How do patient perceptions about their treatment and
relationship with their doctor relate to patient medication
adherence?
Thesis Marketing
31 August 2015 Studiejaar 2014/2015 Semester I, Blok III
Rik van der Sloot 10025979
Statement of originality
This document is written by student Rik van der Sloot, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents
1. Introduction………...P5 1.1 Patient medication adherence………...P5 1.2 Research question and sub-questions………..P8 1.3 Contributions of this study………..P9 1.4 Limitation of this study………..P10 1.5 Structure of this study………P12 2. Literature review………...P13 2.1 What is patient medication non-adherence?...P13 2.2 What factors constitute patient attributes? And what is their effect on patient
medication adherence?...P15 2.3 What factors constitute the patient perception about the physician-patient
relationship? And what is their effect on the patient medication
adherence?...P21 2.4 What factors constitute the patient beliefs about the treatment? And what is
their effect on the patient attributes?...P27 2.5 The mediating effect of patient role clarity and patient motivation………..P30
3. Research methods………..P34 3.1 Sample and data collection………P35
3.2 Measurement scales………...P36 4. Analysis and results………...P40 5. Discussion and conclusion……….P51 5.1 Interpretation of the results………P51 5.2 Unexpected results……….P54 5.3 Theoretical implications………P58 5.4 Practical implications………P59 5.5 Limitations and recommendations for future research………..P60 5.6 Conclusion……….P62 6. Apendix………...P63 6.1 Apendix A: Facebook and Email messages………...P63 6.2 Apendix B: Both the Dutch and English questionnaires………...P65 6.3 Apendix C: SPSS output for both factor analyses……….P73 6.4 Apendix D: SPSS output for testing for the assumptions………..P77 6.5 Apendix E: SPSS output process macro mediation analyses………P84 7. Bibliography……….P102
Abstract
Patient medication non-adherence remains a major issue in the global health industry. Therefore over the years many attempts have been made to explain why patients do or do not adhere to their medication. The research question of this study was: How do patients’ perception of the physician-patient relationship and the (medication) treatment influence patients' medication adherence? Patient participation, tailored communication and provider expertise were included as factors constituting the patient-physician
relationship and patients’ beliefs about the necessity of their medication, their concerns about their medication and the severity of their disease were included as factors
constituting patients’ perceptions of their medication treatment. The data of this study was collected through the use of a survey method. Eventually, 135 patients participated in this study by filling-out the questionnaire. In total six simple mediation analyses were conducted with the SPSS process macro from Hayes (2012), in order to test the proposed direct effects of patient participation, tailored communication, provider expertise,
patients’ necessity beliefs, patients’ concerns about their medication and disease severity on patient medication adherence. Furthermore, patient role clarity and patient
motivation were included in this study as mediators on these relationships. The results from these six regression analyses indicated that especially patients’ necessity beliefs and patients’ concerns about their medication were important predictors of patient
medication adherence. Lastly, it was found that patient motivation mediated the relationship between patients’ concerns about medication and patient medication adherence.
1.Introduction
1.1 Patient medication non-adherence
Patient medication non-adherence is a major issue in the healthcare industry. According to, Forissier & Firlik (2012), pharmaceutical industries lose billions every year due to patient medication non-adherence. According to their estimations pharmaceutical industries lose 188 billion annually in the US alone and 564 billion globally. Also, medication non-adherence is responsible for 290 billion in spending that could otherwise be used for different medical activities (Forissier & Firlik, 2012). Their report represents one of the most accurate estimations of pharmaceutical revenue losses due to medication non-adherence of patients and represents the fact that medication non-adherence is a major problem that needs to be solved.
The World Health Organization (WHO) defines adherence as: “the extent to which the patient follows medical instructions”. The emphasis here is on the extent to which a patients’ behavior corresponds with agreed recommendations from a healthcare provider, with regard to: taking medication, following a diet and/or executing lifestyle changes. Jin et al (2008) describe several types of non-adherence, including: taking incorrect doses of medication, taking medication at the wrong times, stopping the treatment too soon and failure to follow (other) doctor’s instructions. Of course, besides the previously mentioned negative financial consequences for healthcare providers, non-compliance has negative effects on patients as well. Since non-adherence to prescribed medication reduces the possibility of a patient achieving the desired healthcare outcomes (Jin et al, 2008). For example, Sokol et al (2005), state that symptoms of several diseases may worsen if a patient only partially adheres to the prescribed treatment. According to
costs for the healthcare provider, but also negatively affects the patients’ well being. Since the threat of complications will grow significantly if a patient receives suboptimal treatment due to medication non-adherence (Sokol et al, 2005)
There is a commonly held assumption that the cost of medication and
forgetfulness are the two major drivers behind medication non-adherence (Firlik, 2013). This assumption, however, seems to overlook the possible psychological and
motivational challenges among patients to pick up their medication and stick to their prescribed medication regime. Since, the use of reminders or even giving away free medication, have not proven to be successful solutions. One major issue here is the fact that most medication needs to be taken over long periods of time without offering short-term benefits to patients, for example increases in health (Firlik, 2013). This can create strong psychological barriers for patients and in order to solve the problem of medication non-adherence these barriers need to be overcome.
The article of Gottlieb (2000) also contradicts the assumption that forgetfulness and the cost of medication are the two major drivers of medication non-adherence. According to Gottlieb (2000) non-adherence rates are the highest among patients that are ‘symptom free’. Another interesting finding is the fact that adherence rates seem to be higher for patients taking medication in order to cure disease, compared to patients taking medication to prevent disease. And lastly, when medication needs to be taken for longer periods of time (chronic diseases), adherence rates drop even further. According to Roselund et al (2004), only 24 percent of patients who said they did not take their medication as described, mentioned forgetfulness as the reason. Instead, most patients actively chose to disregard the instruction from their physician. Which is alarming, since
this problem needs a more comprehensive solution than simply sending emails to remind patients to take their medication. Roselund et al (2004) state that how and why these patients make these decisions varies substantially. These variations are caused by three factors: gender, nature of the patient’s condition and the patient’s involvement in
healthcare decisions. The authors conclude with the notion that there are large benefits to reap for healthcare providers that succeed in tackling these aforementioned issues.
In order to further examine the underlying causes of patient non-adherence, Jin et al (2008) also explored and evaluated factors (or barriers) causing therapeutic non-compliance. They distinguish between factors centered on the patient (demographics, motivation and beliefs), the therapy (complexity, duration), social support (family and friends), financial situation (income) and factors about the disease (symptoms and
severity). The authors state that healthcare providers need to consider all these factors (or barriers), when designing a therapy plan, in order minimize patient non-compliance. They conclude with dividing the previously mentioned factors in ‘hard’ and soft’ factors. They find that these ‘hard’ factors are more easily quantifiable and can be directly linked to adherence. On the other hand, ‘soft’ factors are more difficult to measure and link to patient medication adherence, for example a patient’s beliefs about medication or the patient’s motivation. However, the authors conclude, these ‘soft’ factors, need to be addressed adequately as well, If healthcare providers want to improve the adherence rates.
Another study that examines variables that influence patient adherence is the study from Dellande et al (2004). They investigated the relationship between role clarity,
which customer compliance is essential for these services to be successful. More
specifically, the authors gathered data from service providers and clients in a weight-loss clinic and found that patients that understand what is expected of them, posses the ability to perform these behaviors and are motivated to do so, are more likely to comply. Also, the results show that increased client compliance resulted in greater client satisfaction. However, these authors do not have a strong focus on elements related to the interaction that takes place between the patient and the physician. Yet, this is what Hausman focuses on in her article. Hausman (2004) describes the importance of interpersonal elements in the patient-physician relationship. The author develops and tests a model that
demonstrates how interpersonal elements, together with communication and participation contribute to positive outcomes, such as compliance with medical advice. Participation and communication between patient and physician are proven to positively affect patient satisfaction and compliance, but Hausman (2004), does not describe the mechanism through which these elements stimulate compliance. It is this gap that this thesis is trying to fill. This is also in line with what Kendler et al (2014) state:, namely that: efforts to improve patient adherence to medication are most likely to be successful if they focus on the perceptual and practical factors that influence a patient’s ability and motivation to adhere.
1.2 Research Question
Based on preliminary research the following research question has been formulated to clearly discuss the problem above: How do patients’ perception of the physician-patient relationship and the (medication) treatment influence patients' medication adherence?
1.2.1 Sub questions
In order to answer this research question several sub questions have been designed to split the problem into more manageable parts. These sub questions are designed as followed:
• What is medication-adherence?
• What factors constitute the patient attributes? And what is their role in explaining the effect of patient perceptions about the physician-patient relationship and patient (medication) treatment perceptions on patient medication adherence? • What factors constitute the patient perception about the physician-patient
relationship? And what is their effect on the patient attributes?
• What factors constitute the patient perception about the (medication) treatment? And what is their effect on the patient attributes?
In the second chapter, more background information will be provided about the different concepts included in the sub questions.
1.3 Contributions.
1.3.1. Theoretical contributions
As mentioned above, non-adherence behavior is a major issue in the healthcare industry. Many attempts have been made to solve this issue, however the definitive answer has not been found yet. This study will try to improve the existing knowledge about patient adherence, by combining a set of variables in a way that has not been done
as mediators, like in the study of Dellande et al (2004). However, a new set of independent variables was chosen in order to attain new insights into what factors influence patient role clarity patient, motivation and ultimately patient adherence to medication. These variables are: patient participation, tailored communication, provider expertise, necessity beliefs and concerns and disease severity. Investigating the
relationships between these independent variables, the mediators role clarity and motivation and the dependent variable medication adherence, will provide new insights into why patients do/do not adhere to their medication. The goal of this study is to examine the mechanisms through which patient medication adherence is influenced by patient role clarity, patient motivation, patient participation, provider expertise, tailored communication, necessity beliefs and concerns and disease severity. This knowledge will improve the existing knowledge on patient medication adherence since these variables have not been studied in this combination before.
1.3.2. Managerial contributions
As mentioned above, the aim is to investigate variables that influence patient’s behavior. Patient participation, provider expertise and tailored communication together play a central role in the relationship between patients and their physician. By improving our understanding of the effect of these variables on patient medication adherence, doctors and other healthcare practitioners can design and improve the quality of their interactions with their patients in order to improve their medication adherence.
Furthermore, the beliefs patients have about the necessity of their medication; the concerns they might have about having to take medication and their beliefs about the
severity of their disease will influence patients’ in whether or not to take their medication. Insights into the relationship between these factors will help doctors and other healthcare providers when they address the issue of patient medication adherence.
1.4 Limitations of the study.
As with most studies this study had several limitations as well. This study includes the variables: patient participation, tailored communication, provider expertise, necessity beliefs and concerns and disease severity as independent variables and patient role clarity and motivation as mediators. However, there are a lot more variables that might influence patients’ adherence to medication. For example, the severity of possible side effects may have a strong influence on patients’ decision to whether or not take their medication. However, since addressing the issue of side effects would involve asking patients several sensitive questions, it was not possible to include the potential effects of side effects in this study. Another factor that might have an effect on patient medication adherence is patients’ financial situation. However, as with potential side effects, involving patient’s financial situation in this study, would increase the sensitivity of the questions and therefore the choice was made not to include patients’ financial situation in this study.
1.5 Structure of the thesis
In chapter two, the different variables and relationships that will be the focus of this research together with the conceptual model that will be tested, will be discussed. In chapter three the methodology of this study will be discussed, including information about the gathering of data and the measurement of the variables included in this study. In chapter four the gathered data will be analyzed for potential relationships, followed by the discussion section and conclusion in chapter five.
2. Literature review
2.1 What is patient medication non-adherence?
Over the years several definitions related to medication non-adherence have been developed. Therefore the first step of this research is to develop a clear understanding of what is meant exactly with medication non-adherence. Lehane & McCarthy (2008) describe compliance, concordance and adherence as three (partially) overlapping phenomena. According to their review, the first of these phenomena is compliance. Compliance used to be associated most commonly with patient medication taking. For instance, Hausman (2004) described compliance as the extent to which patient follow their doctor’s order. Specifically in terms of: taking medication, following diets, or executing other adjustments to the patients’ lifestyle.
A second phenomenon, that largely replaced compliance, is concordance (Lehane & McCarthy, 2008). Most recently it has been defined as “a process of prescribing and medication taking based on partnership”. Compared to compliance, concordance takes a more patient-centered approach to the delivery of health services (Lehane & McCarthy, 2008). Here specifically, knowledge, partnership and a supportive relationship between patient and physician are central to this approach.
The last phenomenon is adherence, which will be the subject of this research. As mentioned in the introduction, the World Health Organization (WHO) defines adherence as: “the extent to which the patient follows medical instructions”. With a strong emphasis on the patients’ behavior, taking medication, following a diet and/or executing lifestyle changes, corresponds with agreed recommendations from a healthcare provider.
and the patient (Lehane & McCarthy, 2008). The authors state that the concept of adherence not only facilitates the care given to patients and the assessment patient medication taking, it also allows for these assessments to be quantified, which allows healthcare providers to evaluate patient medication adherence. Lastly, the authors state that, the concept of adherence is believed to be flexible enough to incorporate the patients’ values and beliefs, regarding their medication, into account. Linn et al (2014), also qualify adherence as a relatively broad concept, where the focus shouldn’t be on a single behavior. Instead, the authors state that adherence should be viewed as a three-step process: the initiation phase, the implementation phase and the discontinuation phase. Where, in the first phase the patient takes the first dose of medication. In the second phase, the patient adjusts its behavior accordingly to the prescribed regimen. And lastly, in the third phase, the patient’s treatment is ended. Furthermore, Linn et al (2014), establish the importance of patient beliefs about their medication. Especially during the initiation phase, patients may develop certain beliefs about the necessity or concerns about their medication. This underlines the importance of patient involvement and communication between the healthcare provider and patients, since positive patient beliefs regarding their medication is associated with higher levels of adherence. Ahmed & Aslani (2014) also mention the importance of the need of patient involvement in their paper describing the concept of patient medication adherence. In a terminology overview the authors describe patient adherence as a multi-faceted construct. Adherence is believed to evolve around (1) the patients’ understanding of the severity of the illness, (2) their belief in the effectiveness of a certain treatment and (3) their ability to control their symptoms by utilizing this treatment. According to Ahmed & Aslani (2014) the term
adherence is currently favored over compliance by many experts in the field, because of its focus on the agreement between the patient and the physician in developing a course of action. Since the patient-physician relationship and the patients’ values and beliefs also play a central role in this research, the term adherence will be used from now on in this paper and is defined as the extent to which patients behavior corresponds with agreed upon medical instructions, with a specific focus on prescribed medication.
2.2 What factors constitute patient attributes? And what is their effect on patient medication adherence?
Dellande et al (2004) describe several variables affecting the likelihood of customer compliance (note that the authors use the term compliance in their paper,
therefore adherence is adapted here as compliance). Three of these variables are customer role clarity, customer ability and customer motivation. The authors empirically test a conceptualization of healthcare services in which customer compliance outside the service organization is necessary for successful health outcomes. They find that customer role clarity, ability and motivation serve as mediators, mediating the effect of several provider and customer characteristics on compliance, such as, provider expertise, homophily (degree to which certain people are similar on certain attributes, beliefs or values).
The study of Bowen (1986) stated an important difference between service firms and manufacturing firms. This difference represents the fact that with service firms, customers are often present as the service is offered, whereas with manufacturing firms,
challenge) for service firms to employ management practices that simultaneously keep customers satisfied with the service and productive during the creation of the service (Bowen, 1986). This has led to the suggestion that customers should be viewed as ‘partial employees’. Since customers often take an active role during the service delivery process, or even after the service delivery has occurred. For example with patients doing recovery exercises at home after a successful operation. A general model of the determinants of employee behavior resolves around three basic questions (Vroom, 1964):
• Do they understand what is expected from them (role clarity)? • Are they able to perform as expected (ability)?
• Are there valued rewards for performing as expected (motivation)? By viewing customers as ‘partial employees’ or even ‘co-producers’, Bowen (1986) suggests that these determinants apply to customer behavior in service settings as well. Moreover, Bowers, Martin & Luker (1990) have developed a three-step approach to apply these determinants to customer behavior. The authors explain that as a first step, the customer’s (patient) job should be defined. This means that it should be entirely clear to the patient, which behavior is expected. With the second step, the patient is
trained/schooled so the patient becomes able to perform this kind of behavior. Lastly, through positive reinforcement the patient stays motivated to show the desired behavior. With the formulation of these three steps the authors assumed that role clarity would lead to ability, which ultimately would lead to motivation. Dellande et al (2004) test this proposition empirically, in a health context, stating that as customers gain in role clarity, their ability to perform desired behavior increases, as will their motivation to do so. More specifically the authors gathered data from service providers and clients in a weight-loss
clinic and found that patients that understand what is expected of them, posses the ability to perform these behaviors and are motivated to do so, are more likely to comply. Also, the results show that increased client compliance resulted in greater client satisfaction with the weight-loss program.
This line of reasoning will be applied in this research as well. If patients know how to perform the desired behaviors, their ability to perform these behaviors increases, which ultimately will lead to patients being more motivated to perform the behaviors desired by the healthcare provider.
Together with the effect these patient attributes have on each other, Dellande et al (2004) also find direct effects of role clarity, ability and motivation on customer
compliance. This implies that the greater a customers role clarity, ability and motivation, the more likely this customer is to comply with the desired behavior. This effect has also been tested in a market setting by Meuter et al (2005). They investigated the adoption of self-service technologies (SST’s) by consumers across different industries. The authors explored key factors influencing the initial SST trial decision by customers, specifically focusing on actual behaviors in situations where consumers have a choice between delivery modes. They find that role clarity, ability and motivation, which they qualify as ‘consumer readiness variables’, act as key mediators between consumer’s innovation characteristics and individual differences and the likelihood of a trial. They state that consumer participation in a trial can be constrained because consumers experience insufficient clarity in terms of their understanding of their role in the service process. Furthermore, they state that the possibility of a trial is unlikely when consumers do not
to perform a task. With the notion that ability refers to what a consumer can do and not to what a consumers ‘knows to do’ or ‘wants to do’. According to the authors, consumers are less likely to engage in an SST trial when they believe they are incapable of
performing the necessary behaviors. Lastly, the authors describe motivation as ‘the willingness to perform certain behaviors’. Here the authors state that consumers can either be intrinsically (feelings of accomplishment) or extrinsically (rewards such as discounts) motivated to engage in an SST trial, but that the engagement in a trial is unlikely when these consumers do not have the motivation to perform these behaviors. Comparably, Moorman and Matulich (1993) describe health ability as consumers’ resources, skills or proficiencies to perform certain health behaviors. Furthermore, the authors state that knowledge, which can be linked to role clarity, is an important variable affecting consumer health ability. Secondly, Moorman and Matulich (1993) refer to health motivation as: consumers’ goal-directed arousal to engage in preventive health behaviors, also stating that health motivation focuses on consumers’ ‘willingness’ to perform certain behaviors. As with the previously described articles, the authors expected (and found) a positive relationship between consumers’ health ability, health motivation and their actual engagement in health preventive behaviors.
These confirmed direct effects of role clarity and motivation will be tested in this research as well. If patients posses the desired know-how (role clarity) and/or are
motivated to use this know-how, the likelihood of adherence will increase. The variable ‘ability’ will not be included in this research’s conceptual model. According to Dellande et al (2004) patient ability focuses on whether or not patients posses the necessary skills to perform the instructions they are given by their doctor or healthcare provider. Since the
main focus of this study is on whether or not patients take their prescribed medication and not whether or not they can make adjustments to their lifestyle, such as exercising, quit smoking or adjusting their diet, including patient ability would seem meaningless. On the other hand, as proposed by the text above, patient motivation and patient role clarity are more suitable in a setting where patient medication adherence is the central theme. Since patients need to understand exactly why, how, and when they need to take their medication, plus patients need to be sufficiently motivated to do so accordingly. Therefore the first hypothesis of this research is:
H1: the patient attributes (a) patient role clarity and (b) patient motivation, will have a positive effect on patient medication adherence.
The importance of social marketing
Social marketing began as a discipline in 1970 when Kotler and Zaltman argued that commercial marketing strategies, that have been proven successful in selling products, could also be used to promote socially beneficial ideas, attitudes and behaviors (Evans & McCormack, 2008). In the health industry context, social marketing tries to
promote/stimulate healthy behaviors, by using proven marketing techniques, used to promote commercial products. So for example, social marketing could be used to realize changes in a population’s smoking, drug use or sexual behavior (Evans & McCormack, 2008). Furthermore, Morris & Clarkson (2009) also state that social marketing tries to learn from the techniques of commercial marketing. The authors argue that the goal of social marketing is to ensure alterations in displayed behavior among a certain
important concept on the public health agenda in several different economies in the world. The authors describe the following six principles of social marketing:
• Behavioral goals: the aim of social marketing is to create measurable changes in consumers’ behavior. Setting behavioral objectives is a strategy for achieving this goal.
• Customer insight: The focus is here is to gain understanding in why people do what they do, why it benefits them, what influences them and might stop them. This insight is essential in order to understand what people need to alter their behavior into the desired state.
• Segmentation and targeting: Understanding that the market consists of different people, that will need different interventions is central to this principle.
• Competition: the focus here is on every possible force that might stand in the way. It refers to all the factors that might prohibit people from changing or altering their behavior.
• Exchange: the central element of any intervention is creating attractive and motivational exchanges with target audiences.
• Marketing and intervention mix: both consist of the four P’s: product, price, promotion and place. With the possibility of a fifth P: policy.
Morris & Clarkson (2009) conclude that their social marketing framework provides a useful framework for understanding the barriers to individual behavior change and designing interventions accordingly. This is in line with McKenzie-Mohr (2000) who emphasizes that a successful social marketing program starts with an understanding of the barriers for people to displaying certain behavior. Only after these barriers are identified,
can a social marketing program be implemented, in order to tackle these barriers and stimulate a population’s desired behavior. Therefore, in this research, several variables have been identified, that are supposed to affect patients’ adherence behavior. These variables have been grouped under either patients’ perceptions about the physician-patient relationship or physician-patients’ perceptions about their treatment and will be discussed in the next paragraphs.
2.3 What factors constitute the patient perception about the physician-patient relationship? And what is their effect on the patient medication adherence?
As mentioned in the previous paragraph, if physicians want to improve their patients’ adherence to medication, they will need to change their patients’ behavioral patterns. It is at this point, that social marketing comes into play. In the patient-physician relationship behavioral goals, customer insight and targeting and segmenting seem most important in stimulating patients to perform the desired behaviors according to Morris & Clarkson (2009). Therefore in this section, the focus will be on these factors contributing to the patient-physician relationship. In describing this relationship, the focus will be on providing patients with tailored communication, encouraging them to participate in deciding on a treatment process and the effect of provider expertise. These factors will be explained in this paragraph.
Patient participation
principles, the setting of goals can help to achieve this objective. The paper from Strecher et al (1995), discussed the benefits of goal setting in realizing behavioral change. The authors state that goals do not necessarily increase one’s motivation. The first step is to ensure the patient is interested in achieving the goal and make sure that achieving the goal does not conflict with the realization of other goals; for example, the goal to quit smoking might conflict with other personal goals such as maintaining weight or dealing with stressful situations. However, according to the authors, once a person is interested in achieving a goal, and this goal achievement is free of goal conflict, goal setting can generate higher performance than when no goals are set. Furthermore, Bassett and Petrie (1999) examined the effect of treatment goals on patient compliance. They investigated the effect of treatment goals, as a motivational tool, on patient compliance to exercise programs. The authors tested this effect among sixty-six patients starting a new physical exercise program. The authors made a distinction between mandated goals set by the therapist and goals that were set collaboratively between the patient and physician. Their analysis proved that participants in the collaborative group showed significantly higher levels of compliance than participants in the physician mandated group. This finding not only indicates the importance of goal setting in stimulating patient adherence, but also underlines the importance of patient participation in the process of goal setting. Also indicating the importance of participation. According to their results patients were more likely to adhere when they felt encouraged to participate in deciding on a treatment process. Hausman (2004) also investigated the importance of patient participation. According to the author, you can speak of patient participation, when patients are becoming actively involved in the decision-making process. This means that patients
need to be given the opportunity to solicit their own opinions and ideas about their treatment. According to Hausman’s (2004) study this involvement plays a critical role in the attempts to decrease patient noncompliance, since patients have become more and more medically sophisticated. Lastly, Loh et al (2006) examined the effect of
participation in a healthcare setting. Their objective was to test the effect of patient participation in shared treatment decision-making on patient adherence and clinical outcomes in the care of depression. According to the authors, patient participation is realized if, on the side of the physician, the patient was helped to understand all the information, the physician understood what was important to the patient and answered all of the patient’s questions. On the side of the patient, participation was realized, if the patient is sufficiently involved in decisions about the treatment and decides about further treatment together with the physician. The aim of the study was to test the impact of patient participation on treatment adherence for antidepressant pharmacotherapy and clinical outcome. The authors tested these effects by administering surveys to 30 general practitioners and 207 depressed patients. Their results indicated that patient participation had a significantly positive effect on patient medication adherence. These findings reveal the importance of patient participation as a key factor for improving treatment adherence. Therefore the second hypothesis of this research is:
H2: patient participation will have a positive effect on patient adherence to medication.
Tailored communication
The second and third principles from the social marketing theory are customer insight and segmenting and targeting. According to Morris and Clarkson (2009) these principles mainly emphasize that markets consist of different people and that in order for marketers to realize changes in the behavior of target audiences they need to understand that these different people will require a different approach. Here, the link can be made to the work of Rimer and Kreuter (2006). They describe tailored health communication as: ‘a process of creating individualized communication between patient and physician’. The authors qualify tailoring as an assessment-based approach in which data about a specific individual are used to determine the most appropriate information or strategy to meet that person’s unique needs. The authors state that tailored health communications could be more effective if they are developed with a focus on the differing patient needs instead of using uni-dimensional approaches to facilitate behavioral changes. According to the authors, tailored health communications have a modest success in changing a number of cancer related behaviors, such as quitting smoking, exercising, dieting and cancer screening. Which implies a positive effect of tailored health communications on patient adherence. Furthermore, Albada et al (2009) also studied interventions in which people were provided with information about cancer risks. According to the authors, in cancer risk communication, the presentation is critical for the patient’s understanding and subsequent decision-making. Tailoring this risk communication to individual risk perceptions and risk beliefs might increase the effectiveness of this cancer risk
information. A systematic literature review was performed, in which several databases were used. Which resulted in the inclusion of a total of forty studies. The authors
conclude that information that was tailored increased patients’ perception of cancer risks and knowledge of cancer compared to information that was generic (untailored). Which implies that tailoring communication to an individuals’ specific needs will increase the effectiveness of the communication. Another study that focused on the quality of patient-physician communication is the study of Linn et al (2014). In that study the quality of nurses’ communication was examined in terms of patient satisfaction. Tailoring has been advocated to be a promising strategy to improve patient medication adherence. Linn et al (2014) suggest that communication between the patient and physician, about the
treatment and medication, should be tailored to the specific individual patient needs and preferences, in order to ensure maximum patient satisfaction and therefore patient medication adherence. Zolnierek & DiMatteo (2009) also examine the importance of communication in the patient-physician relationship. These authors conducted a meta-analysis that allowed them to estimate the overall effects, of both correlational and experimental studies, involving patient-physician communication. Their results show that poor communication between patient and physician leads to a 19% higher risk of non-adherence among patients. As with the Linn et al (2014) article, these authors conclude that improving the quality of communication between patient and physician will be effective in improving patient medication adherence. Lastly, Hausman (2004) also investigates the effect of communication on patient compliance. Here, communication is defined as ‘formal and informal sharing of meaningful and timely information between patient and physician’. The author proposes that open communication between patient and physician will improve patient compliance with the physician’s medical advice.
tested, using the survey method. Prove was found for the positive effect of communication on patient compliance, indicating the importance of physicians’ demonstration of concern for patients as individuals (Hausman, 2004). These propositions will be applied in this research as well. If the communication between patient and physician is tailored to the patient’s individual needs and preferences, patient adherence will increase. Therefore the third hypothesis is:
H3: tailoring communication to patient’s individual needs will have a positive effect on patient adherence to medication.
Provider expertise
The last factor that will be discussed is this section is provider expertise, which in this case will be physician expertise. Physician expertise is a concept that is examined
frequently in the literature about patient medication adherence. Expertise is defined is having a particular skill or knowledge that represents the mastery of a particular subject or field (Stewart, 1989). Furthermore, Simons, Berkowitz and Moyer (1970) state that greater provider expertise will result in a greater behavioral change towards the desired position. This implies that greater physician expertise will have a positive effect on patient behavior. For instance, Wu et al (2008) qualitatively investigated factors
influencing medication adherence among patients with heart failure. After interviewing around twenty patients with heart failure, the authors concluded that patient’s perceptions about their healthcare provider’s expertise facilitated patient medication adherence. Another study that investigated the effect of physician expertise is the study of Kim et al (2004). These authors investigated the effects of physician empathy on patient
satisfaction and compliance. They did so through a survey among 550 patients, in a large university hospital in Korea. Part of this survey, was an examination of the effect of patient perceived provider expertise on patient compliance. Their results indicated that among the variables that were included in the research (trust, physician empathy, patient-physician partnership), patient perceptions about their providers’ expertise was actually one of the best predictors of patient compliance. This indicates that if patients have positive perceptions about their physician’s expertise, they are more likely to adhere to their physician’s instructions. This relationship will be tested in this research as well. Therefore the fourth hypothesis of this research is:
H4: Perceived physician expertise will have a positive effect on patient adherence to medication.
2.4 What factors constitute the patient beliefs about the treatment? And what is their effect on the patient attributes?
According to Roselund et al (2014), most patients actively chose not to adhere to their prescribed medication or their doctor’s orders. Some of these patients decide so because they experience the side effects to be worse than the disease itself. Other patients decide to forgo on the prescribed medication because of general negative beliefs they hold against pharmaceuticals or prescribed medication, or because they fail to see the necessity for them to take their medication (Menckeberg et al, 2008). Therefore in this section, treatment necessity beliefs and concerns and disease severity will be discussed as important factors constituting the patients’ beliefs about the treatment.
Menckeberg et al (2008), test patients’ beliefs about medication as another personal factor influencing whether or not a patient will stick to its medication. They distinguish between specific beliefs about prescribed medication and more general beliefs about pharmaceuticals as a form of treatment as well as the fear for or experience of (potential) side effects of the medication (Menckeberg et al, 2008). They find that not only self-reported adherence but also adherence measured by prescription refill-records correlate with patients’ beliefs about medication. Furthermore, adherence to medication is strongly influenced by patients’ beliefs about their need for that medication and concerns about possible adverse effects (Menckeberg et al, 2008). They developed four different attitudinal types based on patients’ beliefs about medication: accepting, ambivalent, indifferent and skeptical, of which adherence measures were highest among patients in the accepting and the ambivalent groups and the lowest among patients in the skeptical and the indifferent groups (Menckeberg et al, 2008).
Horne & Weinman (1999) describe patients’ beliefs about medication and the role these beliefs play in the patients’ adherence to medication and their treatment. Adherence decisions are most frequently based on a cost-benefit assessment, in which patients balance personal beliefs about necessity of medication to maintain or improve health, against concerns about potential adverse effects of taking the medication. A key finding of the Horne & Weinman (1999) study, was that medication beliefs were stronger predictors of patient adherence to their medication than for example frequently
researched demographic factors such as age and gender. The findings also support the necessity-concerns framework the authors proposed. Further proving the notion that patients should be respected as active decision-makers who will be more instructed if
their belief in the necessity outweighs their concerns about taking the medication. Lastly, the authors find that patients with higher concern scores than necessity scores reported significantly lower adherence scores (Horne & Weinman, 1999). This necessity-concern framework will be tested in this research as well. Meaning that it is assumed that if patients hold positive beliefs about the necessity of their medication, while their concerns can be kept at a minimum, these patients will be more likely to adhere to their
medication. Therefore the fifth and sixth hypotheses of this research are:
H5: Patient beliefs about the necessity of taking their medication will have a positive effect on patient adherence to medication
H6: Patient concerns about the adverse effects of taking their medication will have a negative effect on patient adherence to medication.
The last factor that will be discussed, in this section, as an important factor
constituting the patients’ perceptions about the treatment is the patients’ perception about the severity of the disease. DiMatteo, Haskard & Williams (2007) conducted a meta-analysis in order to integrate statistical research findings from many individual studies. With a specific focus on the patients beliefs about the disease severity threat, as well as the actual severity of their illness condition. Results of their analysis state that
non-adherence is 1,5 times greater among patients that do not perceive a disease threat and the odds of adhering are 2,5 times higher if patients believe the disease to be treated is severe and a potential threat. This effect will be investigated in this research as well. Results from the study of DiMatteo, Haskard & Williams (2007) imply that if patients perceive
their disease as threatening these patients are more likely to adhere to their medication and treatment.
Therefore the seventh hypothesis of this research is:
H7: Perceived disease severity among patients will have a positive effect on patient adherence to medication.
2.5 The mediating effect of patient role clarity and patient motivation.
The factors discussed in paragraph 2.3 all constitute to the perceptions patients have towards their relationship with their physician. As mentioned above, patient participation, tailored communication and provider expertise are all hypothesized to have a positive effect on patient adherence to medication. However these factors are also assumed to influence patient role clarity. Therefore in this section the mediating effect of patient role clarity will be discussed.
First of all, Golin et al (1996) investigated the role of patient participation when visiting a doctor. They introduced a model of determinants of adherence to diabetes self-care, which incorporated the effect of patient participation. The model not only assumes a direct effect of patient participation on patient adherence, it also introduces an indirect effect through role clarity. The authors state that patient perception will affect patient adherence, directly, but also through an increased patient understanding of their treatment regime. Which implies that patient participation will have a positive effect on patient role clarity.
Second of all, Heisler et al (2002) studied the importance of patient perceptions about the communication with their healthcare provider. The authors surveyed a total of
2000 patients receiving diabetes care. One of their findings was that patients with positive perceptions about the communication between them and their healthcare provider also had a better understanding of their role in the diabetes treatment. This implies that improving the communication between patient and physician has a positive impact on patient role clarity.
Lastly, Dellande et al (2004) examined (and found a positive effect of) provider expertise on role clarity. According to the authors, expert providers are supposed to take a more active role in structuring information for their patients and clarifying the patient’s role. This will lead to increased patient understanding of the types of behaviors (e.g. how to take their prescribed medication) which are expected from them during the treatment
As described these factors all contribute to the quality of the communication between the patient and the physician and are supposed to positively affect the patient’s role clarity. However, there seems to be little up to date research investigating the mediating effect role clarity might have on the relationship between provider expertise, patient participation, tailored communication and goal setting and adherence to medication. Therefore the eighth hypothesis of this research is:
H8: Patient role clarity will mediate the relationship between, (a) patient participation, (b) tailored communication and (c) provider expertise and patient medication adherence.
The findings from Horne & Weinman (1999) and their proposition of the necessity-concern framework show considerable overlap with the Health Belief Model as discussed
by Rosenstock et al (1988). According to these authors, The Health Belief Model (HBM) proposes that health related action is dependent on the occurrence of three factors:
• The existence of sufficient motivation (or health concern) to make health issues salient or relevant.
• The belief that one is vulnerable to a serious health problem due to that disease or illness. This is often described as the perceived threat.
• The belief that following a particular health recommendation would be beneficial in reducing the perceived threat at an objectively acceptable cost. These costs refer to perceived barriers to overcome, in order to follow the health
recommendation. It includes, but is not restricted to, financial outlays.
Rosenstock et al (1988) compared these factors with the social cognitive theory and stated that the social cognitive theory made an important contribution in predicting behavior. The addition of efficacy expectations (belief that one can successfully complete the desired behavior) and outcome expectations (completing these
behaviors will lead to outcomes that will benefit one’s health) expectations provide further insight in predicting behavior and both efficacy and outcome expectations need to be positive in order for patients to be motivated to engage in the
recommended behaviors. Dellande et al (2004) also rely on the HBM to provide support for the relationship between motivation and compliance. The authors argue that the HBM postulates two key elements that determine the likelihood of patients to engage in recommended health behaviors: (1) readiness to take action and (2)
evaluation of the feasibility of the action. Based on their results Dellande et al (2004) are able to confirm their hypothesis regarding the positive effect of patient motivation
on compliance. However, it seems interesting to further look into the relationship between patient beliefs about necessity, patient concerns and patient beliefs about disease severity and motivation, since these factors are supposed to affect the degree to which patients are motivated to adhere to their treatment. Therefore the ninth hypothesis of this research is:
H9: Patient motivation will mediate the relationship between (a) patient necessity beliefs, (b) concerns about medication and (c) patient beliefs about the disease severity and patient medication adherence.
The proposed relationships are summarized in the figure (1) below.
H8a, b & c H9a, b & c
H1a H1b
H2, 3 & 4 Adherence H5, 6 & 7
Patient-‐Physician-‐ Relationship: -‐Patient participation -‐(Tailored) Communication -‐Provider Expertise
Beliefs about the Medication Treatment:
-‐Necessity-‐Concerns trade-‐ off
-‐Disease severity
3. Research methods
In this section, an overview will be given of the research methods that will be used to test the hypotheses, as discussed in the previous paragraph. The data for this survey will be gathered through the survey method. Due to the limited amount of time and resources available for this research, this choice was made. Since the constructs of patient participation, tailored communication, provider expertise, necessity beliefs and concerns, disease severity, role clarity, motivation and adherence have been measured extensively in previous studies. In these studies data was gathered mostly through the use of either a survey method or (semi) structured interviews. However, as mentioned above, the time available for data collection was limited, therefore the survey method was believed to be most appropriate. Therefore, for this study, a questionnaire was composed. This questionnaire consists of various scales, subtracted from several other studies. These scales were already tested on their reliability and validity in previous studies. The
questionnaire was handed out in either English or Dutch, since this would allow non-Dutch speaking patients to participate as well. Since the collected scales were all in English, they needed to be translated to Dutch. Through the use of back-translation, another Dutch student from the VU, raised with English as a second language, translated Dutch scales back to English again. This was done to ensure the quality and accuracy of the initial translation. During this process no considerable translation issues occurred so no further adaptations were needed. Lastly, the questionnaire was pilot tested with three participants, these participants received and read both the English and the Dutch scales, to see if all questions were clear and understandable. These participants consisted of a female medicines student of 23 years old; a 26 years old male bar manager with no
further education and lastly, a female hospital director of 62 years old. All three participants stated the questionnaire was clear and understandable to them. Therefore both questionnaires were accepted for submission.
3.1 Sample
The participants for this study are gathered through two separate routes, in order to find a sufficient amount of respondents. The first route was through the use of
Facebook. Here, a message was posted online for seven days, for it to be seen by a network of 625 people, asking friends and family to participate in this study. In order to maintain the quality of the sample, only individuals that had been prescribed medication during the last twelve months, or were taking prescribed medication at the moment the questionnaire was distributed, were allowed to participate. After potential participants replied to the posted message, they were sent an email with information about the study together with the questionnaire. Both the posted message on Facebook and an example of the email will be posted in appendix A.
The second group of participants was gathered at the cardiology outpatient clinic at the Academic Medical Centre (AMC). This route was chosen since most of the potential participants at this clinic are/were taking any sort of prescribed medication and most of the potential participants do/did see a doctor on a regular basis. Therefore, these participants would not have problems with recalling certain situations, while filling-out the questionnaire. Over the course of five days, patients were asked to participate in this study and fill-out a questionnaire, after the appointment with their cardiologist. The
participate were brought. Here, these participants filled-out the questionnaire by paper and pencil and had the opportunity to ask the researcher as many questions as necessary.
3.2 Measurement Scales
As mentioned above, existing scales were used for developing this study’s questionnaire. For all scales a five-point Likert-type scale was chosen, ranging from strongly disagee (1) to strongly agree (5). The scales were placed in the questionnaire according to the following order: Patient participation, tailored communication, provider expertise, necessity beliefs, concerns, disease severity, role clarity, motivation and adherence. These scales are showed in table 1. The full questionnaire (Dutch and English) is posted in appendix B.
Table 1: Measurement scales.
Variable Items Questions Reference
Patient participation
5 items
1. My doctor asked my advice and council, regarding treatment options.
2. I helped the doctor in planning my treatment.
3. My doctor encouraged me to make suggestions about the appropriate treatment of my illness.
4. Both the doctor and I participated extensively in planning the treatment of my illness.
5. Together, my doctor and I set goals and discuss treatment options.
Hausman (2004)
(tailored)
Communication
1 item 1. If you should grade your provider’s communication on a scale from 1 to 5, where 1 indicates very poor and 5 indicates very excellent, what grade would you give your doctor regarding the extent to which the
communication was tailored to your personal situation?
Linn et al (2014)
Expertise items trained and qualified.
2. My doctor carries out his tasks competently.
3. I believe my doctor is highly skilled at his job.
4. I feel good about the quality of the care given to me by my doctor. Necessity
beliefs
5 items
1. My life would have been impossible without the prescribed medication.
2. Without my prescribed medication, I would have been very ill.
3. My health, at present, depended upon my prescribed medication. 4. My health in the future would have depended on my prescribed
medication.
5. My prescribed medication
protected me from becoming worse.
Horne et al (1999)
Concerns 3
items
1. I sometimes worried about the long-term adverse effects of my prescribed medication.
2. Having to take my prescribed medication worried me.
3. I sometimes worried about becoming too dependent on my prescribed medication.
Horne et al (1999)
Disease severity 3 items
1. The illness, for which my
medication was prescribed, will last a short time.
2. The illness, for which my
medication was prescribed, is likely to be permanent, rather than
temporary.
3. The illness, for which my
medication was prescribed, will last a long time.
Weinman et al (1995)
Role clarity 3
items
1. My doctor has not made it clear to me how to properly take my
prescribed medication.
2. My doctor has made it clear to me how to properly take my prescribed medication.
3. My doctor has made it clear to me
Dellande et al (2004)
Motivation 2 items
1. I feel motivated to take my medication as prescribed by my doctor.
2. I do not feel motivated to take my medication as prescribed by my doctor. Dellande et al (2004) Adherence 6 items
1. When I’m ill, I always take all the medication as prescribed by my doctor.
2. I forget/forgot to take the prescribed medication.
3. I changed the dosage of my prescribed medication.
4. I stopped taking the prescribed medication for a while.
5. I decided to skip one of my prescribed medication dosages. 6. I take my medication less then as prescribed by my doctor.
Hausman (2004); Tommelein et al (2014)
Independent variables
Patient participation was measured with the scales from the study of Hausman (2004). An example question is: My doctor asked my advice and council, regarding
treatment options. Tailored communication was measured with the scale from the study
of Linn et al (2014). An example question is: If you should grade your provider’s
communication on a scale from 1 to 5, where 1 indicates very poor and 5 indicates very excellent, what grade would you give your doctor regarding the extent to which the communication was tailored to your personal situation? Provider expertise was measured
with the scales from the study of Dagger et al (2007). An example question is: I can rely
on my doctor to be well trained and qualified. For the beliefs patients have about the
necessity of their prescribed medication or concerns about their prescribed medication, the study of Horne et al (1999) was used. Example questions are: Without my prescribed
effects of my prescribed medication. Disease severity was also measured with scales from
the study of Horne et al (1999). An example question is: the illness, for which my
medication was prescribed, is likely to be permanent, rather than temporary. Mediators
The mediators of this study, patient role clarity and patient motivation, were measured with adjusted scales from the study of Dellande et al (2004). These scales were adjusted in order to improve their suitability to the context of this study. Example
questions are: my doctor has made it clear to me how to properly take my prescribed
medication/ I feel motivated to take my medication as prescribed by my doctor. Dependent variable
Lastly, the dependent variable adherence was measured with scales from the studies of Hausman (2004) and Tommelein et al (2014). Two sets of scales were combined here in order to measure the adherence construct more broadly with more questions. This choice was made order to enhance the suitability of the scales regarding this thesis’ context. An example question is: when I’m ill, I always take my medication
as described by my doctor.
In the next chapter the data, collected through the questionnaire described above, will be analyzed. This chapter will discuss the different statistical models that will be used for the analysis together with the results from these analyses.
4. Analysis and results
The program SPSS statistics version 22.0 from IBM was used for the analysis of this thesis. The first step of the analysis was to check the data for missing values. Since the researcher was actually present when participants at the AMC filled out their survey, and respondents who participated through email were asked to completely fill out the survey, only a small amount of missing values were found. A total of only eight missing values were found among the email respondents. Therefore the choice was made to ‘exclude these missing values list-wise’ from the analysis. This way the analyses will only run on cases, which have a complete set of data. Furthermore, cases with at least one missing value in the specified variables will be dropped from the analyses. The second step was to recode the reverse-coded items in the survey. In the survey, disease severity (DS1), role clarity (RC1), motivation (Mot2) and adherence (Adh2, 3, 4, 5, 6) were qualified as reverse coded. These values of these items have been recoded before proceeding with the analysis.
After these two steps, the general information provided by the participants was analyzed, in order to get a clear picture of the sample. Of the 135 participants, 78 participated by filling in the survey in person, at the AMC and 57 participated in this study through email. In this sample, 48% of the participants was male and 52% was female. The average age of the participants was 50.43 years. Most participants indicated ‘university’ as their highest level of education with 29.9%, followed by HBO with 20.9% and MBO with 20.1%. Only 14.9% of the participants quit their education after high school and 8.2% indicated that they received another sort of education. Regarding their health condition, 50.8 % rated their current health condition as normal (which indicates
their health as neither being good nor poor) and 31% even rated their health condition as good. Lastly, Of the 135 participants 18.2 % rated their current health status as poor. Most respondents were/are being treated for a chronic disease with 69.4% as compared to 30.6% being treated for a non-chronic disease. 87.4 % of the participants had last seen a doctor within the past six months and 12.6% had last seen a doctor between six months and a year ago. Finally, 88.1% were taking any sort of prescribed medication and 11.9% was not, at the time of participating in this study. Table 1 shows an overview of the characteristics of this sample.
Table 1. Sample characteristics
Average age 50.43 years
Gender Male: 48%
Female: 52%
Education 1. High School: 14.9%
2. MBO: 20.1% 3. HBO: 20.9% 4. University: 29.9% 5. Other: 8.2%
Health condition Good: 31%
Normal: 50.8% Poor: 18.2%
Disease type Chronic: 69.4%
Non-chronic: 30.6%
Last doctor’s visit Within the past six months: 87.4 %
Between six months and a year ago: 12.6% Currently taking
medication Yes: 88.1% No: 11.9%
Note: N=135.
Next, two factor analyses were conducted in order to check if the individual items in the survey clustered/grouped together into the intended constructs of: patient
participation, tailored communication, provider expertise, necessity and concern beliefs, disease severity, role clarity, motivation and adherence. Initially the factorability of all 32
found with an eigenvalue exceeding one. The results of this analysis indicated that the items RC1, Adh1 and Adh2 did not load on the other items of their intended construct. An examination of the Kaiser-Meyer Olkin measure of sampling adequacy suggested that the sample was factorable (KMO= .765) and Bartlett’s test of sphericity was significant (χ2 (469) = 2514.181, p < .01). The Pattern matrix and scree-plot of this analysis can be found in appendix C.
After the first factor analysis, the reliability of the scales was tested as well. In order to test the reliability of the constructs used in this study, the cronbach’s alpha was calculated for each construct. These values are shown in brackets in table 2. For the constructs patient participation, provider expertise, necessity beliefs and concern and disease severity the cronbach’s alpha was initially already sufficiently high, therefore no items needed to be deleted from these constructs. The initial cronbach’s alpha of role clarity was .548, but this could be improved to .851 by deleting item RC1. The initial cronbach’s alpha of adherence was .753, but this could be improved to .799 by deleting item Adh2. Therefore the items RC1 and Adh2 were excluded from the analysis. The cronbach’s alpha of the construct motivation was insufficiently high with .464. However, since motivation is one of the key mediators in this study the choice was made to include it in the regression analysis for now (please note that this will be further discussed in the limitations section). Lastly, since the construct of tailored communication only consisted of one item, no cronbach’s alpha could be calculated for this variable. After excluding the items RC1 and Adh2 from the analysis, a second factor analysis was made. A second Principal Axis Factor was conducted, in order to check the effect of deleting the items mentioned above and a total of eight factors were found with an eigenvalue exceeding
one. An examination of the Kaiser-Meyer Olkin measure of sampling adequacy
suggested that the sample was factorable (KMO= .777) and Bartlett’s test of sphericity was significant (χ2 (435) = 2424.161, p < .01). The Pattern matrix and screeplot of this analysis can be found in appendix C. This pattern matrix showed only the item Adh1 did not load on to the other Adherence items. Therefore the choice was made to exclude this item from further analyses. No other unexpected tendencies were found in this analysis; therefore the other constructs were taken through the regression analysis as intended. Lastly, for each of the constructs patient participation, tailored communication, provider expertise, necessity beliefs and concerns, disease severity, role clarity, motivation and adherence, the mean was calculated, for inclusion in the regression analysis.
Before starting with the regression analysis inter-correlations between the construct and the cronbach’s alpha were calculated. These results are shown in table 2. Table 2. Scale Means, SD's, Intercorrelations, and Reliabilities.
