How will patient intrinsic motivation, perceived capability
and own influence impact on patient adherence:
with the moderating effects of patient emotional reliance
and neuroticism personality trait
Author: Chang Liu Student No.: 10830448
Thesis supervisor: Dr. Antoon Meulemans Date: Aug 18, 2017
Statement of originality
This document is written by Student Chang Liu 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
Abstract
...
3
1. Introduction
...
4
2. Literature review
...
8
2.1 Patient (non-)adherence
...
8
2.2 Patient self-efficacy
...
10
2.3 Internal health locus of control (ILOC)
...
14
2.4 Intrinsic motivation
...
16
2.5 Patient perceived empathy
...
18
2.6 Patient trust
...
19
2.7 Neuroticism personality trait of patient
...
21
3. Theoretical framework
...
24
4. Methodology
...
27
4.1 Sample
...
27
4.2 Measurement of variables
...
28
4.2.1 Dependent variable (Adherence)
...
28
4.2.2 Independent variable - Intrinsic motivation
...
28
4.2.3 Independent variable - self-efficacy
...
29
4.2.4 Internal health locus of control
...
29
4.2.5 Moderator - Patient perceived empathy
...
30
4.2.6 Patient Trust
...
31
4.2.7 Moderator - Neuroticism
...
32
4.2.8 Control variables
...
33
5. Results
...
33
5.1 Missing values and recoding
...
33
5.2 Factor analysis and reliability test
...
34
5.3 Hypothesis testing
...
37
5.3.1 Hypothesis 1
...
37
5.3.2 Hypothesis 2
...
38
5.3.3 Hypothesis 3
...
39
5.3.4 Hypothesis 4
...
40
5.3.5 Hypothesis 5
...
42
6. Discussion
...
44
List of tables and figures
6.1 Hypotheses results
...
44
6.2 Theoretical and managerial implications
...
46
6.3 Limitation and future research
...
47
7. Conclusion
...
49
References
...
50
Appendix 1. English questionnaire
...
61
Appendix 2. Dutch questionnaire
...
63
Appendix 3. Chinese questionnaire
...
67
Appendix 4. SPSS results
...
71
Figure 1. Commonsense Model of Self-regulation
...
11
Figure 2. Protection Motivation Theory
...
12
Figure 3. Needs Continuum.
...
17
Figure4. Theoretical model
...
26
Table 1. KMO and Bartlett test for 30 items
...
34
Table 2. Factor analysis for variables
...
34
Table 5. Principal component analysis
...
35
Table 7. Skewness and Kurtosis
...
36
Table 8. Correlation matrix
...
36
Table 9. Regression_H1
...
37
Table 10. Regression result_H2
...
38
Table 11. Regression result_H3
...
39
Table 12. Regression result_H4
...
41
Table 13. Regression result_H5
...
42
Figure 5. Refined model
...
44
Table 3. Eigenvalues of 26 items
...
71
Table 4. Scree plot for 26 items
...
72
Table 6. Descriptive statistics
...
72
Abstract
The problem of patient adherence has perplexed healthcare providers for years, with many researches been done to see the factors influencing adherence and many results have been reached. This study will explore this issue based on the Need Continuum in (Ingwer, 2012)’s empathetic marketing theory, which helps give new explanations by treating patients as customers and find the influence of their inner emotional needs. The main relationships being studied are the effects of different levels of patient intrinsic motivation, self-efficacy and internal health locus of control (ILOC) on patient adherence. Patient perceived empathy, trust and patient personality have been investigated as moderators with either positive or negative effect on the main relationships. By conducting a survey among 325 participants mainly from China and Netherlands, the author have reached several conclusions. Patient intrinsic motivation is proved significantly positively correlated with patient adherence. Surprisingly, ILOC exhibit a significant negative impact on patient adherence. However, the impact of self-efficacy as well as the two moderators (patient emotional reliance and neuroticism personality) haven’t been proved in this study. Patient emotional reliance, composing of patient perceived empathy and patient trust hasn’t been proved to moderate the relationship between patient intrinsic motivation and patient adherence. This study contributes to the current understanding of customer behavior in a rather rarely tapped-into context, when involved in a medical interaction. It gives explanation for patient adherence from a marketing perspective, which provides healthcare providers with new insights when developing interventions to improve adherence. All the unproved relationships in this research point out new directions for future investigation on patient adherence.
Key words: patient adherence, intrinsic motivation, internal health locus of control,
self-efficacy, big five, neuroticism, emotional reliance, perceived empathy, patient trust, patient perceived capability and influence,
1. Introduction
Patient adherence is that “they take their medication as prescribed and continue to take during the required period” (Ho, Bryson & Rumsfeld, 2009). Many researches have been done to explore the factors that enhance patient adherence and those might trigger non-adherence, however, mostly in a medical or psychological perspective. As patient adherence is a special kind of customer behavior, the author would like to find out whether the overlap of this topic with marketing will give some new insights. So far, the research of patient adherence using a marketing thought, especially through the emphasis of customer emotional needs is done very little and this study aims at contributing in this new branch.
Though it seems that patients have no choice but to follow whatever their doctors tell them to do to recover from illness, they often choose to behave in a non-adherent way. For example, Brown & Bussell (2004) pointed out the white-coat adherence, which refers to that patients improve their medication-taking several days before or after the appointment with their doctors. It’s of great importance that patients adhere to the treatment as it will lower the chance of rehospitalization and help insure a more positive health outcome and surely save extra medical costs for the patients. With so much benefits, this research is conducted with great value to find out truly powerful predictors of patient adherence. Taking a look at past researches, Health Belief Model has the most widespread influence. Patients perceived threat, perceived benefits and barriers as well as self-efficacy are the main elements in the model (Rosenstock, Strecher & Becker, 1988). Perceived threat, benefits and barriers are mainly
related to disease itself and often result in contradicting findings when tested in different contexts. Although widely used, disease features, drug, symptoms, prognosis and clinical setting are reported as poor indicators of compliance (Urquhart, 1999; Vermeire, Hearnshaw, Van Royen & Denekens, 2001). Apart from Health Belief Model, WHO elaborates there are five categories of reasons causing poor adherence, relating to socioeconomic status, health care team and system, disease, therapy and patient (Brown & Bussell, 2011). Low income, limited health literacy, prescription coverage, limited visiting time, restricted access and lack of health information technology all contribute to poor adherence (Browne & Merighi, 2010; Brown & Bussell, 2011).
Health belief model is criticized by Becker & Maiman (1975) as only addresses the negative aspects of health and they suggested that positive health motivation also accounts for variance in adherence. It is argued by Ryan & Deci (2000) that people get intrinsically motivated when feel the satisfaction of three fundamental needs of humanbeing, which are autonomy, competence and relatedness. They go to two opposite directions, to be autonomous and competent and at the same time feel connected (Ingwer, 2012, p.p 30). In his book Empathetic Marketing, Ingwer (2012, p.p 39) demonstrated “six principal emotional needs of customer - control, self-expression, growth and recognition, belonging and care”. He drew the Need Continuum (see figure 1). This helps lay the foundation for current research, which is to find out the underlying elements facilitating patient adherence when marking the importance of emotional needs. It is believed to be requisite that the emotional needs of patients are satisfied before they truly committed to their treatment regimen, whether severely or slightly ill. So far, the impact of patient intrinsic motivation on adherence hasn’t been explored and this research will fill this gap.
Vlasnik, Aliotta & Delor (2005) said that “ideal patient adherence should be performed as an knowledgable choice within a helpful surrounding”. Therefore, patient
should be seen as an active problem solver rather than only a complier to the medical regimen, which is exactly self-efficacy (in the Health Belief Model) addresses. As most researches are done to explore HBM generally and only a few focused on the influence of patient self-efficacy on adherence, an empirical study focusing on self-efficacy is needed. Besides self-efficacy, patients may demonstrate their own influence in the medical interaction through locus of control. This is in line with the control need of customer in Ingwer (2012)’s work. Internal Health Locus of Control (ILOC) is to describe how much a person perceive his or her expected health status is dependent on him or herself. While External Health Locus of Control is to describe how much a person perceive his or her expected health status is relying on powerful outsiders (Christensen & Wiebe, 1996). These two concepts are sometimes confused to be the same thing, although actually they are not, therefore the individual effect of ILOC and self-efficacy is going to be tested in this research and they are grouped into a superior concept - patient perceived capability and influence.
Patient adherence can be seen as a special kind of customer loyalty taking place in a medical context. Positive emotion facilitates customer loyalty (Yu & Dean, 2001). Patients may receive technically competent care and inadequate ‘psychological care’ as stated by Squier (1990). If patients are not psychologically fulfilled, they may respond with a non-adherent behavior pattern. This is in line with the need for belonging and care in Ingwer (2012)’s Need Continuum. Patients may perceive empathy from doctor or social support, together with patient trust, it will make patients feel emotionally supported. There are researches about patient trust and adherence, hitherto, little research has combined it with patient perceived empathy and they are grouped together as one superior concept - patients’ emotional reliance in this study.
Booth-Kewly & Vickers (1994) argues that personality works as a better predictor of health behavior because it describes the general behavior predispositions and tends to
remain stable in adulthood. Big five personality model provides five dimensions to measure individual differences, which are conscientiousness, neuroticism, extraversion, openness and agreeableness (Barrick & Mount, 1991). Among which, Neuroticism is the dimension most closely related to emotion. It describes the tendency to experience negative, distressing emotions. The direct influence of personality on adherence has already been tested and neuroticism turns out to work in a negative way in most cases (Piette, Heisler, Krein & Kerr, 2005; DiMatteo, Sherbourne, Hays, Ordway,…& Rogers, 1993; Christensen & Smith, 1995). So far, it hasn’t been tested whether neuroticism will exert certain influence on self-efficacy or ILOC. In this research, whether high neuroticism trait of a person will weaken the positive effect of patient self-efficacy or ILOC will be studied.
The research question will be How will patient intrinsic motivation, perceived capability and own influence impact on patient adherence: with the moderating effects of patient emotional reliance and neuroticism personality trait. The sub questions are: 1). How will intrinsic motivation influence patient adherence? 2). Will ILOC and patient self-efficacy hold a positive influence on patient adherence? 3). How will patients’ emotional reliance (perceived empathy and trust) moderate the relationship between patients’ intrinsic motivation and adherence? 4). Will the Neuroticism trait of the patient negatively influence the relationship between patient self-efficacy as well as ILOC and patient adherence?
Theoretically, this study will be the first one to group testing the influence of intrinsic motivation, self-efficacy and ILOC on patient adherence. It will contribute to the understanding of the importance of emotional needs (control, care, relatedness) fulfilling in a special customer context - medical treatment. By adding in patient’s emotional reliance and neuroticism personality trait in the model, it will test how much the patients will be emotionally influenced by outsiders and themselves in the adherent process. Is it more important that the patients get intrinsically motivated or getting enough qualified emotional
reliance? If the patient holds a neuroticism character, will their self-efficacy become weakened? As these questions will be answered in this research, it will help healthcare providers to identify truly important factors influencing patient adherence and work out certain procedures to improve adherence. Improved adherence will help patients reach satisfied health outcome, lower health risk and save the cost of rehospitalization both for the health care system and most importantly for patients themselves.
In chapter 2, the current literatures relating to topic will be reviewed. In chapter 3, the conceptual framework and hypotheses will be presented. Chapter 4 will discuss about methodology and data analysis and results will be presented in chapter 5. Discussion will be presented in chapter 6. Chapter 7 is the conclusion of the study.
2. Literature review
2.1 Patient (non-)adherenceWhen talk about patient adherence, another word compliance is often seen as its synonym. Compliance, however, used in a doctor-centered perspective and this power imbalance give up the chance for patient self regulation (Conrad, 1985) and self management (Zullig et al., 2013). The term adherence reflects the power balance between patients and doctors and convey the importance of collaboration in patient-doctor interaction (Vermeire et al., 2001).
There are different ways to measure adherence, from patient self-reported to biomedical examination. The white-coat adherence, which refers to patients improve medication-taking before medical encounter, will make biochemical indicators such as plasma drug concentration not reliable to prove patient adherence performance (Urquhart, 1999).
Later, microprocessor attached to drug bottles has been invented to detect ‘drug holidays’ (Vermeire et al., 2001).
Poor adherence will aggregate budget waste because of rehospitalization and intensive treatment for emergency for both mental and physical illnesses. Irregular drug use among schizophrenia patients predicts number of hospital days (Svarstad, Shireman & Sweeney, 2001). Long-term adherence among patients who have had a cardiovascular event is still poor (Brown & Bugsell, 2011). On the other hand, regular intake of drug strongly reduces the risk of coronary heart events (Urquhart, 1999). Forty-five percent medical cost of severe asthma patients will be saved through optimal adherence (Vlasnik et al., 2005). Although adherence will lead to higher drug costs, the much larger reduction in medical costs it produces is of high leverage benefit for the entire health system (Sokol, McGuigan, Verbrugge & Epstein, 2005).
Unintentional non-adherence refers to “patients’ intention to follow the prescription is blocked by unpredicted events, for example, they just didn’t remember to take the medication” (Brown, Rehmus & Kimball, 2006). When a regimen is too complicated, it might cause unintentional non-adherence easily (Bentford, Milligan, Pike, Anderson, Piercy & Fermer, 2012). When reduce pill burden by fixed-dose, the poorly adherent diabetic patients become fairly adherent (Bentford et al., 2012). Intentional non-adherence is patients deliberately decide not to follow the prescribed regimen, they may even overuse to achieve shorter treatment duration (Brown et al., 2006). Clifford, Barber & Horne (2008) found that unintentional non-adherers are not significantly different from adherers.
Non-adherence is one kind of problem health behavior. Prochaska, Velicer, Rossi, Goldstein, …. & Rossi (1994) explored the decisional balance influences in different stages of change among twelve problem health behaviors and they addresses the sequence of changing stages: “pre-contemplation, contemplation, preparation, action and maintenance”. Staying
adherent through the duration of regimen is certainly a kind of maintenance. Maintenance requires the patient to feel confident about self-capability as well as always to stay motivated in the treatment duration. In the following sections, patient self-efficacy, ILOC and intrinsic motivation will be discussed.
2.2 Patient self-efficacy
When address self-efficacy in medical context, the cognitive mediating process of Rogers’ Protection Motivation theory (See Figure 3 in appendix) confirms its importance (Floyd, Prentice-Dunn & Rogers, 2000). Protection motivation refers to a person’s desire to perform in an adaptive way (beneficial health behavior) or maladaptive way (harmful health behavior) (Armitage & Conner, 2000). People calculate the potential threat and their coping potential and then formulate their final protection motivation. If both are perceived high by a person, his or her protection motivation will be high, thus, later involve in adaptive coping. To stay adherent is among adaptive coping decisions. If adapt protection motivation theory in the adherent behavior, threat appraisal is that the patient will compare the perceived severity and vulnerability with the rewards to not stay adherent. For example, if a patient perceives high rewards from smoking and the perceived threat causing by smoking is considered to be low, it will result in low threat appraisal. Then he or she will act in a maladaptive way - won’t quit smoking. This is similar to perceived threat and benefits in Health Belief Model. Leventhal categorizes five features to be considered in representation of health threat according to his Commonsense model (see figure 2), which are “identity, timeline, cause, consequence, cure/ control” (Ross, Walker & MacLeod, 2004).
Figure 1. Commonsense Model of Self-regulation
The common-sense model of illness representation: Theoretical and practical considerations. Diefenbach, M. A., & Leventhal, H. (1996). Journal of social distress and the homeless, 5(1), 11-38. p.p21.
As for coping appraisal in Protection Motivation theory (see figure 3), the patient will compare the perceived usefulness of adherence (response efficacy) as well as self-efficacy to stay adherent and the cost correlating to staying adherent. Thus, self-self-efficacy is of same importance as the factors relating to disease threat highlighted in HBM. Self-efficacy is the belief that the action required to produce the expecting result can be executed (Kaplan, Atkins & Reinsch, 1984). To stay adherent might require the patient to take medication on time, do routine examination, on diet, if the patient feel incapable to fulfill the task (low self-efficacy), the chance to go non-adherent is high. In Bandura (1989)’s social cognitive theory, he defined three kinds of efficacy judgments: outcome expectancy, self-efficacy as well as expectancies about environmental cues (Kaplan et al., 1984). Environmental cues expectancy is also similar to the perceived susceptibility and severity in Health Belief Model (Rosenstock
et al., 1988). Outcome expectancy is “the belief that the recommended behavior will bring an appreciative outcome” (Kaplan et al., 1984), which is similar to response efficacy in Protection Motivation theory. As we see so much overlapping in concepts, self-efficacy appears important when people involving in health behavior, however, rarely been used to predict adherence.
Figure 2. Protection Motivation Theory
A meta-analysis of research on protection motivation theory. Floyd, D. L., Prentice-Dunn, S., & Rogers, R. W. (2000). Journal of applied social psychology, 30(2), 407-429. p.p 410
Ajzen’s Theory of Planned Behavior addresses self-efficacy in a general context. Ajzen came up with Theory of Planned Behavior (TPB) by adding the third element - Perceived Behavioral Control (PBC) in Theory of Reasoned Action (TRA). Theory of Reasoned Action is criticized as only addressing volitional behavior - the success of performance only requires the formation of an intention (Armitage & Conner, 2000). In both TPB and TRA, intention is the proxy lead to behavior. In TPB, besides attitudes and subjective norms (social pressure), intention is co-determined by perceived behavioral control as well. In Ajzen (2002)’s Hierarchical Model of Perceived Behavioral Control, PBC is composed of two subordinate components: controllability and self-efficacy. Controllability is the belief whether performing the behavior is up to themselves or not. Ajzen (2002) found that self-efficacy significantly predict intention but does not predict behavior, which again
forwards the uncertainty that patient with high self-efficacy will be more adherent, which will later be tested in this study. Kaplan et al. (1984) said that efficacy judgments significantly correlate with internal locus of control and uncorrelated with external locus of control. Ajzen (2002) clarifies that self-efficacy does not correspondingly link to internal barriers and facilitators. This is why health locus of control will be discussed in the next part, as it is different from self-efficacy and there should be an enlarged effect when combine the influence of ILOC.
Bandura suggests there are several sources of self-efficacy and the most influential one is enactive performance attainments, which is based on individual past success (Rosenstock et al., 1988). People who have repeated success desire more information and decision-making power to further improve their social competence (Braman & Gomez, 2004). Repeated success and self-efficacy will work together in spiral up direction. Kaplan et al. (1984) argued that limit of performance attainment is that self-efficacy won’t be strengthened through success of dissimilar behavior. The second source of self-efficacy is from vicarious learning by observing others’ performance result and the third source is verbal persuasion, which can be realized by health education in adherence issue (Rosenstock et al., 1988).
In Protection Motivation theory, rewards for maladaptive coping is parallel with the influence of self-efficacy and reward is often mentioned in operant studies. Operant theorists argue that how often the behavior is performed depends on the corresponding reinforcements and only the temporal link between a certain behavior with its reward will sufficiently increase the performing chance (Rosenstock et al., 1988). The effectiveness of operant theory for maintaining behavior is criticized, Leventhal & Cameron (1987) found that when stimuli to elicit adherence or the rewards to reinforce adherence being removed, high relapse happened as the cues for non-adherence still existed. When Godin & Kok (1996) tested the predicting power of Theory of Planned Behavior in both health-promoting
behaviors (exercise, screening) and problem health behavior (eg. addictive substance use, risky behavior) and found self-efficacy was as important as attitude towards health behavior. People with high self-efficacy are more likely to execute and pay the effort to maintain certain behavior (Ajzen, 2002). Self-efficacy is especially required when there are complex behavior requirements in the regimen (Rosenstock et al., 1988). The less the self-efficacy, the less effort the person will pay and the more easily they will quit when facing obstacles (Kaplan et al., 1984). Mann, Ponieman, Leventhal & Halm (2009) stated that disease-specific self-efficacy (feeling confident to control current disease) is much more related to adherence than traditional generic self-efficacy (feeling confident to control future health). Christensen & Wiebe (1996) defined perceived health competence as a person’s perceived self-efficacy to successfully influence own health outcomes. They also found that perceived health competence does not exert an influence on adherence if the patient hold high internal health locus of control. In the next part, we are going to discuss the so-far found effect of internal health locus of control on adherence.
2.3 Internal health locus of control (ILOC)
Standing on one end of the need continuum (Ingwer, 2012), control is believed to be important for customer. “When people purchase with a sense of control, even if later the perceived worth is not as much, the purchase still give them a feeling of satisfaction” (Ingwer, 2012 p.p 50-51). In the medical interaction, customer control is elaborated as health locus of control (HLOC). Internal health locus of control is the belief that one’s own behavior is the determinant of one’s own health status, while external health control beliefs are multi-facets that external influencers (god, powerful others, chance) are the determinant of one’s health status (O’hea, Grothe, Bodenlos, Boudreaux, White & Brantley, 2005). For example, chance HLOC is the degree to which people perceive fate or fortune is in control of their health status
(Braman & Gomez, 2004). Situation-outcome expectancy in Bandura’s social cognitive theory is that environment, rather than the individual himself accounts for the outcomes (Armitage & Conner, 2000), is a kind of external locus of control. An individual with high level external HLOC isn’t correspondingly with low level ILOC. As patients who hold both high external and internal health locus of control show most favorable adherence (Christensen & Wiebe, 1996). A slightly different finding from O’hea et al. (2005) is that type two diabetes patients who held high internal HLOC and meanwhile low chance HLOC show best adherence. Internal ILOC is the critical HLOC belief when predicting health behaviors as when patients hold both high chance HLOC, the ones with high internal HLOC will be more adhere (O’hea et al., 2005). Though with high self-efficacy, if the patient holds external HLOC belief, he or she will probably not adhere because they are not convinced it will lead them to a desirable outcome (Rosenstock et al., 1988). Young patients hoping for miracle (external HLOC) are less likely to follow the medical regimen and rely on avoidant coping (Sherbourne, 1992). In Stanton (1987)’s research, hypertensive patients with internal HLOC are more favorably adherent.
Perceived control may positively influence health through increasing coping efforts from chronically ill patients and help to reduce stress (Helgeson, 1992). Seligman (1975)’s Learned Helplessness theory showed that after repeated control efforts that ends in failure, the person will generate a feeling of helplessness (feel deprived of control), negative affect and non-adherence (Helgeson, 1992). Besides actual control, the perception of control is also seen as with certain advantages (both internal and external). Rothbaum, Weisz & Snyder (1992) defined illusory control as an “unreasonable belief that either chance or skill-determining situations is within personal control by spending own effort. They proposed one different opinion of external believers. They want to be lucky and see fortune as an outcome which can be achieved through personal trying (Rothbaum et al., 1992). When patients
perceive primary control is not possible, they choose to submit the power to healthcare provider to realize a perception of secondary control, which help them maintain the sense of own influence over health status (Christensen & Wiebe, 1996). The extent of positive cognitive distortions, Helgeson (1992) argued that, including illusions of control can only do good to patients when they are not exaggerated too much from real facts.
2.4 Intrinsic motivation
Motivation is essential when maintained change is demanded (Ryan & Deci, 2000) and patients need to continuously overcome barriers during the required duration of treatment (Armitage & Conner, 2000).
Ryan & Deci (2000) set up the Self-Determination Continuum model with motivation and Intrinsic Motivation standing at the two extremes. When a person act out of amotivation, it’s suppose that he or she does not value the activity and feel incompetent while performing, at the same time, without expectation of a desirable result. Intrinsic motivation is “the innate inclination to extend and execute one’s capabilities, a human tendency toward learning and creativity” (Ryan & Deci, 2000). The transition from amotivation to intrinsic motivation is a process of internalization, through four different stages of extrinsic motivations, which are external regulation, introjected regulation, identified regulation and integrated regulation. Internalization is a person gradually transform extrinsic contingencies and social values into personal values and later by performing certain behavior, a person can achieve inherent satisfaction and enjoyment (Ryan & Deci, 2000). Three basic human needs - autonomy, relatedness and competence are essential for the internalization process. Alienation, hindered perceived competence, feeling being controlled too much will all undermine intrinsic motivation by causing the failure during internalization. If the three basic needs of patients are satisfied, they will become intrinsically motivated to adhere. This is also
in line with Ingwer (2012)’s customer Need Continuum (Figure 1). The need for control, self-expression, growth/recognition, belonging and care appears in order from the individuality end of continuum to the connectedness end of continuum. Customers need all the needs fulfilled at the same time. This is proved in Ryan & Deci (2000)’s theory, as they argued that the need for belonging and connectedness is important for the internalization process to intrinsic motivation and feeling of competence will better facilitate this process with a sense of autonomy or internal perceived locus of causality.
Figure 3. Needs Continuum.
Ingwer, M. (2012). In Empathetic Marketing (pp. 39). Palgrave Macmillan US.
As relatedness is needed during internalization process, Ryan & Deci (2000) argued that when perceive certain behavior is prompted or taken seriously by important people they are related to, a person will be willing to perform certain job. This has been tested in the medical interaction as well. Davis (1968) found that if physician gave no feedback after collecting information from patients, the patients was likely to react with reciprocal non-adherence. Becker (1975) also stated that patient adherence decreases when physician gave no feedback after a long interview. It might be patients’ need for belonging is ignored by the deviant emotionally unresponsive style from physician and this feeling of detachment erodes patients’ intrinsic motivation to adhere.
The importance of relatedness and attachment is also being stressed in Rothbaum et al. (1992)’s Two-Process Model of Perceived Control. People will seek secondary vicarious
control by linking themselves to powerful others when the strive for primary control didn’t work out. By self-defining to a intrinsically desired group of people, the person will also realize secondary vicarious control. This perception of secondary control (one kind of control cognition distortion mentioned in 2.3) is realized through connectedness - one core customer emotional need. It’s important to fulfill patients’ emotional needs as Watson & Pennebaker (1989) suggested that emotional inhibition (insufficient emotional needs fulfilling for a person with high social desirability) would cause negative health consequences.
2.5 Patient perceived empathy
Buller and Buller (1987) categorized two communication styles of physicians: affiliative style - the physician tries to establish and maintain a positive relationship by acting in a friendly, empathetic, honest and devoted and nonjudgmental way. Although patients’ satisfaction with physicians’ communication determined three fourths variance in their general satisfaction with medical care, when the anxiety with patients increases, which is common among more severely ill patients, the importance of communication style is reduced (Buller & Buller, 1987). It’s explained by Larson & Yao (2005) that patients’ need for empathy goes beyond mere communicating skills from doctors. Doctors’ general responsiveness to patients’ emotion and the capability to convey their emotional responses strongly correlated with patients’ satisfaction and satisfied patients are more likely to comply (Squier, 1990). Becker (1975) found that when patients felt that their complaints understood by doctor, their adherence would increase. The more the patient is satisfied with the affective concern rather than only the technique competence from their physician, the more committed they are to the medical regimen (Segall & Burnett, 1980). Physician empathy is a kind of emotional labor, which requires both internal and external emotion management and is different from manual or mental labor (Larson & Yao, 2005). Physicians either show empathy
sincerely or use deep acting to meet patients’ empathy expectations (Larson & Yao, 2005). It is only when the physician pay the effort to actively role-taking in the patients’ perspective that they can respond in a sincerely empathetic way. Most of the time, it’s easier for the physician to involve empathetically when there are some similarities between the patient and the physician, however, with a much more powerful situation (eg. the patient is severely ill), individual similarities won’t matter so much (Larson & Yao, 2005). Squier (1990) suggested the presence of empathy is essential for a good practitioner-patient relationship.
Patients also perceive empathy when receive social support. Sherbourne (1992) suggested that it is the quality of social support is much more important than mere existence of support when influencing adherence. DiMatteo (2004) confirms this with the conclusion that practical support shows significant positive effects on patient adherence, while structural support - whether living with someone else, only exert slight influence on adherence. Apart from tangible practical support, affectionate/emotional support is needed by patients and most of the time comes from family cohesiveness (Sherbourne, 1992; DiMatteo, 2004). Good quality social support help relieve negative feeling, foster self-esteem and a sense of control (Sherbourne, 1992).
2.6 Patient trust
Pearson & Raeke (2000) defined patients’ trust in physician is “their belief that the physician will behave in line with their expectations”. Patient trust is different from patient satisfaction. A patient may be satisfied with several interactions with the physician but still not feel the establishment of trust in physician. On the contrary, trust in physician can not be easily changed even if there are several visits after which the patient is not satisfied (Thom, Hall & Pawlson, 2004). Patient satisfaction is to look backward, based on judgments about
past experience with physician. Patient trust is to look forward as it’s the expectation of physician in future interaction (Thom et al., 2004).
Kao, Green, Zaslavsky, Koplan & Cleary (1998) claimed that enough selections of own doctor, longer relationship with their doctor and trust in the entire care organization help build patients’ trust. Patients’ trust also engenders from satisfied assessment of physician rapport, understanding, honesty and compassion (Pearson & Raeke, 2000). Thom et al. (2004) clarified three important attributes of physician that leads to patient trust, which are technical competency, interpersonal competency and acting in patient’s interest.
Piette et al. (2005) suggested that patients’ trust in physician has more predicting power than patient satisfaction when to predict later adherence performance and found that low income leads to cost-related poor adherence only when patients does not trust their physicians. Patients’ trust in physicians is positively correlated with patients’ engagement in the treatment recommendations and regular medication (Thom, 1999; Pearson & Raeke, 2000). Vermeire et al. (2001) contrasts that the importance of patients’ own experience of illness and medication is beyond their perception and expectation of health care providers. Ciechanowski (2001) mentioned four kinds of attachments in medical practice defined by Bowlby: secure, dismissing, preoccupied and fearful. The two dimensions to form the attachment type between patients and care provider are “view of others -whether physicians can be trusted” and “view of self - worthy of care”. Secure attachment refers to a patients hold positive views of both self and others. Dismissing attachment is that the patient think he or she is worthy of care, however, does not trust the physician. Preoccupied attached patients hold the exact opposite view to dismissing attached patients. And it’s found that diabetes patients with dismissing attachment showed severely worse adherence than patients with preoccupied or secure attachment, which again indicates the importance of patient trust on medical adherence.
Patients tend to place higher value on prescription when they trust their physician (Piette et al., 2005), that is to say patient trust positively correlates to adherence through increased perceived effectiveness of care. When the patient find the physician as trustworthy, he or she will expect better health outcome from physician’s recommended treatment (Lee & Lin, 2009).
2.7 Neuroticism personality trait of patient
Bandura argued that specific expectancies in certain situations are not generalized personality characteristics (Kaplan et al., 1984). To predict patient adherence, besides patient self-efficacy and control expectancies, patient personality is also important to be taken into consideration. It’s been justified in past studies. Cardiac patients scoring high on psychotics show poor adherence (do not quit smoking) (Wiebe & Christensen, 1996). Patient personality predicts preference for patient-doctor relationship, whether in a paternalistic in which doctor do not share too much information and make treatment decisions or egalitarian in which patients possess access to information and the share the right in decision making. According to Big Five model developed by Costa & McCrae (1985), there are five dimensions to describe personality: conscientiousness, agreeableness, extraversion, openness and neuroticism.
A person scoring high on agreeableness is more likely to shown an accepting gesture rather than acting cynically and hostilely (Digman & Takemoto-Chock, 1981). Barrick & Mount (1991) assigned the following traits to agreeableness: trusting others, polite and humble, easygoing, compassionate, forbearing, indulgent and helpful. Agreeableness is related to better self-care and better blood glucose control among diabetes patients (Booth-Kewly & Vickers, 1994). Greater cynicism, mistrust and general oppositional style of agreeableness is correlated with poor adherence (Wiebe & Christensen, 1996). Openness is
also referred as Intellectence (Peabody & Goldberg, 1989). Openness is frequently used to describe a personnel exhibiting the following traits: creative, full of curiousity, unconventional, liberal, and high in aesthetic sensitivity (Barrick & Mount, 1991). People scoring high on this dimension is found to be more easily commit to substance abuse (Booth-Kewley & Vickers, 1994). Brown (2012) found in his study that openness significantly accounted for stigma in those with a severe mental illness (SMI). Major depression patients show better adherence for taking antidepressant drugs when they perceive low level stigma of being mentally ill (Sirey, Bruce, Alexopoulos, Perlick, Friedman & Meyers, 2001). Perceived stigma in Sirey’s study is the fear that others may forward a unfriendly and judging gesture when knowing that they are among a discredited group of people who are severely mentally ill. Sirey et al. (2001) predicted that perceived stigma may be even more influential among physically ill patients. Traits belong to extraversion are: being convivial, active and assertive (Barrick & Mount, 1991). Hogan (1986) pointed out the two outstanding elements of extraversion are being ambitious (to take lead, reckless) and sociable (self-expressive, like to show-off). People who are extraverted are more likely to involve in substance taking (smoking, alcohol) (Booth-Kewly & Vickers, 1994). Davis (1968) found that patients score high on assertiveness (high score on extraversion) might fall in a deviant interaction pattern with their physician showing solidarity and are less likely to adhere in which the friendly physician is ignored. John & Srivastava (1999) suggested that conscientiousness describes the different level of inclination to follow subjective norm of restrictions for impulsiveness. Clark & Watson (1999) suggests that the six traits of conscientiousness possess different predictive power of health behavior: self-control and traditionalism are most consistent predictors, followed by virtue and responsibility, industriousness and order are the least relevant. Discipline and persistence are the key roles in maintaining positive health behavior. To be future-oriented and willing to involve in long-term target-chasing are both among the essence
of conscientiousness (Booth-Kewly & Vickers, 1994). This is later confirmed by Bogg & Roberts (2004) that conscientiousness-related traits would prevent an individual from any kind of dangerous health behaviors and facilitate all kinds of promoted health behaviors. It is because people score high in constraint are more concerned with the accumulated effects of their behaviors, which prevents them from involving in risky health behaviors (Clark & Watson, 1999).
Neuroticism refers to individuals' different level of inclination towards “negative, distressing emotions” and is supposed to be predictive for the cognitive or behavior aspects of an individual (Watson & Pennebaker, 1989). Health distress refers to frustration, despair and other negative emotions (Sherbourne et al., 1992). Neuroticism will elicit maladaptive health practices and prohibit adaptive health practices (Booth-Kewly & Vickers, 1994). It is also mentioned by Wu et al. (2008) that health distress has a negative effect on adherence. It’s found that patients with depressive symptoms are twice as likely than those without depression to perform cost-related underuse by Piette et al. (2005). DiMatteo et al. (2000) found substantial relationship between depression and noncompliance while the average association between anxiety and noncompliance is small and nonsignificant. Neuroticism also influences patient symptom perception, which causing reporting bias (Wiebe & Christensen, 1996). Negative affectivity (NA) is correlated with subjective health indicators (such as health complaints) and largely unrelated to objective health indicators (Watson & Pennebaker, 1989). So NA could influence patients’ health behavior. But it’s not sufficient to argue that patients who complain more would be more adherent. On the contrary, trait anxiety and somatic focus patients show poor adherence as their cognitive judgment about symptoms is misled (Wiebe & Christensen, 1996). Christensen & Smith (1995) reported that the significant negative effect of neuroticism on adherence will disappear when control the conscientiousness dimention. Patients with health distress are more likely to behave in a non-adherent way as
perceive no beneficial effects of the behavior (Sherbourne et al., 1992), accordingly, perform in a non-adherent way. Depression often triggers patient hopelessness and it’s hard for patients holding little optimism to consider regimen worthwhile which would lead to patient noncompliance (DiMatteo et al., 2000). In this study, one dimension - Neuroticism of Big Five personality will be focused.
3. Theoretical framework
Customers possess six core emotional needs - control, self-expression, recognition, belonging and care according to Ingwer (2012, p.p 44)’s need continuum. The facilitation to intrinsic motivation, internal health locus of control and self-efficacy, together, elaborate part of the continuum. These three factors are treated as independent variables in this study. Operant theorists focuses on extrinsic motivation and facilitate it through contingent rewards (Ryan & Deci, 2000) and causing high relapse rate (Leventhal & Cameron, 1987) of non-adherence. Williams, Rodin, Ryan, Grolnick & Deci (1998) indicate that patients perceiving high autonomy see few obstacles to adhere. Intrinsic motivation of patients is predicted by this study to result in lasting positive effects on adherence. The first hypothesis of this study is as follows:
H1. Patient intrinsic motivation is positively related to patient adherence.
Patients’ self-efficacy and their internal HLOC are grouped as one superior concept in this study - patient perceived capability and influence, and are supposed to positively relate to adherence:
H2. Patient self-efficacy has a positive effect on patient adherence.
When doctors are able to create a relationship in which they share their patients’ concerns about illness, the patients’ motivation to get better will be strengthened (Squier, 1990). Adherence will increase when patients feel sincere concern and sympathy from physicians (Squier, 1990). Social isolation and withdrawal from the essential persons who provides the patient with emotional support and assistance will cause patient depression (DiMatteo et al., 2000). DiMatteo (2004) suggests that social support benefit patients’ adherence through relieving depression and stress, ameliorate interpersonal affective states and on the other hand, by enhancing patients’ self-efficacy, intrinsic motivation, personal control and self-esteem. Patient perceived empathy is sourced from outsiders, patient trust in physician generated from physician’s interpersonal competency. Trust and emotional attachment are found to be powerful predicting factors for behavioral loyalty (Sui & Baloglu, 2003). In this study, as perceived empathy and trust both elaborates on a sense of relatedness, they are grouped together as - patient emotional reliance and supposed to strengthen intrinsic motivation’s positive effect on adherence. Thus, it’s hypothesized that:
H4. Patient emotional reliance (perceived empathy and trust in doctor) will positively moderates the relationship between intrinsic motivation and patient adherence.
Yu & Dean (2001) explored how emotional satisfaction could contribute to consumer loyalty and found that affective components (including both positive and negative emotions) work better than cognitive components when predicting customer loyalty, among which positive emotion is with the highest predicting power. They explained that a person’s decision to continuously involve in certain events is influenced by positive emotions (Yu & Dean, 2001). If in this case, patient adherence is also one kind of loyalty - stay loyal to treatment, positive affections might support the patient to adhere constantly. In other words, high scoring in neuroticism may influence adherence in a negative way as neuroticism is to describe a person easily filled with negative affection and emotions. Apart from illness
representation (patient’s accurate linkage of symptoms to a health condition, patient belief of causes, perception of consequences and belief of timeline and controllability), Leventhal’s Commonsense Model of Health and Illness Self-regulation (see Figure 2) also addresses the emotional coping process of patient (Diefenbach & Leventhal, 1996). Patients’ emotion coping is of comparable importance and is independent of their cognitive coping (Leventhal & Cameron, 1987). Browne & Merighi (2010) found that depression has an impact on cognition, later causing cognitive impairment and limit patients’ ability to make sound judgements. However, they didn’t mention what kind of cognitive impairment. In this study, as self-efficacy and internal health locus of control are cognitive thinking of patients, it’s worthwhile to test whether patients’ neuroticism will weaken the possible positive effect of self-efficacy and internal HLOC on adherence. Thus, it’s hypothesized that:
H5. Neuroticism will negatively moderates the effect of patient perceived self-capability and influence on patient adherence.
4. Methodology
4.1 SampleThe survey method is used in this study considering its advantage: time-saving, cost-efficient and reaching a sizable population (Saunders & Lewis, 2012). The survey is conducted in three languages, as the original survey questions are in English, it is translated into Dutch and Chinese. The Chinese questionnaire is translated by the author and examined by a fellow Chinese student at UvA and proved to be fully expressed the meaning of questions in English and easy to understand. The Dutch questionnaire is translated by several Dutch-speaking student at UvA and using a translation and back-translation method to ensure the quality of the questionnaire. Chinese questionnaire is spread out and collected data on a most frequently used academic survey website sojump.com, Dutch and English questionnaires are both online spread through qualtrics.com and collected face to face by the author in hospitals. The respondents either answer the survey according to the current disease condition or based on a past medical experience.
Convenience sampling and snowball sampling (Saunders & Lewis, 2012) are used in combination during the survey collecting process. The author reaches the Dutch respondents at Amsterdam Medical Center and Chinese respondents through social media. Then the initial Chinese respondents spread the survey to suitable acquaintances. During a period of one week, 325 questionnaires are collected both practically and in digital. The ones which didn’t fill in carefully are excluded and left with 278 completed questionnaires, reaching a response rate of 85.5%. Participants are mainly from 19 to 55 years old, taking a percentage of 90%. There are 203 female participants (73.02%) and 75 male participants (26.98%). The majority (81%) of the participants acquired a bachelor degree or higher and 12% of the participants received the education till high school. More than half of the
participants are in a normal health condition, 31% of them are self-reported as in a good or excellent health condition and 12% of them are in a poor health condition.
4.2 Measurement of variables 4.2.1 Dependent variable (Adherence)
One item measuring the dependent variable in this study is from Morisky, Green & Levine (1986)’s self-reported measure of medication adherence. The other items are from the General Adherence Scale developed by DiMatteo, Hays, Gritz, Bastani, Crane,…& Marcus (1993):
Among which, item 2 and 4 are reverse-coded questions. 4.2.2 Independent variable - Intrinsic motivation
The items to test the extent to which the patients feel intrinsically motivated to follow the treatment regimen are adapted from Choi, Mogami & Medalia (2009)’s Intrinsic Motivation Inventory. Choi et al. (2009) validated a self-report intrinsic motivation scale originally used in a treatment adherence context. The scale was highly associated with relevant constructs of motivation for health behaviors, including autonomous treatment engagement and perceived capability for treatment challenges. The original Intrinsic Motivation Inventory that Choi et al. (2009) based on includes 6 subscales, which are subjective experiences of interest/enjoyment, effort, value/usefulness, pressure/tension,
Patient adherence: DiMatteo et al. (1993) Morisky et al. (193)
1. I took all medications recommended.
2. When I feel better, I sometimes stop taking my medicine.
3. I found it easy to do the things my doctor suggested me to do.
4. I didn’t follow my doctor’s suggestions exactly.
relatedness and perceived selection freedom. And the items testing effort, enjoyment, pressure are included in this study:
Among which, item 1 is reverse-coded. 4.2.3 Independent variable - self-efficacy
Self-efficacy items is adapted from Anderson, Funnell, Fitzgerald & Marrero (2000)’s study measuring self-efficacy among diabetes patients. Self-efficacy can also be measured by the items from Perceived Health Competence Scale (PHCS) (Marks & Lutgendorf, 1999):
Among which, no items are reverse-coded. 4.2.4 Internal health locus of control
The items measuring internal HLOC are sourced from the Multidimensional Health Locus of Control Questionnaire originally developed by Wallston, Strudler & DeVellis
Intrinsic motivation: Choi et al. (2009)
1. The activities involved in the regimen do not hold my attention at all.
2. It is important to me to do well at all the tasks for my treatment.
3. I tried very hard while following treatment regimen.
4. I did not feel pressured while following my treatment requirements.
Self-efficacy:
Anderson et al. (2000); (Luszczynska & Schwarzer, 2005)
1. In general, I believe that I can ask for support for caring for my disease when I need it.
2. When taking care of my disease, in general, I believe that I know which part(s) I’m not satisfied with.
3. When taking care of my disease, in general, I believe that I know which part(s) I’m ready to change.
4. I’m generally able to accomplish my goals regarding to my health.
(1978). In this study, items in form C version developed by Wallston, Stein & Smith (1994) are used:
Among which, item 2 and 4 are reverse-coded. 4.2.5 Moderator - Patient perceived empathy
As patients get a certain feeling of empathy not only during the interaction with physician but also when they receive social support. The questions to test patient perceived empathy are separated in two parts: degree of perceived empathy from physician as well as from social support.
The items to test the degree of patient perceived empathy from physician are sourced from (Kim, Kaplowitz & Johnston, 2004). Kim et al. (2004) modified items originally developed by Barrett-Lennard (1981) in his Relationship Inventory for perceived-physician empathy. The items are separated in two parts, which are cognitive empathy and affective empathy. For this research, only affective empathy items are included and the Cronbach’s alpha for these items in Kim et al. (2004) are below 0.70.
Internal HLOC: Wallston et al. (1994)
1. The main thing which affects my health is what I myself do.
2. Regarding my health, I can only do what my doctor tells me to do.
3. If I take the right actions, my condition should improve or at least not get any worse. 4. Whatever improvement occurs with my condition is largely a matter of good fortune.
Patient perceived empathy from doctor: Kim et al. (2004)
1. My doctor responds to me mechanically. 2. My doctor tries to keep me from worrying. 3. My doctor shows caring about my psychological well-being.
4. My doctor shows great concern for my well-being.
Patient perceived empathy from social support is adapted from one item from Broadhead, Gehlbach, De Gruy & Kaplan (1988)’s measuring scale of functional social support. Thus, with a focus of patient perceived empathy from social support. It is supposed the more social support they receive, the higher their perceived empathy from social support.
Among which, only item 1 is reverse-coded. 4.2.6 Patient Trust
Hall, Zheng, Dugan, Camacho, Kidd, Mishra & Balkrishnan (2001) identified the five dimensions of patient trust: loyalty, expertise, honesty, privacy and global. Hall, Dugan, Zheng & Mishra (2002) came up with a measurement of patients’ trust in their primary care providers. They chose the 10 best-performing items to form a new multidimensional scale and the validation of the questions have already been tested. The items turned out be strongly related to patients’ loyalty - their desire to remain with a physician and willingness to recommend to friends.
Among which, only item 4 is reverse-coded. Patient perceived
empathy from social support:
Broadhead et al. (2004)
5. I get help when I’m sick in bed.
6. I get people who care what happens to me. 7. I get chances to talk to someone about problems at work or with my housework when I’m ill.
Patient trust: Hall et al. (2002)
1. My doctor will do whatever it takes to get me all the care I need.
2. My doctor is extremely thorough and careful.
3. My doctor is totally honest in telling me about all of the different treatment options available for my condition.
4. My doctor does not pay full attention to what I’m trying to tell him or her.
4.2.7 Moderator - Neuroticism
The test of overall personality traits (including all five dimensions) is conducted in this study to prevent the respondents from focusing too much on the questions about Neuroticism dimension, which will causing bias. McCrae, Costa & Martin (2005) modified their old NEO Personality Inventory and the instruments show better inter-correlation and are more understandable. Also the items from Gosling, Rentfrow & Swann (2003) and Rammstedt & John (2007)’s scale measuring personalities are also used in this study.
Among which, item 2 and 3 are reverse-coded. Neuroticism:
McCrae et al. (2005); Gosling et al. (2003); Rammstedt & John (2007)
1. I see myself as anxious, easily upset.
2. I see myself as someone who is relaxed, handles stress well.
3. It doesn’t bother me too much if I can’t get what I want. Conscientiousness, Extraversion, Agreeableness, Openness: McCrae et al. (2005); Gosling et al. (2003); Rammstedt & John (2007)
1. I see myself as generally trusting.
2. I see myself as someone who does a thorough job. 3. I see myself as critical, quarrelsome.
4. I see myself as someone who has few artistic interests. 5. Sometimes I don’t stand up for my rights like I should. 6. I try to go to work or school even when I’m not feeling well.
7. I act forcefully and energetically.
8. I would rather keep my options open than plan everything in advance.
9. Once I find the right way to do something, I stick to it. 10. I see myself as someone who is outgoing and sociable. 11. When making laws and social policies, we need to think about who might be hurt.
Among which, item 1, 3 (reverse-coded) and 11 are measuring Agreeableness. Item 2, 6 and 8 (reverse-coded) are to measure Conscientiousness. Item 5 (reverse-coded), 7 and 10 is to measure Extraversion. Item 4 (reverse-coded), 9 (reverse-coded) and 12 is to measure Openness.
4.2.8 Control variables
The control variables in the survey are general health condition (from “1” very poor to “5” excellent), gender and age. The educational level, current disease or disease treated before and the times see the doctor in the last year are also being asked.
All the questions from 4.2.1 to 4.2.6 are being rated in a 7-likert scale (“1” Totally disagree to “7” Totally agree). The complete questionnaire in English Dutch and Chinese are listed in appendix 1, 2, 3.
5. Results
5.1 Missing values and recoding
The survey sent out in paper are examined when collected back and all the missed out questions are being asked again. The questions on digital platform are set in forced answering mode except question asking the current disease, thus no missing value detected. As the question about current or before treated disease is relating to privacy, the participants are given the freedom to not answer and about 140 participants did not answer this question. The answers concerning this part are rather scattered, the most often mentioned bothering disease is gynaecology with seven participants. Other commonly seen diseases are allergy, cardiovascular disease, neurology, etc. And all the answers towards the reverse-coded questions in the survey are coded back. The SPSS 24.0 is being used for data analysis.
5.2 Factor analysis and reliability test
The overall Kaiser-Meyer-Olkin measure of sampling adequacy for the 30 items is .814, well above .7, the items are considered adequate. Bartlett’s test of sphericity for all the items is .000, which indicates the significance standard is satisfied (see table 1). The KMO value for the items measuring Neuroticism and internal HLOC are both below .6, which is not satisfying and this will be addressed in the limitation part. This might due to the large data set (278 participants) this study included and different nationalities of sample. As in this test, patient perceived empathy and patient trust is combined as one variable - patient emotional reliance to be tested in the hypothesis 4, the combined items are tested also.
Table 1. KMO and Bartlett test for 30 items
Table 2. Factor analysis for variables
KMO and Bartlett’s Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy .814
Bartlett’s Test of Sphericity Approx. Chi-Square 2677.574 Df 435 Sig. 0.000 Variable Items no.
KMO Bartlett’s test for sphericity Cronbach’s alpha Neuroticism 3 .574 .000 .544 Adherence 4 .642 .000 .609 Intrinsic motivation 4 .635 .000 .615 Self-efficacy 4 .631 .000 .65 Internal ILOC 4 .569 .000 .503 Perceived empathy 7 .758 .000 .765 Patient trust 4 .776 .000 .816 Patient emotional reliance 11 .881 .000 .861
Next, for all the independent variables and moderators, further factor analysis is conducted for dimension reduction. Excluding the 4 items measuring adherence, the 26 items are tested. Seven items score over 1 on initial eigenvalues and in all explained 60% of the variance. It’s also shown on the scree plot that the most important 2 items explains a lot variance with a steep drop on the third most important item (See table 3. Eigenvalues of 26 items, table 4. Scree plot for 26 items). Principle component analysis for items are done, and 3 items for Neuroticism is connected to each other and works out as one component, which proves the items are qualified to stand for the moderator variable - neuroticism and suitable for further analysis.
Table 5. Principal component analysis
The descriptive information (Std, Mean) for each item can be found in table 6 (see appendix 4).
All the items standing for one variable are being added together and given a short name for next step analysis. First, the distribution of the data is tested. For neuroticism, the skewness is .071, and kurtosis is -.674, so the tail of the data is slightly to the left direction of the mean value and it’s not normally distributed and platykurtic. The kurtosis for self-efficacy is 2.593, thus in a shape of leptokurtic. The skewness of self-efficacy is -1.026, the data is tailing to the left and concentrate on the right side of the mean value. The others are shown below.
Table 7. Skewness and Kurtosis
Component I get anxious or upset easily. .679
Big5_N2 .762
Big5_N3 .660
Skewness Kurtosis
A correlation analysis is done to see the possible relationships between each pair of variables.
Table 8. Correlation matrix
All the Pearson correlation value is below .7, so the chance of multicollinearity is low and next the regression analysis to test the hypothesis is to be conducted in next part.
Intrinsic motivation -.173 -.246
Internal HLOC -.396 .493
Self-efficacy -1.026 .593
Patient emotional reliance (EMOS)
-.16 .204
Neuroticism .071 -.674
ADR IntriMo ILOC SelE Big5_N
IntriMo Pearson Correlation .498** 1 Sig. (2-tailed) .000
ILOC Pearson Correlation -.017 .166** 1
Sig. (2-tailed) .78 .006
SelE Pearson Correlation .186** .379** .291** 1
Sig. (2-tailed) .002 .000 .000
Big5_ N
Pearson Correlation -.026 -.236** -.109 -.161** 1
Sig. (2-tailed) .666 .000 .07 .007
EMOS Pearson Correlation .219** .325** .125* .356** -.266**
Sig. (2-tailed) .000 .000 .038 .000 .000
** 0.01 significance level * 0.05 significance level
5.3 Hypothesis testing 5.3.1 Hypothesis 1
It’s hypothesized that patient intrinsic motivation will be positively correlated to patient adherence. A hierarchical linear regression analysis is being conducted. The three control variables, which are age, gender and general health condition are entered first to ensure the relationship is explained by intrinsic motivation alone.
Table 9. Regression_H1
In both steps, the relationship is significant, with only the control variables, p = . 05 (age, general health condition, gender) and p = .000 (intrinsic motivation). In table 6, for general health condition (p = .000, β = .135) which means that general health condition is significantly positively related to adherence. R2 change for intrinsic motivation is .23 (p = .
000) (F = 23.691), it indicates that by controlling other variables, intrinsic motivation explains 23% variance in adherence. With β = .495, intrinsic motivation is significantly positively related to adherence, and hypothesis 1 is supported. This indicates that the more the patients are intrinsically motivated, the more likely they adhere to the treatment regimen.
Model R R2 St. Err. R2 change F β t sig. a. 1 .167 .028 3.263 .028 2.63 .089 1.419 .157 a. 2 .135 2.235 .026* a. 3 .106 1.724 .086 b. .508 .3 2.856 .23 23.691 .495 9.191 .000**
a. 1) Age, 2) general health condition, 3) gender
b. Age, general health condition, gender, intrinsic motivation Dependent variable: ADR
