The effect of the communication of perceived behavioral control and the integration of age-related adjustments on the compliance rate of a preventive E-health intervention for older employees

(1)

The effect of the communication of perceived behavioral

control and the integration of age-related adjustments on the

compliance rate of a preventive E-health intervention for older

employees

Master thesis D.M. van Doorn

Author: Daniëlle van Doorn Student number: 10673067 Primary supervisor: Dr. René Bohnsack 2nd_{reader: Francesca Ciulli Date of submission: 27-06-2014 Track: Strategy}

(2)

Preface

Amsterdam, June 27th, 2014

As part of my master program in Business Studies at the University of Amsterdam I have written a thesis related to my specialization Strategy and my interest in health management. The thesis concerns means to improve the compliance of older employees with an online intervention to help them improve their health status. This research has led to interesting results and implications. I have been privileged with the supervision of Dr. René Bohnsack. He has given me the freedom to specify this research in the areas of my interest and has helped me to reach this final result. I would like to thank my family, girlfriends and boyfriend for their unstoppable patience and support. And especially I would like to thank my parents. This thesis could not have been written without their support and love.

Kind regards, Daniëlle van Doorn

(3)

Abstract

The upcoming years, the European Union (EU) will generate workforces that are the oldest in history (Arandjelovic, Milic, Radevic, Lekovic, & Gavrilovic, 2008). To be able to retain healthy workforces, health interventions should be implemented at the workplaces of older employees. In this research the effect of the communication of perceived behavioral control and the integration of age-related adjustments on the compliance rate of a preventive E-health intervention for older employees has been investigated. The results of this research show that there is a positive effect of the communication of perceived behavioral control and the integration of age-related adjustments on the compliance rate of older employees of an E-health intervention. First, the communication of perceived behavioral control in an intervention increases the perceived ownership of skills and the non-existence of barriers. Second, the integration of age-related adjustments into the intervention increases older employees their confidence and motivation to use the intervention. Therefore, developers of E-health interventions that integrate these two factors are able to create more value for their customers, differentiate themselves from their competition, and are therefore able to generate higher compliance rates of their E-health interventions in the future.

(4)

List of figures, tables and graphs

List of figures

Figure 3.3.1 Elements of SCT……….. 20 Figure 3.3.2 Elements of TPB……….. 20 Figure 3.3.3 Elements of HBM………. 21 Figure 3.3.4 Elements of FBM………. 22

Figure 3.3.5 Elements of the integrative model of behavioral prediction……… 23

Figure 3.5 Derived model, based on Fishbein (2000) and literature on age-related changes….. 29

Figure 4.1 Different steps in this research……… 34

Figure B(1) Home page of the E-health intervention in the survey, without ARA………. 67

Figure B(2) Home page of the E-health intervention in the survey, with ARA………... 69

Figure B(3). Video in the survey, showing an exercise for shoulders and neck……….. 70

Figure B(4) Situation sketch, home page and video shown in the survey……….... 71

List of tables

Table 3.1.1.Employer objectives driving their use of WHPPs……… 13

Table 3.2.2 Articles on the effectiveness of E-health interventions………. 16

Table 3.4.4 Age-related changes and subsequent adjustments for E-health interventions…….. 26

Table 3.5 2X2 matrix intention-behavior relationship and implications for interventions…….. 28

Tabel 3.6 Conditions of the experiment performed in this research……… 30

Table 4.3(.1) 2x2 factorial design of this thesis……… 36

Table 4.3(.2) Independent variables, constants and dependent variable in this research………. 36

Table 5.1(.1) Average age of the participants (that filled in their age) in absolute numbers and percentages……… 39 Table 5.1(.2) Number of participants per condition in absolute numbers and percentages……. 40

Table 5.1(.3) Average WTP per age group………... 40

Table 5.3 Correlation matrix between the variables age, WTP, ARA en PBC……….... 41

(7)

Table 5.4 Bootstrapped BCa 95% confidence intervals per variable……… 43 Table 5.6 Mean per condition in Euros……….. 44 Table 6.3 Revenues per WTP amount……… 48

List of graphs

Graph 1.1 Age distribution of the workforce in the EU in 2030………... 7 Graph 3.6.2 WTP for a health intervention……… 32

(8)

1. Introduction

The upcoming fifty years, the demographic distribution of the European Union (EU) will change dramatically. The statistical office of the EU (Eurostat) predicts that in 2060 about 30 percent of the population in the EU will be 65 or older (Eurostat, 2012). This trend is caused by the combined effect of low fertility and longer life expectancy of people (Rechel et al., 2013). In the EU, the mortality rate among older people is fallen considerably since 1970. This fall in mortality among older people is largely explained by improved lifestyle, and prevention and treatment of cardiovascular disease by means of the expansion of health care systems (Mackenbach, Slobbe, & Looman, 2011). Therefore, the ageing population can be seen as the outcome of, and the challenge for, European health systems (Rechel et al., 2013).

Reasonably, this older population will have an impact on the demographics at the workplace. During the next few decades, the EU will face an era in which their workforces will be the oldest in history (Arandjelovic et al., 2008). This is the result of baby boomers reaching their retirement and the fact that only a small proportion of the total population will be of working age in the future. These two factors will result in a change in structure of European workforces (See graph 1.1).

As a result of this change in demographic distribution, the competitiveness of the EU will depend on the contribution of older employees (Arandjelovic et al., 2008). Pension reforms and other actions of the EU member states are planned to encourage older people to work for a longer period of time. The effect of a person his or her work on their health is considered as one of the main factors of influence on the decision to work at older age (Arandjelovic et al., 2008). Hence,

Graph 1.1 Age distribution of the workforce in the EU in 2030 (Arandjelovic et al., 2008)

(9)

the most important place where change must happen is the workplace itself. The EU must acknowledge the effect of work on health and determine which regulations and policies will extend workability and health at older age.

Employees and employers should work together to tackle health issues. Employees can extend their working life by improving their individual health and lifestyle. Employers can invest in employees their health by using health promotion programs to change employees their health behavior (Soler et al., 2010). Preventive workplace health interventions can make a difference in an older employee his or her health and therefore the health of the total workforce. At this moment, about 42 percent of the employers use workplace health interventions (Hall, Hunt, & Ratcliffe, 2012). However, interventions like personal trainers, fitness gyms and personal nutrition coaches bring with them tremendous costs. Next to this, it is not for certain that these interventions will actually make a difference in employees their health behavior because a great deal of these interventions simply fails (Anderson et al., 2009; Hutchinson & Wilson, 2012; Rongen, Robroek, van Lenthe, & Burdorf, 2013a).

In recent years, progress has been made to decrease the costs of workplace health interventions. One medium that has a great deal in bringing these costs down is the internet (Anderson et al., 2009; Hutchinson & Wilson, 2012; Rongen et al., 2013a). By using the internet, health interventions can reach a greater audience (i.e. workforce) and interventions can be bundled into compact online systems. Consequently, the internet has become an often used tool for health promotion (Korp, 2006). Online health interventions are often mentioned as E-health interventions. E-health interventions vary from simple applications on the smartphone to more complex multicomponent online systems (Das & Svanæs, 2013). Examples of E-health interventions that could improve employees their health behavior are online dieting coaches, relaxation exercises and advice on posture during work.

Although the internet has advanced health interventions, still the results of these interventions show inconsistencies. One great problem which causes these inconsistencies is the compliance rate of E-health interventions (Lankton & Wilson, 2007; Neuhauser & Kreps, 2010). The majority of people is great at forming positive intentions to eat healthier or to exercise more with the help of an E-health intervention. However, most people fail at turning this intention into real behavior (Sheeran, 2002).

In this thesis there will be a focus on older employees, because this group will increase in size and therefore importance in the future (Eurostat, 2012). There will be a closer look to the factors that increase compliance with an E-health intervention; especially to how this intervention could create more value for older employees. Therefore, the central research question in this thesis

(10)

is How to increase compliance of older employees with a preventive E-health intervention in

order to change their health behavior?

This research could contribute to the field of health management because there is an increasing need to develop E-health interventions based on theory (Pingree et al., 2010). Ideally, E-health interventions should be developed by first identifying the outcome(s) to target. Second, studying the mechanisms that affect these outcomes and then work backwards to design the E-health intervention to activate at least some of the mechanisms (Pingree et al., 2010). This research will integrate this approach by first looking at the target outcomes of the E-health intervention of interest, second studying human behavior theories and literature on older employees to finally come up with suggestions to establish a high compliance rate of the E-health intervention.

The focus of this research will lie on the effects of the communication of perceived behavioral control (PBC) and the integration of age-related adjustments (ARA) on the compliance rate of an E-health intervention. Previously, studies have been done on the effect of integrating ARA into health interventions to be suitable for older people (Brauner, Valdez, Schroeder, & Ziefle, 2013; Gerling, Schild, & Masuch, 2010; Ijsselsteijn, Nap, & Kort, 2007; Kaufman et al., 2006). Next to this, several studies have communicated PBC within their E-health interventions for older people (Ijsselsteijn et al., 2007; Kaufman et al., 2006). However, the effect of the integration of these variables on the demand of older employees for the intervention is not yet researched. This thesis will look at the individual effects of both variables and, if possible, at their interaction effect. It is assumed that when both variables are included, the E-health intervention creates most value for its users and therefore the demand and compliance of older employees with the E-health intervention will increase.

To be able to answer the main research question, chapter two will first discuss the E-health intervention of interest. Chapter three will review literature on workplace E-health promotion, E-health, human behavior theories, older employees, the derived model, the hypotheses and compliance and willingness-to-pay. Chapter four will describe the methodology. Chapter five will discuss the results of this research. Next, chapter six will contain the discussion. Finally, chapter seven will elaborate on the conclusions of this research.

(11)

2. The E-health intervention of interest

This chapter will introduce the E-health intervention of interest in this research. The focus of this preventive E-health intervention is on improving the well-being of older employees with a sedentary function. It is a preventive intervention because it wants to avoid physical and physiological issues at older age by regular exercise and mental training starting at a relatively early age (+55). The intervention includes four pillars: physical training; stress management; nutritional balance and workplace ergonomics. These four pillars lead together to an enhanced quality of life in the working environment of older employees in addition to an increase in their overall wellbeing.

The first pillar of the E-health intervention is physical training. Today, multiple organizations offer their employees the use of sport facilities. However, these facilities bring with them tremendous costs and therefore these facilities are often not an option for smaller or less fortunate organizations. Another way of promoting exercise in organizations is to provide employees information about small exercises to perform on their own by means of videos on the internet. However, a common problem of this approach is that there is no personal feedback on the correctness of the movements during the exercise. The E-health intervention of interest tackles this problem by means of a 3D-sensor which detects the movements of the user. Therefore, feedback about the accuracy of the movements is gathered and tips can be given to improve the exercise. The exercise is delivered in a traditional way via videos on the internet or via exergames. An exergame is a combination of videogames and exercise. This term will be explained in more detail in chapter 3.2.3. The exercises in this intervention are all based on the Fogg behavioral model; this model will be explained in chapter 3.3.4.

The second pillar of the E-health intervention concerns the workplace ergonomics. Workplace ergonomics consists of guidelines for optimal distances to the desk or monitor et cetera. When attention is given to ergonomic standards, back pain or postural deformity, but also vision problems and tensions can be prevented. With the earlier mentioned 3D sensor, the position of employees can be tracked and, if necessary, tips are offered to improve the current position. Next to this, relaxation exercises can be provided for the muscles and eyes of the users.

The third pillar concentrates on the nutritional balance. The intervention offers an individual nutrition scheme and reminders, for example to consume a piece of fruit or to drink more water. Food is recognized by means of a food recognition unit, by which the food is recognized through a picture (via a camera) or through spoken information. Because of the fact

(12)

that most employees consume their foods in a canteen, there will be a mobile application that recognizes and tracks their food intake.

The fourth pillar of the E-health intervention concerns stress management. This intervention prevents the occurrence of job stress in two ways. First, by increasing the internal resources, especially the cognitive and physical abilities, by providing relaxation exercises and exercises focused on physical activity. Second, the intervention prevents job stress by increasing the social interaction among employees through playing exergames together or in competition with each other.

These four pillars are integrated into one online system, which visualizes the effects of performing exercises and nutrition intake. Next to this, a desktop application will transfer the data of the 3D sensor and camera to the online system to provide reminders and additional information. In addition, the online system can give personal feedback and establish statistics based on the data of the 3D sensor and camera. The system therefore acts as an online health coach.

This chapter has introduced the E-health intervention of interest. The next chapter will discuss important literature regarding workplace health promotion, E-health, human behavior theories, older employees, the derived model, hypotheses and compliance and willingness-to-pay.

(13)

3. Literature review

This chapter will discuss literature on workplace health promotion, E-health, human behavior theories, older employees, the derived model, the hypotheses and compliance and willingness-to-pay.

3.1. Workplace health promotion programs

This part of the chapter will discuss the concept of workplace health promotion programs (WHPPs). First, the concept is defined and the major reasons to use WHPPs are explained. Second, the history of WHPPs will be summarized. Third, the effectiveness of WHPPs will be discussed.

3.1.1 Definition of WHPPs and reasons to use for employers

WHPPS aim to improve lifestyle and therefore health, work ability and productivity of employees at their workplace (Rongen, Robroek, van Lenthe, & Burdorf, 2013b). This concept incorporates the improvement of the work organization and environment, the promotion of active participation of all stakeholders in the process and the encouragement of personal development (ENWHP, 2007). Consequently, the implementation of preventive WHPPs is expected to bring about positive changes in employees their health status (Nöhammer, Stummer, & Schusterschitz, 2010)

The workplace in particular provides several advantages for health promotion in comparison with other settings (Task force on community preventive services, 2010). First, the potential for intervention exposure, promotion, recruitment and participation is high because of a relatively stable and large target audience. Second, social support systems and peer influences among co-workers can reinforce employees their efforts.

The results of a global survey on workplace wellness and health promotion strategies show that in Europe about 42 percent of all organizations offer WHPPs to their employees (Hall et al., 2012). Figure 3.1.1 shows that the main reasons for European managers to implement WHPPs are to 1) improve workforce morale and engagement 2) reduce employee absence due to sickness or disability 3) improve employee productivity (Hall et al., 2012). The major European trends for WHPPs are the use of online health coaching, on-site health coaching, personal health records, disease management programs and the access to healthier vending machines (Hall et al., 2012).

(14)

3.1.2 History of WHPPS

Today, the topic of workplace health promotion is widely known and of common interest of researchers (Burton, 2010) . However, it was quite recently that health promotion was

specifically linked to the workplace.

In 1978, the declaration of Alma-Ata was signed in which more attention was given to the consumers of health services and the wider community (Burton, 2010). Next to this, this declaration stated that primary health care should bring national health care closer to where people live and work. In 1986, the document The Ottawa Charter introduced the concept of health promotion as “the process of enabling people to increase control over, and to improve, their health” (Burton, 2010 p.11). Next to this, this document acknowledged the workplace as one of the key settings to improve a person his or her health. In 1997, consensus was reported on the concept of workplace health promotion (Burton, 2010). The European Network for Workplace Health Promotion has defined this concept as “the combined efforts of employees, employers and society to improve the health and well-being of people at work. This can be achieved through a combination of: improving the work organization and the working environment; promoting active participation and encouraging personal development” (Burton, 2010 p.12). In 2005, for the first time workplace health promotion was presented as a requirement for good corporate practice; all corporations should incorporate practices to improve employees their health at their workplace (Burton, 2010).

Table 3.1.1 Employer objectives driving their use of WHPPs (Hall et al., 2012)

(15)

3.1.3 Effectiveness of WHPPs

Quite a lot of studies have looked at the effectiveness of WHPPs. These studies show that the impact of WHPPs differs greatly. Several articles mention only small effects (Anderson et al., 2009; Hutchinson & Wilson, 2012; Rongen et al., 2013a). However, some other studies show that WHPPs do have moderate positive effects on employees their health status. For example, Muto & Yamauchi (2001) found that a four-day health promotion program with a follow-up provided over one year improved the BMI, systolic blood pressure, cholesterol and triglycerides of the employees significantly in comparison with the control group (P<0.05). Next to this, the systematic review of Soler et al. (2010) stated that an assessment of health risks of employees, with providing feedback on the results had a positive effect on tobacco use, alcohol use, dietary fat intake, cholesterol, health risks estimates, employee absenteeism and health care use. Furthermore, Hanlon et al. (1998) its observational study showed that a health check with feedback and personal health education resulted in 46% of the employees changing their behavior in the desired direction: they quitted smoking, reduced alcohol consumption, improved diet or increased the amount of exercise.

This part of the chapter has provided an introduction on the topic of WHPPs. One specific development regarding WHPPs is the use of the internet as a tool to improve the quality of the health interventions and to reduce their costs (Dansky, Thompson, & Sanner, 2006). This topic of E-health will be discussed in the next part of the chapter.

3.2 E-health

This part of the chapter will explain the importance of the internet as a tool for health promotion. Next to this, the concept of E-health will be explained and its different types will be discussed. 3.2.1 The internet as a tool for health

In recent years, the internet has become a tool for health-related information and communication between consumers (Korp, 2006). Next to this, health care institutions and health promotion agencies are recognizing the opportunities the internet has to offer for health promotion. The internet challenges the traditional one-way information delivery by the development of interactive environments “with the potential for independent action and discovery” (Manning, 1997, p.72). The internet could be seen as an intervention channel for improving consumers their health (Evers, 2006). This intervention channel could provide the consumer health information in addition to interventions or applications to make changes in their health behavior.

The rapid expansion of online health interventions is the results of multiple factors (Evers, 2006). First, technology has spread itself widely, computers have become more available and

(16)

faster browsers are now accessible for the public (Fotheringham & Owen, 2000). Second, the delivery of online interventions is frequently associated with relatively low costs (Fotheringham, Owies, Leslie, & Owen, 2000). Third, consumers are increasingly willing to manage their own health actively (Baker, 2001) as well as they are willing to use the internet as a tool for this (Eng, Gustafson, & Henderson, 1999).

3.2.2 Explanation on E-health

Online interventions are often mentioned in the literature as E-health interventions. Eysenbach (2001 p.20) defines E-health as “an emerging field in the intersection of medical informatics, public health and business, referring to health services and information delivered or enhanced through the internet and related technologies”. E-health interventions vary from simple applications on the smartphone to more complex multicomponent online systems (Das & Svanæs, 2013). Employers could implement these E-health interventions to address common workplace health problems as incorrect postures, high stress levels and unhealthy eating habits.

During the last years, the amount of research on the effectiveness of E-health interventions increased substantially (Kreps & Neuhauser, 2010). Although the overall effect of E-health is positive on a person his or her health, there are inconsistencies in results of different research publications (See table 3.2.2). For example, in a review of twelve randomized controlled trials (RTCs) and quasi-experimental designs of a computer-tailored dietary intervention, seven out of twelve studies showed positive results on dietary behavior (Neville, Ohara, & Milat, 2009). Next to this, a review of different RTCs on dietary and physical behavior showed that twenty out of twenty-six studies had positive results (Kroeze, Werkman, & Brug, 2006). However, only three of the eleven physical activity RCTs were successful. Moreover, the study of Vandelanotte, Spathonis, Eakin, & Owen (2007) reviewed RCTs and quasi-experimental studies of website-delivered physical activity interventions and found that in eight of the fifteen studies improvements of physical activity was proven. However, these positive results diminished after a period of six months.

The previous summary of several RCT and quasi-experimental studies on E-health interventions show that several factors are likely to increase the success of an intervention (See table 3.2.2). These success factors are tailoring to the needs of the consumer, the use of theory, computer-delivery and high intervention intensity. Next to this, the article of Kreps & Neuhauser (2010) propose four suggestions to develop effective E-health interventions. First, E-health interventions should be designed in a way that the interaction with users and their active involvement in health promotion is maximized (Kreps & Neuhauser, 2010).

(17)

Table 3.2.2 Articles on the effectiveness of E-health interventions

When users actively participate with the intervention, they are more likely to internalize the intended message of the intervention (Cassell, Jackson, & Cheuvront, 1998). Therefore, one of the advantages of E-health interventions is the ease to integrate feedback mechanisms to promote this interaction (Kreps & Neuhauser, 2010). An example of a feedback mechanism is when an employee carries out an exercise routine and the intervention gives feedback on posture, speed and endurance. Next to this, questions and Answers (Q&A’s), the use of avatars, live-chat and customer support can increase the personal interactivity of health interventions (Kreps & Neuhauser, 2010). Second, E-health interventions should be designed to work effectively and transparently across different settings while appealing to a wide range of users (Kreps & Neuhauser, 2010). This criterion is especially important for national interventions or institutionalized (e.g. hospital) interventions. Therefore, this criterion will not be discussed in detail. Third, E-health interventions should be designed to personally engage users to maximize message exposure and influence (Kreps & Neuhauser, 2010). Interventions should match the unique characteristics, interests and cultural orientations of the target audience. Hence, effective interventions should be strategically designed for clearly segmented, homogeneous groups of people (Kreps, 2008). For example, interventions targeting young and active people should have

(18)

different characteristics than interventions targeting older and less active people. Fourth, E-health interventions should both use mass information as interpersonal information to be able to make changes on individual, institutional and social levels (Hornik, 2002; Napoli, 2001). Hence, the advantages of mass information (relative low costs, broad reach) should be complemented by the advantages of interpersonal information (personally engaging, effective in changing behavior). An example of a health information system that combines these two forms of information is a website through which users can ask questions and get personal answers, while the information on the website is intended for the mass of users.

3.2.3 Types of E-health solutions

There are several types of E-health solutions for consumers available. Examples are health information, computer-assisted learning, blogs and social media, biometric assessment and monitoring and exergames (Neuhauser & Kreps, 2010). Each of these examples will be discussed in the following paragraphs.

Health information on the internet

There is an increasing amount of health information available to make people actively involved in their health care (Ferguson, 2002). This information is offered by different stakeholders like insurance companies, physicians, hospitals, patient associations and patient support groups (Rozenkranz, Eckhardt, & Rosenkranz, 2013). Next to the health information offered by these mentioned professional instances, information is also widely shared by (non-professional) individuals. This health information is often shared through blogs and social media.

Computer-assisted learning (E-learning)

E-learning is defined as “the use of interactive electronic media to facilitate teaching and learning on a range of issues, including health issues” (Edwards et al., 2010 p.2). Distinctive advantages of E-learning are the possibilities to 1) tailor E-learning programs to individual circumstances (Edwards et al., 2010) 2) translate complex information into video, graphics and audio fragments (Murray, 2008) 3) cut costs because there are no face-to-face meetings with health practitioners (Edwards et al., 2010). E-Learning interventions have been divided into three generations: first generation interventions using computers to tailor print materials; second generation interventions using interactive technology on computers and third generation interventions using portable devices, such as mobile phones, for direct interaction and feedback (Neville, O’Hara, & Milat, 2009).

(19)

Blogs and social media

Blogs are defined as “personal web sites with content displayed in reverse-chronological order” (Adams, 2010 p.e90). Blogs offer people the opportunity to manage and share their own related information and experiences with others. Blogs can implicitly or explicitly mention health-related information (Adams, 2010). Implicit health blogs mention health-health-related issues and information as part of the blogger its daily life. Explicit health blogs are set up with a particular health situation in mind, for example coping with diabetes.

Social media could be seen as a channel for social support and as a channel that brings a sense of connectedness among people (Korda & Itani, 2013). These media let users share consumer-centric and consumer-controlled information enabling some kind of anonymity or personal connection as preferred. Next to this, social media is an inexpensive way to reach a large audience. Therefore, social media provides an opportunity for health promotion. Social network sites are one of the most popular forms of social media like Facebook and Twitter and health-specific network sites that focus on health conditions and services (Korda & Itani, 2013). These sites encourage people to ask questions and to enter discussions with other people about their health.

Biometric assessment and monitoring

The US Centers for Disease Control and Prevention defines biometric screening as “the measurement of physical characteristics such as height, weight, BMI, blood pressure, blood cholesterol, blood glucose, and aerobic fitness that can be taken at the worksite and used as part of a workplace health assessment to benchmark and evaluate changes in employee health status over time” (American College of Occupational and Environmental Medicine, 2013 p.244). When biometric screenings are included in WHPPs, their goal is to reduce employees their health risk, improve their health status, reduce health care costs, and improve the productivity and performance of the entire workforce (ACOEM, 2013). Biometric screenings usually consists of three parts (ACOEM, 2013). The first part is a health risk assessment, which refers to a questionnaire of self-reported life style. The second part of the screening consists of biometric measurements such as height, weight and BMI. The third part of the screening encompasses blood testing to get data about the cholesterol- and glucose level.

(20)

Exergames

Exergame is a term that is derived from the combination of exercise and digital gaming. This term describes digital games that ought to promote caloric expenditure and elevate heart rate through exercise (Staiano & Calvert, 2011). These exergames could support the health status of people; establish health-related behavior; increase awareness of healthy living habits and support the awareness of people to take good care of themselves at older age (Brauner et al., 2013).The great influence of these games is explained by behavior-related concepts like self-control, self-efficacy and motivation.

The article of Gerling, Schild, & Masuch (2010) has summarized important factors that could increase the value of exergames for older people: 1) adjustable game speed and difficulty, and an accessible interface (Weisman, 1983; Whitcomb, 1990) 2) visually adjustable (size of fonts, windows, colors and contrast), multimodal feedback and multichannel delivery (Ijsselsteijn et al., 2007) 3) appropriate cognitive challenges, a simple user interface and motivational feedback (Flores et al., 2008).

This part of the chapter has elaborated on the use of the internet for health promotion, E-health and the different types of E-E-health interventions. Theories of human behavior could improve the understanding of the possible impact of these E-health interventions. Therefore, the next part of the chapter will elaborate on several human behavior theories.

3.3 Human behavior theories

This part of the chapter will discuss five human behavior theories. Each theory has a slightly different perspective on what drives human behavior. These theories could provide more insights into the opportunities for preventive E-health interventions to change the behavior of older employees towards making healthier choices.

3.3.1 Social cognitive theory

A first model that tries to explain human behavior is called social cognitive theory (SCT). Bandura (1986) argues that behavior is influenced by personal, behavioral and environmental factors. First, personal factors that influence behavior are knowledge, perceived self-efficacy and outcome expectations. Second, behavioral factors are plans and goals an individual develops to carry out a particular behavior at a future point in time. Third, environmental factors encompass social support and potential barriers in the person his or her environment. All these factors have an impact on future behavior change (See figure 3.3.1). However, this theory has its main focus on the construct of perceived self-efficacy. Bandura (1977) argues that the more efficacious persons are about their goals, plans or behaviors, the more likely they will act on it. Next to this,

(21)

perceived self-efficacy has an impact on all other factors, for example environmental barriers can be overruled because a person has strong perceived self-efficacy. The formulation of this construct has had a large impact on research on consumer behavior and has influenced a lot of subsequent theories. For example, Ajzen (1991) states that his term perceived behavioral control, in the theory of planned behavior, is derived from the concept of Bandura’s perceived self-efficacy.

3.3.2 Theory of planned behavior

A second model that tries to explain human behavior is the already mentioned theory of planned behavior (TPB) (See figure 3.3.2). Ajzen (1991) argues that behavior depends on a person his or her intention to perform a particular behavior and its perceived behavioral control. The intention to perform a behavior consists of three factors (Ajzen, 1991). The first factor that influences the intention is the attitude towards the behavior which refers to the degree to which the person appraises the behavior positively or negatively. The second factor is the so called subjective norm and this presents the perceived social pressure to (not) perform a particular behavior. The third factor is the earlier mentioned perceived behavioral control; this construct refers to the degree that the behavior is easy or difficult to perform.

Figure 3.3.2 Elements of TPB (Ajzen, 1991) Figure 3.3.1 Elements of SCT (Bandura, 1986)

(22)

3.3.3 Health belief model

A third model, the health belief model (HBM) focuses specifically on health behavior. Gochman (1997, p.3) defines health behavior as “behavior patterns, actions and habits that relate to health maintenance, health restoration and to health improvement”. The HBM asserts that a health action occurs because of three factors: sufficient motivation to make the health issue relevant (health concern); the perception that one is vulnerable to a condition or illness (perceived threat) and the belief that following a recommended health action will result in a lower perceived threat at a reasonable cost (perceived benefits and barriers) (Becker, 1974; Rosenstock, Strecher, & Becker, 1988) (See figure 3.3.3). These three factors are influenced by both demographic (age, gender) as psychological (personality, group pressure) factors. The ultimate preventive action is thus influenced by those three factors besides the factor cues to action. Cues to action are strategies to increase the readiness of a person to perform the behavior. Examples of cues are reminders and how-to information. In 1988, Rosenstock, Strecher, & Becker added the factor perceived self-efficacy (See SCT) to the HBM.

3.3.4 Fogg behavioral model

A quite recent model that is argued to be very useful for designs of persuasive technologies is called the Fogg behavioral model (FBM) (Fogg, 2009). In this model, behavior is a result of three factors: motivation, ability and triggers (See figure 3.3.4). For a person to perform the target behavior, all three factors must be present. This model implies that the factors motivation and ability are often trade-offs that can be influenced by the design of an intervention. For example, people with low motivation could though perform a behavior because it is very simple (high on ability). However, even if ability or motivation (or both) are high, when a behavior is not triggered it will not be very likely to occur (Fogg, 2009).The factor triggers could be seen as

Figure 3.3.3 Elements of HBM (Becker, 1974)

(23)

derived from Becker’s (1974) term of cues to action in the HBM. Examples of triggers are text messages, alarms, or pop-ups that get people their attention to performing the target behavior when they are above the behavior activation threshold. This threshold consists of the two factors motivation and ability, and when this combination is high enough, a trigger will elicit the target behavior (Fogg, 2009).

Figure 3.3.4 Elements of FBM (Fogg, 2009)

3.3.5 Integrative model of behavioral prediction

In the article of Fishbein (2000), it is argued that although there are many theories of behavioral prediction, careful consideration of them concludes that only a limited number of variables have to be considered to predict human behavior (Fishbein & Cappella, 2006). Therefore, Fishbein proposes an integrative model of the HBM, SCT and theory of reasoned action (this last theory resembles closely to the TPB).

When looking at figure 3.3.5, it is shown that “any given behavior is most likely to occur if one has a strong intention to perform the behavior, if one has the necessary skills and abilities required to perform the behavior, and if there are no environmental constraints preventing behavioral performance” (Fishbein, 2000 p. 275). The model suggests that to form a strong intention, three factors should be considered. The first factor is the attitude towards performing the behavior; this can be a positive or negative evaluation (Fishbein, 2000). The second factor is the perceived norm and the motivation to comply with this norm (Fishbein, 2000). This factor includes both the perceptions of what others think one should do as well as perceptions of what others are actually doing. The third factor is a person his or her perceived self-efficacy with respect to performing the behavior (Fishbein, 2000). Next to this, the model integrates beliefs underlying these three factors and background variables that have an indirect impact on behavior (Fishbein, 2000).

(24)

Behavior change is accomplished by changing a person his or her behavioral skills, the environmental factors and/or the behavioral intention. Therefore, it is important to research for a certain behavior and population which of these factors are more important to be able to target these factors in a behavior-change intervention. This knowledge can be used to conceptualize an effective intervention to accomplish favorable behavior change.

This part of the chapter has summarized five human behavior theories, of which four described slightly different theories and one integrated them (partly) into one model. The integrative model of behavioral prediction will be used to derive suggestions for increasing the value of and therefore the compliance with the E-health intervention of interest. The relative importance of all the mentioned psychosocial variables in this model will depend upon the behavior and the population of consideration (Fishbein & Cappella, 2006). Therefore, the next part of the chapter will describe the characteristics of the population of interest, namely older employees.

Figure 3.3.5 Elements of the integrative model of behavioral prediction (Marco Yzer, 2008)

3.4 Older employees

This part of the chapter will elaborate on the characteristics of the population of interest, namely older employees. In 2060 about 30 percent of the population in the EU will be 65 or older (Eurostat, 2012). This growth in the amount of older people is partly the consequence of advances in medical care and improvements in health status (Davies, 2010). When looking at the EU, the life expectancy of both males and females after 65 is still slightly increasing every year. The indicator health life years (HLY) measures the number of years that a person is expected to live in a healthy condition (Eurostat, 2014). While the life expectancy of people at the age of 65 is

(25)

increasing, the HLY for people at 65 is slightly declining. Therefore, health interventions are of great importance, because these interventions would be able to improve the number of healthy years older people will live in the future.

3.4.1 Age-related diseases

There are five age-related diseases: ischemic heart disease, malignant neoplasms, diabetes, osteoarthritis and Alzheimer’s disease (Davies, 2010). These diseases are all somewhat influenced by the lifestyle of a person. Here, lifestyle is defined as the personal choices that influence health such as choices regarding nutrition and physical activity (Davies, 2010). Lifestyle can be positively influenced by interventions targeting healthier choices. The workplace could be one of the prominent places to implement such preventive health interventions to improve older employees their lifestyle and hence their health status.

3.4.2 The definitions of older employees and health promotion

The European Agency for Safety and Health at Work (2014) reports that the rate of employees aged 55–64 rose from 36.9% in 2000 to 46% in the EU in 2009. The term older employee has been defined in a variety of ways. First, the 1967 Age Discrimination in Employment Act considers older employees as 40 or older (Tikkanen & Nyhan, 2006). Second, statisticians often consider older employees between the age of 45 and 64. Third, according to the EU Lisbon benchmark, people between 55 and 64 are considered as older employees.

The article of Naumanen (2006) integrates what occupational health workers perceive as health promotion in the case of older employees. They see health promotion as “interventions that ensure better health, well-being and pleasure, adaptation to age-related changes, and the recognition of factors that promote health and prevent illness. It also means taking care of one’s own health by making use of counseling and advice. Health promotion for older employees is also understood as early prevention of the influence of health hazards” (Naumanen, 2006 p.41). Next to this, they give the advice that the needs of older employees should be taken in consideration with work arrangements. For example, to sacrifice some tasks, to shorten the working hours, to use special equipment or to make existing working conditions healthier. In this way, work tasks and conditions are balanced with the older employees their resources (Naumanen, 2006).

3.4.3 The natural ageing process

Because of the natural ageing process some functional capacities, mainly physical and sensory, decline with older employees (EASHW, 2014). The physical capacities are mainly declining

(26)

between the age of 50 and 60 years. However, this decline can be slowed down by regular exercise (Ilmarinen & Louhevaara, 2001; Ilmarinen, 2001; Lipow, 1997). The psychological and most intellectual functions tend to stay stable until the age of 75 (Ruoppila & Suutama, 1994). Next to this, older employees have more experience in social interactions with other people (Heikkinen, 1994). Therefore, the weakened physical capacities are compensated by enhanced social capacities (Wegman, 1999).

However, ageing processes are not clear cut and are dependent on the person his or her current lifestyle; nutrition; fitness; genetic predisposition to illness; educational level; work and other physical environment factors (EASHW, 2014). Therefore, there are considerably great differences between the rates of functional capacities declines in older people. Next to this, precedent health behavior has a great influence on employees their health status in later life (EASHW, 2014). A healthy lifestyle throughout a person his or her life can slow down and minimize the decline in functional capacities. The workplace can be a prominent environment to promote a healthy lifestyle and support activities that prevent the decline of functional capacities (EASHW, 2014). Therefore, WHPPs are of great importance to improve older employees their health status. Thus, although the employee his or her ability could decrease with age, interventions could usually accommodate for these declines in ability (Warr, 2001).

3.4.4 Age-related adjustments for E-health interventions

For an intervention to be suitable for older people, websites and intervention systems need to be designed according to their needs. There are several age-related changes important for the development and design of an E-health intervention (See table 3.4.4).

First, older people experience a decline in psychomotor and perceptual skills, especially in the dexterity and hand-eye coordination (Kaufman et al., 2006). Next to this, the amount of ‘noise’ in the motor system pathways result in disproportionate development of movement variability (Mead, Batsakes, Fisk, & Mykityshyn, 1999).

In addition, navigating through online health information places heavy demands on fluid cognitive abilities (Garfein, Schaie, & Willis, 1988). Fluid cognitive abilities are processes that allow us to learn and adapt in novel situations. This concept consist of the working memory (the ability to hold content in memory while giving attention to other things) and the spatial abilities (the ability to create and manipulate mental representations such as maps) (Pak, Czaja, Sharit, Rogers, & Fisk, 2006). Because both abilities decline with age, performance on tasks that rely heavily on these two concepts decreases.

(27)

Table 3.4.4 Age-related changes and subsequent adjustments for E-health interventions

Furthermore, the results of the research by Fisk & Rogers (2002) show that older people are limited in developing automated response. Consequently, actions that become highly automated for most adults may continue to exert a significant cognitive load for older people.

Another age-related change is the reduced ability of older people to focus on objects that are situated at short distance (Kaufman et al., 2006). This may limit their ability to clearly see stimuli as cursors and clickable objects on the computer screen. Next to this, older people are more affected by distracting context and this potentially limits their ability to selectively give attention to the relevant screen elements and to perform parallel tasks (Kaufman et al., 2006).

Finally, older people experience a loss in static and dynamic visual awareness, a loss of contrast sensitivity and a decline in color sensitivity (Ijsselsteijn et al., 2007). Such visual impediments make it harder for older people to perceive small elements on a screen, to read small instructions or to locate information on complex screens.

Adjustments to E-health interventions can partially accommodate for the earlier mentioned age-related changes (See table 3.4.4). This includes changing the basic elements of the user interface such as the size of buttons and certain widgets (e.g., links, scrollbars, buttons, menus) (Kaufman et al., 2006). Next to this, it is necessary that the fonts are clearly readable and

(28)

that navigation elements are not too closely spaced (Kaufman et al., 2006); Pak, Price, & Thatcher, 2009). Furthermore, to accommodate for possible visual impediments it is necessary to use relatively high-contrast images (Brauner et al., 2013). In addition, the research of Pak & Price (2008) showed that when websites are designed around key words or tags, instead of in a hierarchy of folders, older people could better find information online. This result is explained by the notion that people their knowledge of vocabulary increases with age and the tag-based interface reduces the demand on spatial abilities.

This part of the chapter has highlighted the importance of WHPPs for older employees. Next to this, several important adjustments for E-health interventions are proposed to make them suitable for this target group. The next part of the chapter will summarize this literature review by means of one derived model that will be used to describe health behavior of older employees.

3.5 Integration of the literature: the derived model

This part of the chapter will describe the model that integrates the previous discussed literature. This model intends to describe the factors that influence health behavior of older employees.

The integrative model of behavioral prediction by Fishbein (2000) explains that behavior change is accomplished by changing a person his or her behavioral skills, the environmental factors and/or the behavioral intention. Therefore, it is important to research for a certain behavior and population which of these factors are more important to be able to target these factors in a behavior-change intervention. In this research the behavior of interest is the use of the E-health intervention and the population consists of older employees.

The model of Fishbein (2000) asserts that a given behavior may not be performed because either a person has formed an intention to perform the behavior but is unable to act upon it or the person has little or no intention to perform the behavior (Fishbein & Cappella, 2006). These two different situations require different types of interventions. Table 3.5 shows the 2X2 matrix consisting of the different options and the implications for interventions. If strong intentions to engage in health-protective behavior are not formed, the intervention should focus on changing attitudes, perceived norms and/or self-efficacy to form an intention to perform the behavior (Fishbein & Cappella, 2006). However, if one has formed an intention but is unable to act upon it because of a lack of skills or the existence of environmental barriers, the intervention should focus at skill building or removing/helping people to overcome environmental barriers (Fishbein & Cappella, 2006).

(29)

Table 3.5 2X2 matrix intention-behavior relationship and implications for interventions (Fishbein & Cappella, 2006)

Because the timeframe of this thesis, investigating the motives of each participant would be too time consuming. Therefore, it is assumed that the research population will consist of older employees that have formed an intention to perform the recommend health behavior, but are unable to act on it. The majority of people is great at forming positive intentions to eat healthier or to exercise more, however most people fail at turning this intention into real behavior. The research of Sheeran (2002) justifies this assumption. Namely, his research shows that intention only predicts 28% of the variance in behavior. Thus, forming an intention is often not enough to actually perform the behavior. This gap between intention and behavior is largely caused by inclined abstainers. Inclined abstainers are people who have formed positive intentions (e.g. I will exercise three times a week) but fail to act on these intentions (Sheeran, 2002). Next, the line of reasoning of Fishbein (2000) is followed, by saying that this gap could be closed by building the required skills and removing/helping people to overcome environmental barriers. For the constructs skills and environmental barriers the proxy perceived behavioral control (PBC) will be used. PBC is the person his or her belief that the behavior is easy or difficult to perform (Ajzen & Madden, 1986). The more resources and opportunities individuals think they possess, and the fewer obstacles or impediments they anticipate, the greater their PBC.

This research will also integrate the background, demographic variable age. The integrative model of Fishbein (2000) shows also background variables that play primarily an indirect role in influencing behavior. When such demographic, socio-economic, or individual difference variables are related to the underlying beliefs, they are likely to be related to the behavior in question. In this research the demographic variable age is assumed to be related to the behavior because, as mentioned before, age has an impact on the health behavior of people. The

(30)

variable age will be integrated by using the mentioned E-health adjustments which accommodate for age-related changes (See chapter 3.4.4).

This research will look if the variables age-related adjustments (ARE) and perceived behavioral control (PBC) can turn the intention to use the E-health intervention into real use of the intervention (See figure 3.5). This part of the chapter has partly integrated the previous mentioned literature into one model. The next part of the chapter will discuss the hypotheses formed in this research.

Figure 3.5 Derived model, based on Fishbein (2000) and literature on age-related changes

3.6 Hypotheses

This part of the chapter will discuss the main- and sub hypotheses that were formed in this research. The main hypothesis stated that:

1) The more variables of the derived model are included in the E-health intervention, the higher the WTP of the older employees will be.

This hypothesis was based on the assumption that the derived model is able to create more value for older employees than the value of an E-health intervention that does not communicate PBC and has no ARA. Consequently, their WTP will be higher for the E-health intervention based on the derived model. In this research it was assumed that when people are willing to pay more for an intervention, because it delivers them more value, the compliance rate of this intervention will go up. This assumption will be further explained in paragraph 3.7.2. Hence, it was hypothesized that in condition four (in table 3.6) the highest WTP is reached and therefore the expected compliance rate of the E-health intervention would be the highest of the four conditions.

(31)

The two sub hypotheses of the main hypothesis were:

Hypothesis 1A- When the variable PBC is communicated in the E-health intervention, the WTP of the older employees will be higher than when this variable is not communicated.

This hypothesis was based on the assumption that the communication of PBC in the intervention could increase the perceived ownership of skills and non-existence of environmental barriers. Consequently, older employees would have a greater perceived control over the behavior and therefore would be more likely to use the E-health intervention. This results in a higher WTP and therefore compliance rate.

Hypothesis 1B- When the demographic variable ARA is integrated into the E-health intervention, the WTP of the older employees will be higher than when this variable is not integrated.

This hypothesis was based on the assumption that the integration of ARA into the intervention increases older employees their confidence and motivation to use the intervention because the intervention is adapted to their capacities and needs. Consequently, older employees will be more likely to use the E-health intervention which results in a higher WTP and therefore compliance rate.

This part of the chapter has discussed the hypotheses formed in this research. The next part of the chapter will elaborate on two other important topics, namely compliance and willingness-to-pay.

(32)

3.7 Compliance and willingness-to-pay

This part of the chapter will first introduce the topic of compliance with an intervention. Second, the concept of willingness-to-pay will be explained.

3.7.1 Compliance

Compliance is defined as “the conformation or adaptation to another person his or her wishes, to a rule or to a necessity” (Bowman, Heilman, & Seetharaman, 2004 p.324). In this research, compliance will be with a necessity, namely with the E-health intervention of interest. To be able to form suggestions to increase the compliance with this intervention, the proxy willingness-to-pay will be used.

3.7.2 Willingness-to-pay

In this research there will be a closer look to the factors that increase compliance with an E-health intervention; especially to how this intervention could create more value for older employees.

Recent literature on economic evaluation in health and health care has shown increasing interest in the use of WTP as a measure of health benefits (Olsen & Smith, 2001). This measure is somewhat replacing the measure of Quality-adjusted-life-years (QALYs) that was often used to indicate the outcomes of a health intervention. The measure QALYs expresses only the improvement in health status. At contrary, WTP can express a total health gain profile which includes more than health status.

Brandenburger & Stuart (1996) argue that to obtain a profit an organization must have favorable asymmetries between the organization itself and other organizations. One value-based strategy that is appropriate for achieving this, is to get a higher willingness to pay (WTP) for their product. The WTP is the amount of money a customer is willing to spend on a particular product or service (Brandenburger & Stuart, 1996). A higher WTP is reached when an organization differentiates itself from other organizations by meeting the customer needs in a better manner. This is also known as the differentiation strategy of Porter (Porter, 1980). The concept of WTP is based on welfare economic theory which states that “the benefit to an individual of a service or an intervention is defined as that individual’s maximum willingness to pay for this service or intervention” (Bala et al., 1999 p.10). To illustrate the concept of WTP for an E- health intervention, it must be assumed that well-being (also known as utility) depends on a person his or her income and health. An intervention is then introduced that moves the person his or her health status from a specific disease state (or less healthy state) (HD) to full health (H*) (See graph 3.6.2). The WTP is the maximum amount of money the person would pay for the intervention that restores to full health while maintaining the same level of overall well-being or utility (Bala et al.,

(33)

1999). If the person had to pay more than this WTP, the loss of income would offset the gain in well-being due to the health change. To estimate a person his or her WTP for an intervention, we start with a person in health state HD with an income of Y0 and a well-being of U*. Next, the lower income (Y1) in full health (H*) is determined that results in the same well-being level (U*) as income Y0 in health state HD. The difference between Y0 and Y1 is the WTP of the person.

Graph 3.6.2 WTP for a health intervention (Bala et al., 1999)

In this research the E-health intervention is assumed to move an older employee from a less healthy state to full health. The WTP of the older employees is the amount they are willing to pay for the intervention to restore their full health while maintaining the same level of utility. The second assumption in this research is that WTP is a reflection of the probability the older employee will acquire and use the E-health intervention. If the person is willing to pay a very small/large amount for the intervention, the probability that this person will acquire and use the intervention will decrease/increase. The third subsequent assumption is that therefore WTP is a good proxy for the compliance rate of an E-health intervention. When the person has a low/high WTP for the intervention, he or she will be more/less likely to acquire and use the intervention and therefore the intervention will have a higher/lower compliance rate.

Concluding, in this research WTP will be used as a measure to be able to form suggestions for increasing the compliance rate of the E-health intervention of interest. This part of the chapter has introduced the concepts of compliance and WTP. The next chapter will cover the methodology.

(34)

4. Methodology

This chapter will describe the research methodology. An experimental, 2x2 factorial design was used to test the hypothesis. The independent variables were the communication of perceived behavioral control (PBC) and the integration of age-related adjustments (ARA). Respective levels of independent variables were communication or no communication of PBC, and integration or no integration of ARA. The dependent variable was the WTP of older employees.

This chapter will start with an explanation on the research design. Second, there will be an elaboration on the target sample. Third, the measures in this research will be explained. Fourth, the method of data collection will be discussed. Fifth, the procedures will be explained. Finally, there will be an explanation on how the data will be analyzed.

4.1 Research design

This part of the chapter will discuss the research design and process used in this thesis. The main research question of this research was How to increase compliance of older employees with an

E-health intervention in order to change their E-health behavior? To be able to answer this question

the research process displayed in figure 4.1 was used. An experiment is a research strategy which looks at causal links between variables, to see whether a change in one independent variable causes a change in another dependent variable (Saunders & Lewis, 2012). The form of experiment was a factorial vignette study. With a factorial vignette study, participants are presented questions based on descriptions of a hypothetical situation (Hox, Kreft, & Hermkens, 1991). This method has multiple advantages because it can prevent social desirable answers (Alexander & Becker, 1978), can enhance the involvement of a participant (Ruzicka, 2013), and can provide standardized stimuli to all respondents by which the internal validity, measurement reliability and the ease of replication is enhanced (Ruzicka, 2013). This part of the chapter has elaborated on the research design. The next part of the chapter will discuss the sample of this research.

(35)

Figure 4.1 Different steps in this research

4.2 Sample

This part of the chapter will discuss the sample of this research. The sample of this research consisted of older employees between the age of 45 and 65 with sedentary jobs in the Netherlands. Self-selection sampling was used as a form of non-probability sampling. The participants were invited to fill in the survey at Facebook, LinkedIn and at four different organizations: a consultancy firm, two health insurance companies and one bank. All participants have sedentary functions within the organization. The participants were randomly assigned to each of the conditions. Each condition was targeted at 50 older employees. Because there are four conditions, the total targeted sample consisted of 200 participants. This part of the chapter has discussed the sample. The next part of the chapter will elaborate on the used measures in this research.

4.3 Measures

This part of the chapter will discuss the measures used in this research. The experiment that was performed contained two independent variables. The first independent variable was the communication of PBC. PBC is the person his or her belief that the behavior is easy or difficult to perform (Ajzen & Madden, 1986). To specify this variable to the behavior of interest an elicitation study was performed. The behavior of interest is the use of an online health intervention at a person his or her workplace. The elements of the elicitation study were based on the construction of TPB questionnaires in the articles of Ajzen & Madden (1986) and Francis et al. (2004). 20 persons in the researcher its direct environment were selected for this elicitation study. These persons shared the characteristics (age 45-65 and with sedentary jobs) of the target sample. Their answers to the survey were grouped into five main associations and these associations were used in the vignette to communicate PBC.

The following questions were asked in the survey:

(36)

• Which factors or circumstances enable you to use this online health program at your work?

• Which factors or circumstances make it difficult or impossible for you to use this online health program at your work?

• Are there any other issues that come to mind when you think about using this online health program at your work?

The complete questionnaire for this elicitation study can be found in appendix A. The results of the elicitation study showed that the five most important PBC variables were:

• A simple program design, which means that it must be easy to find and follow the different components of the system.

• Possibilities at work to execute the program and exercises • Time must be (made) available to follow the program

• The program must not decrease employees their focus on their current tasks • The program must be fun to follow

Subsequently, these five variables were used to communicate the PBC variable in the vignette of the experiment.

The second independent variable was the integration of ARA in the E-health intervention. These adjustments were based on literature regarding age-related changes and resulting implications for E-health interventions (See chapter 3.4.4). Five implications for E-health interventions were included in this research:

• Increase the size of buttons and widgets • Establish clearly readable fonts

• Establish enough space between navigation elements • Use high-contrast images

• Use keywords or tags

The dependent variable was the willingness-to-pay (WTP) of older employees for the E-health intervention. The WTP is the amount of money a customer is willing to spend on a particular product or service (Brandenburger, Stuart, 1996). This vignette study had a 2X2 factorial design, (See table 4.3(.1)). The variables and the constants in this research are displayed

(37)

in table 4.3(.2). This part of the chapter has discussed the measures. The next part of the chapter will discuss the manner of data collection.

Table 4.3(.1) 2x2 factorial design of this thesis

Table 4.3(.2) Independent variables, constants and dependent variable in this research

4.4 Data collection

This part of the chapter will discuss the manner of data collection. The data was collected by means of an online survey (See appendix B), created in the online program Qualtrics. The survey

The effect of the communication of perceived behavioral control and the integration of age-related adjustments on the compliance rate of a preventive E-health intervention for older employees