster analysis as a method, this study approaches the online information and support needs of cancer patients in a completely innovative way on a methodological level. With regard to cancer patients’ motives to seek information and support online, no other research is known to adopt this method. Finally, by identifying different cancer patient profiles according to diverse online information and support needs this study acknowledges the heterogeneity that exists among cancer patients’ motives to seek online.
Use of the Internet
As the rapid growth of the Internet has led to the proliferation of a large amount of information available, an increasing number of patients makes use of the Internet to access health information and receive support (Donovan, LeFebvre, Tardif, Brown, & Love, 2014; Eheman, et al., 2009; Jacobs, Amuta, & Jeon, 2017). As a result, online health information has become influencial in medical care (Herrmann-Werner, et al., 2019) and the term “e-patients” is often used to refer to informed health consumers (i.e., patients and their family and/or friends) who use the Internet to inform or educate themselves about health (Ferguson, 2007; Ferguson & Frydman, 2004; Herrmann-Werner, et al., 2019). In 2005, a survey in seven European countries revealed that 71% of all adults reported to use the Internet to seek health information (Andreassen, et al., 2007). This compared to the United States where 56% to 79% of US Internet users go online for health information (Cotten & Gupta, 2004; Hesse, et al., 2005). In the Netherlands, research has shown that 67% of all adults aged 12 years and over sought online for health information in 2018 (Centraal Bureau voor de Statistiek, 2019). The Netherlands had hereby with 72% of people aged 16 to 74 the highest percentage of all 28 Eurpean countries in 2018.
The Uses and Gratifications (U&G) can be drawn upon to understand why cancer patients use the Internet and no other media sources, such as the television and print media (Rubin, 2002; Tustin, 2010). Although other media sources, such as television and print media, can play a role in the dissemination of health information, the Internet nowadays takes on the most of this function (Carlsson, 2009; Cotten & Gupta, 2004). The U&G proposes that all media users have particular communication needs that can be met by either both face-to-face interactions or certain media, such as traditional and/or new media (Rubin, 2002;
Stafford, Stafford, & Schkade, 2004). Each communication channel is chosen on the basis of their appropriateness to satisfy those needs (i.e., gratification). Cancer patients’ unfulfilled
information needs can thus be considered to be the most important drivers that encourage them to turn to the Internet (Case, 2007; Savolainen, 2012).
Patients, however, do not only make the rational decision to seek online for health information (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010). The perceived advantages of the Internet also contribute to people’s assumption about its ability to satisfy their unfulfilled needs. The Internet gives cancer patients the opportunity to be self-active in seeking online health information (Anderson, Rainey, & Eysenbach, 2003; Anker, Reinhart, & Feeley, 2011). In addition, the Internet has been suggested to be a cost effective and key factor in reducing the need for support from the health care system (Elbert, et al., 2014). Not only does it save time in finding new information, it has also the benefit of maintaining anonymity (Fox & Duggan, 2013). Furthermore, in the eyes of patients, the Internet is a source of empowerment to take a more active role in their own medical care, to make
informed decisions or choices, or even self-medicate (Hesse, et al., 2005; Jirasevijinda, 2015; Simmons, Baker, Schaefer, Miller, & Anders, 2009). Thus, with its convenience to offer up-to-date information and peer support (Yli-Uotila, Rantanen, & Suominen, 2013; Wallin, Mattsson, & Olsson, 2016), the Internet appears to be helpful to many to obtain information and/or support in one’s own time (Kemp, et al., 2018). Lastly, the different modes of
communication and forms of content are a specific characteristic that makes the Internet an appealing sources of information (Anderson, Rainey, & Eysenbach, 2003; Prestin, Vieux, & Chou, 2015). While some of its online health platforms provide the exchange of social support in blogs and on discussion forums (Prestin, Vieux, & Chou, 2015; Sanders, et al., 2020), others provide health information in libraries and question-and-answer pages (Anderson, Rainey, & Eysenbach, 2003; Prestin, Vieux, & Chou, 2015; Verberne, et al., 2019). Moreover, previous research has demonstrated that fundamental differences exist between and within the platforms with regard to its content (Sanders, 2020). Each type of
platform may influence the fulfillment of patients’ information and support needs. Therefore, from the perspective of the U&G, each different online platform might be chosen on the basis of its assumption to accommodate the different types of needs that patients have.
Online Health Information-Seeking Behavior
After diagnosis and during everyday life cancer patients are often in the need to get their questions answered by health practitioners and to receive support to cope with this disease (Goerling, et al., 2019). Adequate provision of information and support can help to manage their disease (Blödt, et al., 2018) and reduce high levels of uncertainty and anxiety (Feltwell & Rees, 2004; Kim, Shah, Namkoong, Mctavish, & Gustafson, 2013). This is important because high levels of distress negatively affect a patient’s overall well-being (Street, Makoul, Arora, & Epstein, 2009). Bensing and Verhaak (2004) propose a stress-coping framework that takes on a patient perspective to medical encounters. According to this model, patients have two basic needs when they encounter a health provider, cognitive and affective needs. Cognitive needs refer to information and the need to know and understand, whereas affective needs refer to the need to feel known and understood. While some patients preferably need information about treatment and the symptoms of their disease, others may feel the need to receive reassurance through social support (Sanders, et al., 2020). Previous research showed that health providers often use patients’ sociodemographic characteristics to determine patients’ information needs (Russell & Ward, 2011). For example, they may assume that female cancer patients have higher information needs than their counterparts based on prior research (Fallowfield, Jenkins, & Beveridge, 2002). Yet, this strategy seems not always to be reliable since research shows that cancer patients turn to the Internet as a result of information and support needs being unmet (Faller, et al., 2016; Goerling, et al., 2019; Jo, Park, & Jung, 2019). Also, each patient experiences their own cancer journey
differently, which causes them to require different information and support (Neumann, et al., 2011). Therefore, in order to be able to understand why people information and support online, it is necessary to first understand why patients have these different motives. An explanation for this can be found in the Social Determination Theory.
The Social Determination Theory
The Social Determination Theory (SDT) can be used to understand the origin of patients’ motives to seek online health information (see Figure 1 Appendix A) (Ryan & Deci, 2000; Lee & Lin, 2016). According to this theory, people can experience three innate
psychological processes that act as basic needs for their initial motivation to go online,
namely competence, autonomy and relatedness. Depending on these needs, patients can either be assertive and engaged or passive in how they surf online. The extent to which cancer patients feel the need to be competent, autonomous and related is also determined by the amount of information and support they receive from their social and external environment (e.g., healthcare environment, family, friends) (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Ryan & Deci, 2000).
In brief, competence refers to the degree to which a patient feels capable to effectively manage his or her health outcomes (Smith, Wallston, & Smith, 1995). More often,
competence is referred to as patients’ self-efficacy to cope with the cancer or to execute coping behavior to manage the disease and its treatment (Kohlmann, Janko, Ringel, &
Renovanz, 2019). Thus, this feeling can be considered rather cognitive in which it is expected that cognitive needs are most dominant. According to DeCharms (1968), feelings of
competence will not enhance patients’ motivation to go online unless they are accompanied by patients’ perceived autonomy. This means that patients must also experience a feeling of self-determination to actually address their cognitive needs by seeking online health
own actions and behavior (i.e., self-governing) (Patrick & Williams, 2018). In the context of patient care, this concept refers to the right that patients have to make informed decision about their own care without interference or influence from others (Beauchamp, 1994; Martinez, Kurian, Hawley, & Jagsi, 2015). The last component of the SDT refers to the need for “relatedness” or in other words, to feelings of being related or close with peers (e.g., cancer patients) and feelings of connection and sharing (Patrick & Williams, 2018; Sundar, Bellur, & Jia, 2012). Thus, this feeling can be considered rather affective in which it is expected that affective needs are most dominant.
Motives to seek online health information and support
Thus, based on the stress-coping model and the SDT, different main motives are identified. Empirical evidence identified different motives that can be grouped under these motives. Previous research showed that many cancer patients use the Internet to prepare themselves before consultations (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Echlin & Rees, 2002). Patients indicate that using the Internet before a consultation not only helps them to understand the medical language and terminology, but also to better understand their diagnosis and discuss treatment. Moreover, patients who have a monitoring coping style (i.e., patients who are generally more in the need of information and experience more
distress) prefer to be involved in the decision-making around their health (Rood, et al., 2015; Rood, et al., 2017). For this reason, high monitoring patients desire to have more information that helps them to make informed decisions. Particularly information about the likelihood of cure, advance of the disease, and the diverse treatment choices are searched for. Thus, patients’ practice of online health information-seeking is crucial to their participatory role. Yet, most often patients seek online after their consultation (Li, Orrange, Kravitz, & Bell, 2014). During consultation health practitioners find it both difficult to determine cancer patients’ information needs as well as to tailor their information provision (Russell & Ward,
2011; Zeguers, et al., 2012). Additionally, they struggle with the brevity of the patient-physician encounter (Hardy, 2008). It is then perhaps no surprise that patients experience unmet information and support needs (Tustin, 2010). The additional information is obtained to complement the information given by their health provider (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Pereira, Koski, Hanson, Bruera, & Mackey, 2000; Protière, Moumjid, Bouhnik, Le Corroller Soriano, & Moatti, 2012) or to reduce uncertainty (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010). Most commonly, patients report to seek and prefer medical information about treatment and side effects, diagnoses and prognosis and new interventions (Ellis & Varner, 2018; Kemp, et al., 2018; Pereira, Koski, Hanson, Bruera, & Mackey, 2000). Next to the need to understand and to reduce uncertainty, feelings of anxiety and the feeling to have a sense of control may lead to the need to complement information (Ellis & Varner, 2018). For example, patients may perceive their health
provider’s decisions as a threat to their autonomy which causes them to passively accept the decisions (Zanchetta & Moura, 2006). Patients will then seek additional health information to reduce anxiety and receive support (Yli-Uotila, Rantanen, & Suominen, 2013). Knowledge in this sense not only helps patients to be able to make informed decisions and engage in shared decision-making, but also contributes to feelings of competence (Zanchetta & Moura, 2006). Also, various factors, such as intrusive thinking, cancer worry and intolerance of uncertainty, can indicate the level of uncertainty and are suggested to affect patients’ information needs (Christianson, 1992). For example, cancer worry has been proven to be motivator for and predictor of patients’ OHISB (Li, Orrange, Kravitz, & Bell, 2014; Van Stee & Yang, 2017)
Furthermore, patients use the online health information to validate and/or challenge the information that is given by their health provider (Attfield, Adams, & Blandford, 2006; McMullan, 2006; Linn, et al., 2019; Sanders, et al., 2020). This can be a result of
Tustin, 2010). Particularly, confidence and general trust in health providers can predict whether patients go online to validate or challenge the information that is given (Bell, Hu, Orrange, & Kravitz, 2011; Gill & Whisnant, 2012). Additionally, patients search the Internet after their consultations to verify the given diagnosis and prescribed treatment (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010). Previous research indicates that
particularly patients who feel that their health providers information does not correspond with their own experiences or with other obtained information from, for instance, family and friends use the Internet after a consultation to verify the information provided by the provider (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Sanders, et al., 2020). Some patients search for alternatives to what was discussed during a consultation. In this case, the Internet serves as a second opinion platfom to seek for other treatment or medication
possibilities (Caiata-Zufferey, Abraham, Sommerhalder, & Schulz, 2010; Lund, et al., 2009). Patients who are expected to prepare, complement and validate are patients who enjoy seeking information (Cacioppo & Petty, 1982). Cancer patients with high NFC tend to be active seekers or detailed health information seekers and are intrinsically motivated to engage in thinking with regard to their disease (Berzonsky & Sullivan, 1992). Patients with low NFC on the other hand, do not engage in cognitive efforts to process information and tend to ignore and avoid new information (Venkatraman, Marlino, Kardes, & Sklar, 1990). In contrast with patients with low NFC, patients with high NFC are more likely to seek online for health information, to reflect upon the information and successfully process this
information under stressful conditions such as the diagnosis of cancer (Oh, Meyerowitz, Perez, & Thornton, 2007). Patients, however, also need the digital health literacy or eHealth literacy skills to seek online to fulfill their information needs (Norman & Skinner, 2006; Van der Vaart & Drossaert, 2017). According to Norman and Skinner (2006), eHealth literacy can be defined as “the ability to seek, find, understand, and appraise health information from
electronic sources and apply the knowledge gained to addressing or solving a health problem” (p. e9). Patients’ eHealth literacy can thus influence not only which information needs they have (Davis, Williams, Marin, Parker, & Glass, 2008), but also to what extent they seek and use online health information (Li, Orrange, Kravitz, & Bell, 2014).
Apart from patients’ NFC and eHealth literacy, various sociodemographic characteristics have been associated with going online (Cotten & Gupta, 2004; Nölke, Mensing, Krämer, & Hornberg, 2015; Shaw, et al., 2008). Despite some differences across studies, cancer patients’ OHISB in general seems to be represented by a similar profile that is indicative by characteristics such as gender, age, education and marital status (hereafter, “living situation”). Women are more likely to have higher information needs than men (Matsuyama, Kuhn, Molisani, & Wilson-Genderson, 2013) and to seek health information on the Internet than men (Nölke, Mensing, Krämer, & Hornberg, 2015). They often have a higher level of education, are of younger age (Ginossar, 2016; Mattsson, Olsson, Johansson, & Carlsson, 2017; Nölke, Mensing, Krämer, & Hornberg, 2015), and have a partner (Nölke, Mensing, Krämer, & Hornberg, 2015).
Next to sociodemographic factors, several health-related characteristics may be associated with the OHISB of cancer patients. Most of the previous studies only examined patients’ information needs based on either a single type of cancer (Echlin & Rees, 2002; Fogel, Albert, Schnabel, Ditkoff, & Neugut, 2002; Kim, Shah, Namkoong, Mctavish, & Gustafson, 2013) or multiple in general (Faller, et al., 2016; Rutten, Arora, Bakos, Aziz, & Rowland, 2005). Yet, although information needs differ by tumor site (Okuhara, et al., 2018), it is questionable whether these differences also apply to the motivations that underly cancer patients’ OHISB (Germeni & Schulz, 2014). Lastly, multiple studies have shown that
information needs arise at every stage of the disease (Matsuyama, Kuhn, Molisani, & Wilson-Genderson, 2013; McCaughan & McKenna, 2007; Sheehy, et al., 2018). Information needs
near the initial diagnosis may be directed at the understanding of the disease and its
treatments, but as time goes by information needs may change. For instance, treatments may be new, changed or ended which can leave patients with new questions at each new phase of their cancer journey.
Lastly, previous findings indicate that patients use the Internet to gain social support and feel related. Especially forums have the potential to provide cancer patients with different types of support (Sanders, et al., 2020). Users answer questions, refer others to external sources, or reach out to relate with one another’s experiences (Verberne, et al., 2019). Patients who experience a high emotional well-being and low perceived social support, are more likely to seek for support online compared to patients with high perceived social support and high emotional well-being. The Internet is often used to follow peer stories (i.e., stories or experiences of fellow cancer patients or people in the same situation) (Nelissen, Van den Bulck, & Beullens, 2018) which causes patients to relate, feel less isolated (Sinha, Porter, & Wilson, 2018), and reduce uncertainty (Donovan, LeFebvre, Tardif, Brown, & Love, 2014).
Materials and Methods
Kanker.nl
The current study gathered data from cancer patients and close relatives that are registered on Kanker.nl and who agreed to be part of a research panel. This online platform is accessible to a wide audience (e.g., current patients, ex-patients, close relatives) and consists of three different parts that each provide a variety of cancer-related information and support (Kanker.nl, n.d.). In the expert-generated ‘Library/ information’ part various medical information can be sought ranging from disease type to treatment and consequences of cancer. Experiences and stories can be read in the peer-generated part ‘Experiences from
others’. This part includes group conversations covering several topics. Users can ask each
other questions, share experiences and respond to others. The third part ‘Help and support’ provides possibilities to ask questions and seek help from professionals. The Kanker.nl panel includes participants that were registered users. Before entering Kanker.nl, the website requires its users to register by filling in their username, valid e-mail address, and their relationship with the disease cancer. Those who give (standard) consent to be approached for research were send an automated e-mail with the online survey invitation.
Participants and Procedure
Before actual data could be retrieved, the current study needed to be ethically
approved by the Ethics Committee of the University of Amsterdam (reference number: 2019-PC-11494). After this, a heterogeneous group of current patients, ex-patients (i.e., survivors) and close relatives of cancer patients received an online survey invitation. Next to examining (ex-)patients’ characteristics and motives, it was chosen to include close relatives into the study as Kanker.nl is also often visited by (close) relatives (Sanders, et al., 2020).
seekers (Sadasivam, et al., 2013). These surrogate seekers are often caregivers, relatives, or someone that is close to the person with a serious medical condition whose online health information seeking seems to show different patters than that of (ex-)patients.
To be eligible for the study participants had to be a registered user of Kanker.nl, at least 18 years of age, have sufficient command of the Dutch language, and be either a current patient, ex-patient or close relative of the first degree (i.e., parents, child, brother, sister, partner). Before the participants could take part in the survey their informed consent was asked. After this informed consent, participants were asked to indicate their relation to the disease cancer as a method to filter out those who were not eligible and to determine the course of the survey questions. Depending on whether the participant was a current patient, ex-patient of close relative, different types or formulations of the questions in the survey arose. Data from Kanker.nl cover the period from December 18, 2019 until January 1st, 2020.
Motives for Cancer Patients’ OHISB
Participants’ motives for seeking certain information and support online were assessed with a variety of hypothetical situations, e.g., Would you visit kanker.nl when you received contradictory information from different health practitioners?”, whereby they could indicate whether they would visit Kanker.nl for this situation. Before statistical analysis, all motives were subordinated into overarching categories. This was done based on prior research and led to three categories, e.g., complement information given by health
practitioner or during consultation (hereafter, “compliment information”; KR-20 a = .73), validate and/or challenge information given by health practitioner (hereafter,
“validate/challenge information”; KR-20 a = .76), and contact with peers (KR-20 a = .84) (see Table 1 in Appendix B for categorization of motives).
Measures of Predicting Factors
Medical background and sociodemographic characteristics
To gain insight into the medical background, respondents were asked to report when they or their relative was diagnosed with cancer, which type of cancer was diagnosed, and when their last treatment ended or whether they were still received treatment. When patients were still in treatment the main goal of this treatment was asked. At the end of the survey, sociodemographic data, such as age, gender, education, and living status (e.g., alone, with partner, with partner and child(ren), with child(ren), with other family members, other) were collected.
Cancer-related psychological distress
As psychological distress can manifest itself in different ways it was measured with the use of three concepts. To examine how respondents react during dramatic events, such as the diagnosis of cancer, the 7-item Impact of Events Scale (IES) (Horowitz, Wilner, & Alvarez, 1979), e.g., “I thought about the disease when I didn’t mean to”, was used with a 4-point scale (e.g., 0 = “not at all”, 1 = “rarely”, 2 = “sometimes” to 4 = “often”, Cronbach’s a = .88). Next, worries about the disease were measured with an adapted 6-item Cancer Worry Scale (CWS) (Custers, et al., 2014), e.g., “During the last month how often have you thought about your chances of getting cancer (again) or further expansion of the cancer?”, whereby a 4-point scale was used (e.g., 0 = “not at all”, 1 = “rarely”, 2 = “sometimes” to 4 = “often”, Cronbach’s a = .90). The higher the score, the higher respondents’ worry about the disease. With the use of a Dutch shortened version of the Intolerance of Uncertainty Scale (IUS) it was assessed how people perceive uncertainties, e.g., “Unforeseen events seriously upset me”, using a 5-point scale (e.g., 1 = “strongly disagree to 5 = “strongly agree”, Cronbach’s a = .88).
Coping
To measure participants’ way of coping and thereby style of OHISB under treat, an adapted version of the shortened Threatening Medical Situation Inventory (TMSI) was used (Miller, 1995). Hereby participants could indicate to which extent the following three items were applicable to them after the diagnosis of cancer (of a close relative). Items encompassed statements, e.g., “I intended to read about my/the disease”, with each a 5-point scale (e.g., 0 = “Not applicable at all”, 1 = “not very applicable”, 2 = “somewhat applicable”, 3 = “quite applicable”, 4 = “very applicable”, Cronbach’s a = .87). In addition, participants’ information satisfaction was assessed by using an adapted version of the shortened Information
Satisfaction Questionnaire (ISQ) (Thomas, Kaminski, Stanton, & Williams, 2004), e.g., “I would like to know as much as possible about my disease and/or treatment, both positive and negative information”. Hereby participants could indicate from four items which description was most applicable to them.
eHealth literacy
A Dutch shortened 10-item version of the Digital Health Literacy Instrument (Van der Vaart & Drossaert, 2017) was used to measure several subskills from participants. First, information seeking skills were assessed, e.g., “When searching for information about the disease cancer on kanker.nl, how easy or difficult do you find it to choose from the
information you find?”, which was followed by the measurement of participants’ information assessing skills, e.g., “When searching for information about the disease cancer on kanker.nl, how easy or difficult do you find it to determine whether the information is reliable?”.
Thirdly, information application skills were measured, e.g., “When searching for information about the disease cancer on kanker.nl, how easy or difficult do you find it to determine whether the information found applies to you?”, after which participants’ online navigation
skills were assessed, e.g., “When searching for information about the disease cancer on kanker.nl, how often does it happen that you lose your way on the website or on one of the pages?”. With the adoption of this instrument each item was measured on a 4-point scale (e.g., 1 = “very easy”, 2 = “quite easy”, 3 = “quite difficult”, 4 = “very difficult”, Cronbach’s a = .85), this in such a way that a higher score indicates a higher eHealth literate respondent.
Tendency to engage in thinking
To weigh the extent to which the participants tend to enjoy and engage in thinking an 8-item need for cognition scale was used (Pieters, Verplanken, & Modde, 1987). Items
encompassed several statements, such as “I like to have the responsibility of handling a situation that requires a lot of thinking”, for which participants could indicate how much they agreed on a 7-point Likert scale (e.g., 1 = “strongly disagree” to 7 = “strongly agree”,
Cronbach’s a = .80). Higher scores of a respondent indicate a bigger tendency to engage in thinking in about the cancer-related information.
General trust in doctors
With regard to participants’ trust in health practitioners, a shortened and adapted version of the Wake Forest Physician scale was used (Dugan, Trachtenberg, & Hall, 2005), e.g., “Sometimes health practitioners place their own interest above the medical interest of a patient”. For each of the five items participants could indicate the extent to which it was applicable to them on a 5-point Likert scale (e.g., 1 = “strongly disagree”, 5 = “strongly agree”, Cronbach’s a = .85). In this case, higher scores indicate more trust in health practitioners in general.
Role preferences in medical decision-making
Based on five different issues participants could indicate which role they preferred in medical decision-making. The Control Preferences scale serviced to assess whether they prefer an active role (e.g.,” I would prefer to determine for myself which treatment I
receive”), a collaborative role with the health practitioner (e.g., “I would prefer to determine, together with my health practitioner, which treatment is best for me”), or a passive role (e.g., “I would prefer my practitioner to decide which treatment is best for me” ) by asking the participants to choose between the five options presented.
Statistical Analyses
To analyse the data gathered by the cross-sectional online survey, cluster analysis was adopted as a statistical method to identify the distinctive profiles. The method is used to group cases into clusters in such a manner that all individuals within one cluster are similar to each other on the basis of specific variables, but significantly distinctive from the individuals in other clusters (Borgen & Barnet, 1987; Field, 2017; Norufis, 2010). Respondents’ motives for their OHISB were measured and then standardized to z-scores to compare variables on an equal level. Next, an explorative agglomerative hierarchical cluster analysis was conducted. Hereby, Ward’s method with a squared Euclidean distance was used to join the respondents into clusters such that the variance was minimized within each cluster (Field, 2017). Ward’s method is a clustering algorithm that has been widely used in behavioral science due to its reliability (Borgen & Barnet, 1987; Ward, 1963). The squared Euclidean distance, as the geometric measure of proximity, is most often applied with this method since it is able to not only reflect the profile shape and scatter of the cluster solution, but also the level (Borgen & Barnet, 1987).
Several methods were used to determine the suitable number of clusters. First, an inversed scree plot was drawn to detect the ‘elbow’ in the graph. However, the five-cluster solution this did not provide a definitive answer. Additionally, three random subsamples were used to search for a stability in cluster solutions. Based on the three inverted scree plots a three-cluster solution and one four-cluster solution were found. Thirdly, cluster variable graphs with z-scores and a dendrogram (see Appendix C) were drawn for both cluster solutions to visually compare them (see Figure 2 and 3). Although, three clusters would visually be the solution here, previous research on cancer patients’ general information-seeking behavior (Protière, Moumjid, Bouhnik, Le Corroller Soriano, & Moatti, 2012) has shown four cluster solutions. Therefore, fourthly, both a three-cluster and four-cluster
solution were inspected based on their cluster interpretability, shape, and level by conducting a one-way analysis of variance (ANOVA) with Turkey’s Honestly Significant Difference (HSD) post hoc comparisons on the predictors of OHISB (e.g., age, intrusive thinking, eHealth literacy). A Turkey HSD post hoc was used to report between which clusters precisely the significant difference occurred. This cross-sectional comparison favored the four-cluster solution above the three-cluster solution for which the four-cluster solution was chosen to continue with.
In addition to the four-cluster ANOVA with Turkey’s HSD post hoc comparisons, Chi-square tests were conducted to make cross-sectional comparisons between cluster membership and gender, education level, living situation, type of cancer, information satisfaction, and shared decision-making. The Chi-square was used to determine whether associations or dependence existed between the membership and variables (Field, Chi-Square Test, n.d.; McHugh, 2013). Hereby three assumptions were taken into account: (1) all the data in the cells should be either frequencies, counts or cases and not percentages or some other transformation of data, (2) All study groups must be independent from each other, and
(3) the value of the cell ‘expected’ should be five or more in at least 80% of the cells, and none of these cells should have an expected count of less than one. In other words, not more than 20% of cells with expected count should have less than five (McHugh, 2013). IBM’s SPSS software version 26 (IBM Corporation, Released 2019) was used to conduct the all analyses. During data analysis significance was set at the p-level of .05.
Results
Sample Characteristics
A total of 199 participants responded to the invitation of which ten respondents were excluded from data analysis since they did not meet the selection criteria, yielding a total of 188 responses available to be analyzed. Ultimately, data of a total number of 178 (ex-) patients and ten close relatives was used for data analysis. Respondents ages ranged from 20 to 88 years (M = 60.67, SD = 9.12) and most of them reported a high level of education (i.e., higher education and academic education; 56.5%). Slightly more women than men were represented in the sample (54.0%). Furthermore, most patients were diagnosed with breast cancer (18.6%) or urological cancer (21.8%). The sociodemographic, health-related and medical characteristics of the overall sample are presented in Table 2 (see Appendix E).
Visiting Kanker.nl and motives
Many people indicated to have visited the website Kanker.nl several times (see Figure 4 in Appendix E). This number of visitations ranged from zero to 60 times in the last month before the survey (M = 4.65, SD = 7.32). The website part ‘Experiences of others’ was the most often visited by the respondents (53.7%), which followed by ‘Library/ information’ (29.2%), and ‘Help and support’ (14.0%). With regard to the online platforms each cluster reported to seek for online health information most often on the ‘Experiences of others’ part of Kanker.nl. Most people reported to be seeking online health information on the website because they received contradictory information from different health providers (N = 159, 89.3%). Moreover, many people indicated to be provided with too little information by their health provider (N = 146, 82.5%). Many want to read the experiences and advice of fellow patients (N = 143, 82.2%). A description of each motivation and motive category is presented in Table 1 in Appendix B.
Figure 2. Cluster profiles for the four clusters identified within the four-cluster solution. -2 -1 0 1 2 Complement information Validate/challenge information Contact with peers z-scores Variables
Cluster 1 Cluster 2: Cluster 3
-2 -1 0 1 2 Complement information Validate/challenge information Contact with peers Z-scores Variables
Cluster 1 Cluster 2 Cluster 3 Cluster 4
Results of the cluster analysis
The current study found four clusters. Based on the stability, interpretability and comparability of the clusters with regard to their scores on the motives and predictors, they could be characterized as ‘average literate seekers, ‘distressed all-rounders’, ‘reasonable literate informers’, and ‘lowest literate supporters’. All four clusters differed significantly from each other with regard to the three different motive categories as determined by one-way ANOVA (see Figure 3). The representation of ‘Complement information’, F(3,159) = 133.15, p = .00, ‘Validate/challenge information’, F(3, 159) = 111.06, p < .01, and ‘Contact
with peers’, F(3, 159) = 81.78, p < .01, could be investigated for each cluster and were scored
either high, moderate or low. Furthermore, overall significant differences between the clusters were found with regard to intrusive thinking, F(3, 159) = 4.77, p < .01, intolerance for uncertainty, F(3, 159) = 2.72, p < .05, and eHealth literacy, F(3, 152) = 4.40, p < .01. All the corresponding results for the four-cluster can be found in Table 3 (Appendix F).
Cluster 1: Average literate seekers (n = 56, 34.4%)
People in this cluster scored significantly higher on the motive ‘complement
information’, ‘validate/challenge information’ and ‘contact with peers’ in comparison to the
other clusters (Figure 3). Additionally, this type of online health information and support seeker scored for both intrusive thinking (M = 2.56, SD = 0.72) and intolerance for uncertainty (M = 2.48, SD = 0.67) lower than average. The Turkey HSD post hoc test
revealed that average literate seekers significantly differed from cluster 2 on the basis of both intrusive thinking (p < .01) and intolerance for uncertainty (p = .04). However, for their average score on cancer worry, coping and eHealth literacy no significant difference between average literate seekers and the other clusters was found. Also, no significance was found for the trust in doctors and need for cognition.
Thus, patients in this cluster are characterized as moderately motivated to complement and validate information, moderately motivated to contact peers online, do not have many disturbing thoughts about their disease and are able to cope with uncertainty.
Cluster 2: Distressed all-rounders (n = 39, 28.7%)
People in this cluster scored the highest of the four clusters on all the motives ‘complement information’, ‘validate/challenge information’, and ‘contact with peers’. For this type of information and support seeker a higher score than average was observed for intrusive thinking (M= 3.04, SD = 0.63). With an average score on intrusive thinking, the Turkey HSD post hoc test revealed that this cluster significantly differed from cluster 4 (p = .03). Also, with regard to intolerance for uncertainty distressed all-rounders significantly differed from the average literate seekers (p = .04). Moreover, for the average score on eHealth literacy a significant difference was observed between distressed all-rounders and cluster 4 (p = .01). Next to a higher score on intrusive thinking and intolerance for
uncertainty, respondents also reported the highest score on cancer worry. In line with these previous results, the cluster scored the highest on coping, indicating that the distressed all-rounders try to take control over their cancer journey and care. Yet, they did significantly differ from the other clusters based on their need for cognition and trust in doctors.
Thus, patients from this cluster can be identified by their high motivation to complement and validate information as well as their high motivation to seek contact with peers online. Also, they possess average eHealth literacy skills. However, they have many disturbing thoughts about their disease and feel less to be able to cope with uncertainty.
Cluster 3: Reasonable literate informers (n = 45, 28.6 %)
People in this profile scored moderate on the motives ‘complement information’ and
‘validate/challenge information’ in contrast to the other clusters. Additionally, compared to
average literate restrainers and distressed all-rounders, this profile was composed of
respondents who scored the lowest on the motive ‘contact with peers’. Based on the slightly higher than average score for intrusive thinking (M = 2.82, SD = 0.61) and slightly lower than average score on intolerance for uncertainty (M = 2.53, SD = 0.79), no significant differences could be observed between reasonable literate informers and the patients in other clusters. The only significant difference was observed for eHealth literacy between lower literate informers and cluster 4 (p < .01). Surprising was, however, the score for trust in doctors among the people which was the lowest of all the clusters. In relation to cancer worry, need for cognition, and coping no significant differences were found.
Thus, patients in this cluster can be described as moderately motivated to compliment and validate information, low motivated to seek online contact with peers and a little less than average eHealth literate.
Cluster 4: Lowest literate supporters (n = 23, 14.1%)
The fourth cluster was characterized by the lowest scores on both the motive category ‘complement information’ and ‘validate/challenge information’. Similar to average literate seekers people within this profile reported a moderate score on the motive ‘contact with
peers’. Furthermore, this cluster scored significantly different from distressed all-rounders
with a low intrusive thinking (p = .03) and the lowest eHealth literacy (p = .01). Additionally, with the lowest eHealth literacy of all clusters, the lowest literate supports also significantly differed from the reasonable literate informers (p < .01). No significance could be found for
intolerance for uncertainty, cancer worry, coping and need for cognition. Lastly, patients reported to have a lower than average trust in doctors.
Thus, patients belonging to this cluster can be characterized by their low motivation to complement and validate information and moderate motivation to contact peers. They have a low eHealth literacy and experience disturbing thoughts about their disease.
Cross-sectional comparison of the clusters
In Table 4 are the results from the cross-sectional comparison of the clusters
displayed. The four clusters did not differ with respect to gender, educational level, type of cancer, information satisfaction, and shared decision-making. This means that there were no associations or dependences found between cluster membership and these variables.
Nevertheless, as is shown in Table 4 some results can be reported on the basis of information satisfaction and shared decision-making. Therefore, regarding information satisfaction most people reported that they want to have as much information as possible about the disease and/or treatment. Although, this finding holds no significance, it does show that the respondents of the overall study sample prefer both positive and negative information. Furthermore, the majority of respondents in all the clusters reported to prefer to engage in shared decision-making with the health provider and wanted to decide together which treatment would be best for them. Even though no significant association was found, this level of shared decision-making shows that people want to have a collaborative role and a say in their care.
Discussion
Main Conclusions
The results of the study suggest that different profiles exist with regard to cancer patients’ motives to seek online health information and support. However, with regard to their predictors only patients’ intrusive thinking, intolerance for uncertainty and eHealth literacy were able to distinguish the four clusters. While on the one hand distressed all-rounders and average literate seekers appear to be highly and moderately motivated to compliment and validate information and contact peers online, reasonable literate informers and lowest literate supporters were motivated to a lesser extent. Reasonable literate informers were moderately motivated to complement and validate information, though they had a low motivation to contact peers online. The opposite applied to lowest literate supporters who were moderately motivated to contact peers online and low motivated to compliment and validate information. This may be in line with previous research as the clusters differed based on their levels of psychological distress (i.e., intrusive thinking) and to the extent to which they were able to cope with uncertainty. It is possible that patients with higher levels of distress may seek more information and support online since as previous research has indicated that feelings of uncertainty often elicit sub-sequent health information-seeking behavior (Dickerson, Reinhart, Boemhke, & Akhu-Zaheya, 2011). Especially when patients are sufficiently eHealth literate, prior research suggests that they make greater use of health information (e.g., treatment specific information) (Li, Orrange, Kravitz, & Bell, 2014). Contrastingly, those who mainly seek support online seem to have the lowest eHealth literacy which may indicate that they find it more difficult to deal and understand with the online health information. This may however be hazard since this information is often not validate and of ambiguous nature (Anderson, Rainey, & Eysenbach, 2003).
Practical Implications
The results of this study have shown that significant differences between the clusters can be observed with regard to their motives to seek online for health information and
support. It has already been suggested that health providers should be aware that patients may be receiving information outside of their healthcare settings (Anderson, Rainey, &
Eysenbach, 2003; Anker, Reinhart, & Feeley, 2011). Particularly important after this study is, however, the additional awareness of health providers that cancer patients differ in their motives to seek online health information and support which may originate from the differs levels of need for competence, autonomy, and relatedness (Lee & Lin, 2016). Finally, the current study has shown which four profiles exist among the patients which can give health providers the opportunity to tailor the referral of their patients to information, support or online platforms relevant to them. This anticipation to patients’ information and support needs, might contribute to the enhancement of information provision since different online platforms were found to be represented by the same motives to seek online health information and support (Sanders, et al., 2020). Finally, tailored referrals may to lead an increase and improvement of patients’ participatory role when it comes to their care (Blödt, et al., 2018), and ultimately to the overall improved well-being of patients (Husson, et al., 2013).
Study Limitations and Recommendations
While the cluster analysis proved to be the best manner to identify different profiles with regard to patients’ motives to seek online health information and support, this study did not succeed to distinguish these profiles based on a wide variety of characteristics. This may be due to some limitations of the current study. First in general, Ward’s method was used to determine the number of clusters. However, often it is not clear which cluster analysis to use (Borgen & Barnet, 1987). As a result, people go for the best known and most used method.
This may not be the best strategy since each type of cluster analysis can result in different outcomes (Borgen & Barnet, 1987). Based on this, the choice of the current study for a hierarchical clustering with Ward’s method seems to be arbitrary. Especially since only few significant differences with regard to the predictors of the motives have been found. In addition, the current study adopted multiple methods to determine the most suitable number of clusters in order to reduce this problem of determining the final number of clusters (Everitt, 1979). Even though some rules of thumb exist, these do not always work well (Borgen & Barnet, 1987) as a cluster analysis is an exploratory process that depends on the subjective interpretation of the researcher and their skills (Everitt, 1979; Field, Cluster Analysis, 2017; Norman, Velicer, Fava, & Prochaska, 1998).
Second, potential limitations exist within in the retrospective nature of the online survey used. The self-reported data could be prone to recall bias since memory could have been distorted by a traumatic event, such as the diagnosis of cancer. Moreover, the time period during which the survey was set out could have influenced whether people fully completed the survey or not. A small amount of data was lacking from the study sample as not every respondent completed the whole survey. During cluster analysis no adjustments were made to account for this missing data which may have affected the formation of the clusters. Also, both the data of (ex-)patients and close relatives were used during data analysis that as a consequence could have biased the cluster analysis and its results. An interesting approach to determine the robustness of the data would be to conduct a sensitively analysis whereby only (ex-)patients are selected as cases. Although, the number of close relatives in the current study was relatively small. Prior research has indicated that
information needs differ between patients and non-patients (Jo, Park, & Jung, 2019; Lavallée, Grogan, & Austin, 2019). Another potential shortcoming may have been the design of the survey. This because participants were only able to indicate whether they would visit
Kanker.nl or not for the proposed motivation. However, reasons why they would not visit the website were not taken into account. As one respondent commented, this closed question was not fully representative of patients’ motives to seek online health information since their reasons for (not) visiting can be concerned with more complexity. In this case it would not be recommendable to solely rely on a quantitative survey, but to make use of triangulation by combining quantitative methods with more in-depth qualitative research methods such as interviews (Anderson, 2010; Fiksdal, et al., 2014).
Thirdly, the skewed sample of the current study might give a biased result. As participation was on a voluntary basis, there could have been a selection bias. People who took part of this study were probably already interested in research, as they were part of the online panel, or the (research) topic and found it necessary to participate in the survey to contribute to the improvement of online information provision. The voluntary participation could have led to bias as the sample was overrepresented by people who were mainly 40 years and over and highly educated. A reason for this may be that even though Kanker.nl is visited by a wide range of information and support seekers, the panel may be represented by an older higher educated group. A more general survey about cancer patients’ OHISB could deliver more representative results with a normal distribution. Participants could be recruited through medical settings, such as general practices. For example, previous research has shown that younger cancer patients have other online and offline information and support needs (Donovan, LeFebvre, Tardif, Brown, & Love, 2014; Goldfarb & Casillas, 2014; Lea, et al., 2018).
Finally, some recommendations for future study remain. First, it is suggested to include the health providers into future study since they need to be well-prepared and competent to deal with the ‘e-patient’ (Herrmann-Werner, et al., 2019). Internet-informed patients may expect them to react quickly and comprehensive during the discussion of online
information. However, not all health providers are able to talk about online health
information (Couët, et al., 2015; Sanders & Linn, 2018) and to support patients’ autonomy (Graffigna, Barello, Bonanomi, & Riva, 2017). Nevertheless, talking about online health information has been demonstrated to improve patient outcomes (Sanders & Linn, 2018) and the patient-physician relationship (Tan & Goonawardene, 2017). Thus perhaps, future
research can contribute to gain more insight into the OHISB of patients by combining both perspectives in the model suggested in Figure 1. Secondly, as the current study did not find any other significances besides eHealth literacy, intrusive thinking and intolerance for
uncertainty, it is possible that other variables and characteristics could be relevant to take into account. For example, prior research has shown that patients, who rate their health
practitioners lower on patient centeredness and experience an increased worry after medical consultation, are more likely to engage in post-visit OHISB (Li, Orrange, Kravitz, & Bell, 2014). Especially, since patient-physician relationships have been proven to be a significant predictor of the level of unmet information needs (Herrmann-Werner, et al., 2019;
Jirasevijinda, 2015; Neumann, et al., 2011). Ultimately, other factors may also be relevant to consider since various studies argue to have found an influence on patients’ OHISB based on patients’, frailty (Ekdahl, Andersson, & Friedrichsen, 2010), future time perspective (i.e., perceived time left in life) (Löckenhoff & Carstensen, 2004), cancer stage (Neumann, et al., 2011), perceived loneliness (Fogel, Albert, Schnabel, Ditkoff, & Neugut, 2002), sense of control, need to resume to normality (Germeni & Schulz, 2014), socio-economic status (SES), migration background (Nölke, Mensing, Krämer, & Hornberg, 2015), and trust in the online information source (Yang, Chen, & Wendorf Muhamad, 2017). Since many prior studies have limit themselves to the examination of only few variables, it would be interesting for future research to combine the many variables to understand not only their potential influence on OHISB, but also the possible interplays between them.
Conclusions
The current study demonstrated that heterogeneity exists among cancer patients’ motives to seek online for health information and support. Based on these motives four distinctive patient profiles could be identified. Health practitioners could use the profiles as tailoring method to refer their patient to online health information or platforms relevant to them. However, no other predictors besides the level of eHealth literacy, intrusive thinking and intolerance for uncertainty could be found. Future qualitative and quantitative research is needed to gain more understanding of cancer patients’ motives for their OHISB.
References
Anderson, C. (2010). Presenting and evaluating qualitative research. American Journal of
Pharmaceutical Education, 74(8), 141. https://doi:10.5688/aj7408141.
Anderson, J., Rainey, M., & Eysenbach, G. (2003). The impact of cyberhealthcare on the physician-patient relationship. Journal of Medical Systems , 21(1), 67-84.
https://doi.org/10.1023/A:1021061229743.
Andreassen, H. K., Bujnowska-Fedak, M. M., Chronaki, C. E., Dumitru, R. C., Pudule, I., Santana, S., . . . Wynn, R. (2007). European citizens' use of e-health services: a study of seven countries. BMC Public Health, 7(1), 53. https://doi.org/10.1186/1471-2458-7-53.
Anker, A., Reinhart, A., & Feeley, T. (2011). Health information seeking: A review of measures and methods. Patient Education and Counseling, 82(3), 346-354. https://doi.org/10.1016/j.pec.2010.12.008.
Attfield, S. J., Adams, A., & Blandford, A. (2006). Patient information needs: pre- and post-consultation. Health Informatics Journal, 12(1), 165-177.
https://doi.org/10.1177/1460458206063811.
Ball-Rokeach, S. (1998). A theory of media power and a theory of media use: Different stories, questions, and ways of thinking. Mass Communication and Society, 1(1-2), 5-40. https://doi.org/10.1080/15205436.1998.9676398.
Beauchamp, T. L. (1994). Principles of Biomedical Ethics. Oxford, NY, USA: Oxford University Press.
Beckjord, E. B., Arora, N. K., McLaughlin, W., Oakley-Girvan, I., Hamilton, A. S., & Hesse, B. W. (2008). Health related information needs in a large and diverse sample of adult cancer survivors: Implications for cancer care. Journal of Cancer Survivors, 2(3), 179-189. https://doi.org/10.1007/s11764-008-0055-0.
Bell, R. A., Hu, X., Orrange, S. E., & Kravitz, R. L. (2011). Lingering questions and doubts: online information-seeking of support forum members following their medical visits.
Patient Education Counseling, 85(1), 525-528.
https://doi.org/10.1016/j.pec.2011.01.015.
Bensing, J., & Verhaak, P. M. (2004). Communication in medical encounters. In J. Weinman, & A. Kaptein (Eds.), Health psychology (pp. 261-287). Malden: Blackwell
Publishing. https://doi: 10.1108/09654280810899993.
Berzonsky, M. D., & Sullivan, C. (1992). Social-cognitive aspects of identity style: Need for cognition, experiential openness, and introspection. Journal of Adolescent Research,
7, 140-155. https://doi.org/10.1177/074355489272002.
Biermann, J., Golladay, G., Greenfield, M., & Baker, L. (1999). Evaluation of cancer information on the internet. Cancer, 86(3), 381-390.
Blödt, S., Kaiser, M., Adam, Y., Adami, S., Schultze, M., Müller-Nordhorn, J., & Holmberg, C. (2018). Understanding the role of health information in patients' experiences: secondary analysis of qualitative narrative interviews with people diagnosed with cancer in Germany. BMJ Open, 8(3), e019576. https://doi.org/10.1136/bmjopen-2017-019576.
Blanch-Hartigan, D., Blake, K., & Viswanath, K. (2014). Cancer survivors' use of numerous information sources for cancer-related information: Does more matter? . Journal Of
Cancer Education, 29(3), 488-496. https://doi.org/10.1007/s13187-014-0642-x .
Bol, N., Linn, A. J., Smets, E. A., Verdam, M. E., & Van Weert, J. M. (2020). Manuscript - Tailored communication for older patients with cancer: Using cluster analysis to identify patient profiles based in information needs. Journal of Geriatric Oncology. Bol, N., Smets, E., Burgers, J., Samii, S., De Haes, H., & Van Weert, J. (2018). Older
patients' recall of online cancer information: Do ability and motivation matter more than chronological age? Journal of Health Communication, 23(1), 9-19.
https://doi.org/10.1080/10810730.2017.1394400.
Borgen, F. H., & Barnet, D. C. (1987). Applying cluster analysis in counseling psychology research. Journal of Counseling Psychology, 34(4), 456-468.
https://doi.org/10.1037/0022-0167.34.4.456.
Brier, M. A., Kazak, A. E., Werba DeRosa, B., Hocking, M. C., Schwartz, L. A., Ginsberg, J. P., . . . Ittenbach, R. F. (2011). Profiles of health competence beliefs among young adult survivors of childhood cancer. Journal of Adolescent and Young adult oncology,
1(4), 195-202. https://doi:10.1089/jayao.2012.0006.
Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and
Sociol Psychology, 41(1), 116-131. https://doi.org/10.1037/0022-3514.42.1.116.
Caiata-Zufferey, M., & Schulz, P. J. (2012). Physicians' communicative strategies in interacting with internet-informed patients: Results from a qualitative study. Health
Communication, 27(8), 738-749. https://doi:10.1080/10410236.2011.636478.
Caiata-Zufferey, M., Abraham, A., Sommerhalder, K., & Schulz, P. J. (2010). Online health information seeking in the context of the medical consultation in Switzerland.
Qualitative Health Research, 20(8), 1050-1061.
https://doi.org/10.1177/1049732310368404.
Carlsson, M. E. (2009). Cancer patients seeking information from sources outside the health care system: change over a decade. European Journal of ONcology Nursing, 13(4), 304-305. https://doi: 10.1016/j.ejon.2009.03.005.
Case, D. O. (2007). Looking for Information: A Survey of Research on Information Seeking,
Needs, and Behavior. London, UK: Elsevier/Academic Press.
Centraal Bureau voor de Statistiek. (2019, January 24). Vaker online op zoek naar informatie
over gezondheid. Retrieved January 2020, from CBS:
