Development of a new model to measure perceived service quality in Dutch inpatient hospital care

(1)

Perceived Service Quality in Dutch

Hospital Care

Development of a new model to measure perceived service quality in Dutch inpatient hospital care

(2)

(3)

Perceived Service Quality in Dutch Hospital Care

Development of a new model to measure perceived service quality in Dutch inpatient hospital care

Rijksuniversiteit Groningen

Faculty of Economics and Business Master thesis

Date: 20-06-2011

Name: Pieter Meijer

Address: Middelweg 5 Postal code: 9451 HE Residence: Rolde Phone number: 0624629131 E-mail: pietermeijer@home.nl Student number: S1340905

(4)

Abstract

Abstract

Due to the 2006 health care reform the Dutch health care sector is becoming a more and more competitive environment in which health care organizations realize the need to focus on service quality perceived by its consumers to improve their service offering.

To make this task even more complicated consumers nowadays are more aware of alternative offers and become increasingly knowledgeable and demanding, not least because of the advent of the Internet (Lim & Tang, 2000; Gill & White, 2006).

For instance, in the Netherlands there are numerous hospital comparison websites such as: independer.nl, zichtbarezorg.nl, ziekenhuizentransparant.nl, dr-yep.nl, and

zorgmarktnederland.nl.

In recent years many models to measure hospital quality have been proposed, but to date there is still no well accepted generic model that is able to completely capture the hospital care service domain. There is still a large variety of approaches regarding the instrument domain, how it is measured, and when perceptions of care are elicited (Castle, Brown, Hepner, & Hays, 2005). Therefore the need for a standardized model to capture the perception of hospitalized patients would be very beneficiary to identify service shortfalls and improve perceived quality. In this thesis we come up with such a new model to measure perceived hospital quality of hospitalized patients. We choose to measure the construct perceived service quality (instead of satisfaction), as we focus on the appropriate items and dimensions and do not want our model to be influenced by cognitive and affective processes.

In addition we wanted our model to have a solid theoretical basis so we analyzed, and

compared, a large number of service quality measurement models. Based on these insights we developed a multidimensional second order formative SERVPERF (service performance) model to measure perceived hospital quality.

In order to create a highly efficient model without making too many concessions to its

effectiveness we selected the most important dimensions and corresponding items patients use in assessing hospital quality. Based on extensive literature review we constructed a model containing 24 items dived between 5 dimensions. Most of the dimensions and items stem from SERVQUAL/ SERVPERF research and the Dutch CQI questionnaire. The dimensions and items found are confirmed by means of a small qualitative study. The dimensions used in our model are: Physical environment, Reliability & Safety, Speed & Assistance, Interaction & Communication, and Outcome. Health care managers and marketers should focus on these dimensions to improve their service offering.

The dimensions and items were analyzed using data from 101 respondents of a convenience sample, which is useful to illustrate the functioning of the model and demonstrate its practical applications. Empirical results from our sample show that the dimensions Outcome, and Reliability & Safety have the greatest significant correlation with perceived hospital quality, implying a high relative importance of these dimensions

(5)

Abstract

We also checked our model for direct and indirect effects of three patient characteristics: age, gender, and self-perceived health status. We found a direct effect between groups

divided by the self-perceived health status on perceived hospital quality. We established that the means vary between respondents with a very good self perceived health status and a good health status, and respondents with a very good self perceived health status and a fair health status. We also found an indirect (moderating effect) of age between outcome-

perceived hospital quality and interaction & communication-perceived hospital quality. Despite the found direct and indirect effect we conclude that based on our convenience sample patient characteristics have minimal predictive power.

Finally we checked whether we could use a single number to express hospital quality to facilitate between hospital comparison. We established a positive and significant relation between perceived hospital quality and the Net promoter Score developed by Reicheld (2003). So, we can use the score on perceived hospital quality as a single number to effectively measure perceived hospital quality, but we can also use the NPS to measure consumer loyalty.

Important suggestions for further research would be to test whether the assumed relationships and correlations in our model are significant for the population by performing a probability sampling technique, and to see if we created a model which is generic for all Dutch hospitals and different hospital wards.

(6)

Preface

Preface

I wrote this thesis as the conclusion of my master degree in Business Administration at the University of Groningen.

When I started this thesis I did not have a clear subject but I knew it should have something to do with health care and marketing.

Health care because I have a job in mental health care, and marketing because it has been my primary field of interest during my study.

When visiting my supervisor Mr Alsem for the first time back in April 2010 we discussed several options for a possible thesis in which health care and marketing are related.

Options were numerous because of the Dutch health care reform in 2006 which has enabled marketing as an important instrument to survive the new competitive market.

After some constructive discussions we decided to focus on the dimensions and items which consumers use to evaluate a health care organizations. Later I further delineated the subject to inpatient consumersof Dutch hospitals.

I have experienced writing my thesis as a challenging final chapter of my study. It was not always easy to filter the overload on information and to make sense of the contradicting opinions on service quality in health care. But in the end I think I managed to deliver a thesis which satisfies the prerequisites and is also understandable to the reader. While writing this thesis I received a lot of support and I would like to take this opportunity to thank the following people. First and foremost, I would first like to thank my supervisor Mr. Alsem for his very useful comments and excellent guidance throughout the process. Secondly, I want to thank my second supervisor Mr.Berger for his advice and precious time, and finally I want to thank my wife Willemijn for her patience and continued support.

(7)

Contents Abstract Preface 1 Introduction 8 1.1 Background problem 8

1.2 Problem statement and research questions 10

1.3 Theoretical and social relevance 11

1.4 Structure of the thesis 11

2 Service quality 12

2.1 Service quality: what is it? 12

2.2 Service quality in health care 13

2.3 The relationship between service quality, satisfaction, 15 service value, and behavioral intentions

2.4 Service quality: how to measure it? 19

3 SERVQUAL and SERVPERF 25

3.1 SERVQUAL: past and present 25

3.2 SERVPERF 28

3.3 SERVQUAL & SERVPERF and modifications for health care 30

4 Other important models for measuring service quality 31

4.1 The Nordic perspective 31

4.2 The Hierarchical perspective 32

4.3 Continuous Quality Improvement (CQI) 34

5 Dimensions and items 37

5.1 Health care dimensions and items 37

5.2 Health care dimensions and items from 38

SERVQUAL and SERVPERF research

5.3 Dimensions and items in other research 42

6 Conceptual model 44

6.1 Building blocks of the conceptual model 44

6.2 The conceptual model 60

7 Research design 62

7.1 Research method 62

7.2 The sample 64

7.3 Plan of analysis 65

8 Results: analysis and discussion of empirical data 66

8.1 Characteristics of the sample 66

8.2 Descriptives 68

8.3 Explorative results 72

(8)

Contents

9 Conclusions 81

9.1 Main conclusions and implications 81

9.2 Limitations and suggestions for further research 84

Literature list 85

Appendix A Dissimilarities between health care and other services 98

Appendix B Linkage of the service quality models 99

Appendix C Decision Rules to identify constructs as formative or reflective 100

Appendix D Conceptual model of service quality 102

Appendix E CQI questionnaire items 103

Appendix F Items from the perception scale of Babakus & Mangold 104

Appendix G Questionnaire in Dutch 105

Appendix H Questionnaire 110

(9)

Chapter 1: Introduction

8

Chapter 1: Introduction

1.1 Background problem

Health care reform

In 2006 the Dutch health care system was radically reformed.

Before the reform Dutch health care insurance was based on a two-pillar system. One pillar consisted of the social health insurance system for people in the lower income brackets, and the other pillar was the voluntary private health insurance system for people with higher incomes (Douven, Mot, & Pomp, 1997). Due to a rapid aging of the Dutch population, a low birth rate, and a lack of efficiency, the Dutch health care market needed a reform to cut costs and increase quality. The new Dutch health care system is predominately based on Alain Enthoven's model of national health insurance based on managed competition in the private sector (Enthoven, 1978). This model is defined as a purchasing strategy to obtain maximum value for

consumers and employers using rules for competition derived from microeconomic principles (Enthoven, 1993). But despite the reform the Dutch health care market is still heavily regulated. For example, entry barriers still exist and public health care organizations are not allowed to make a profit (Leeuwen, 2007). This regulation is very important for the protection of public interests in the Dutch health care market such as: quality, affordability, availability, and solidarity (Atos Consulting Trends Institute, 2010). It ensures that every resident is compelled to have a health insurance, and that every insurer is compelled to accept every consumer for the basic care insurance. Another big impact on Dutch health care is the rise of the internet which makes consumers more demanding and better informed. For example consumer are able to express opinions, and compare between hospitals, on websites like independer.nl, and kiesbeter.nl.

Dutch health care market

The Dutch health care market consists of several submarkets including hospital care. Almost every (sub)market in Dutch health care has the players shown in figure 1.

Figure 1: Actors in a Dutch health care market

Source: Getzen, 2007

Health Insurers

Care Providers

Care delivery market

Population

(10)

9 Hospital care

Hospital care is the largest submarket in Dutch health care and is responsible for more than 25% of total health care expenditure in 2009 (CBS, 2010). Dutch Hospital care consists of polyclinic care, one-day care surgery, and inpatient care. In the Netherlands there are 137 hospital locations and 83 so called outside polyclinics (RIVM, 2010). In outside polyclinics consumers can obtain the same health care services as a polyclinic located in a hospital. The clinics are closely related to a general hospital but not on the same location. The main purpose of these clinics is to reduce waiting lists related to the rapid growth of polyclinic visits over the past few years. The number of these outside polyclinics has increased from 60 in 2009 to 83 in 2010, while the number of hospitals has remained stable.

Focus on inpatient hospital care

In this thesis we focus on inpatient care which can be described as care of patients whose condition requires hospitalization. In addition we mention that hospitalization requires (at least) one overnight stay. In 2009 almost 6,7% of the Dutch population (1.1 million people) was hospitalized for at least one day (CBS, 2009). Inpatient care in a hospital is a widely investigated subject in scientific research predominantly in relation to quality and satisfaction. We choose to focus on inpatient care as it is the most comprehensive form of (hospital) care. Quality in health care

Since 2006 the Dutch government intends to relax regulatory constraints gradually and to turn more and more of the care over to the competitive market (Enthoven & van de Ven, 1997). As competition becomes more intense and environmental factors become more hostile the concern for service quality grows. Therefore many of the successful health care providers of the next decade will position themselves as „high quality‟ (Babakus & Mangold, 1992) According to Jacobson & Aaker (1987), Buzell & Gale (1987), and Rudie & Wansley (1985) high quality service is favored in the marketplace and produces measurable benefits in profit, cost savings, and market share (Parasuraman, Berry, & Zeithaml, 1991). But what is „high quality‟ in a service industry and how to measure it? These two questions are central to this thesis in which our focus is on the measurement of perceived service quality in Dutch inpatient hospital care. That is, perceived service quality by the consumer.

(11)

10

Not least due to the fact that patients (consumers) themselves have fairly inchoate and pliable understandings of what quality means (Sofaer & Firminger, 2005). To address this issue it is important to understand which dimensions and items are important for the consumer in evaluating a health care service. These dimensions and items will play a significant role in this thesis as they serve as quality indicators of an inpatient hospital care provider. We will study the accepted models for measuring service quality and identify the dimensions and items important to the consumers when evaluating a health care service with the aim to construct a new instrument that measures perceived consumer quality in Dutch inpatient hospital care. The need for a new model is eminent as there is a large variety of approaches regarding the instrument domain, how they are measured, and when perceptions of care are elicited (Castle, Brown, Hepner, & Hays, 2005). A standardized model to capture the perception of hospitalized patients would therefore be very beneficiary to identify service shortfalls and improve perceived quality. The urge to improve perceived service quality is important as quality deficiencies are still distressingly common among inpatients (Weingart, et al., 2006).The potential benefits of increased perceived quality for the health care

organization evolve around having more satisfied patients which may lead to: more loyal consumers, less staff turnover, and increased efficiency (Clemens, Ozanne, & Laurensen, 2001). Benefits for health care patients include a more efficient and effective service (Clemens, Ozanne, & Laurensen, 2001).

1.2 Problem Statement and research questions

- Management problem

Service quality in Dutch hospital care is becoming more and more important. Hospital care organizations have to focus on quality perceived by its consumers to make sure they survive the new and more hostile environment with better informed and more demanding consumers. Therefore a service quality model is needed to expose the items and dimensions that define quality in the eyes of the patient. As current models are not efficient, do not allow easy hospital comparison, and often lack a sound theoretical basis there is a growing need for a new efficient and effective model to measure hospital quality which allows for easy

comparison between hospitals.

 Marketing Research problem

How can we construct a new model with a sound theoretical basis which measures perceived hospital service quality efficiently and effectively?

 Research questions:

1) Can we use a single number to express perceived hospital quality to facilitate hospital comparison?

2) Which theoretical assumptions should we use to construct our model to measure perceived service quality in health care effectively and efficiently?

3) What items and dimensions should be included in our model to effectively and

efficiently measure perceived hospital quality of consumers in Dutch inpatient hospital care?

(12)

11

1.3 Theoretical and social relevance

From a theoretical and scientific point of view our research is interesting for three reasons. First, we construct a new model based on existing scientific models and theories to measure perceived quality in Dutch hospital care. Secondly, there has been conducted little research to explore in a grounded manner how patients define and perceive the quality of their health care (Sofaer & Firminger, 2005). Thirdly, health care research into patients‟ perceptions regarding the dimensions of perceived service quality is limited (Clemens, Ozanne, & Laurensen, 2001). From a social point of view the research contributes to a better understanding by hospital management concerning the perception and evaluation of quality factors in inpatient hospital care by consumers. This allows organizations to serve their consumer better to create superior value to the consumer: namely, high quality!

1.4 Structure of the thesis

In this chapter we described the background problem at hand, and presented our management and research problem. In the second chapter we focus on service quality and give an

(13)

Chapter 2: Service quality

12

Chapter 2: Service Quality

2.1 Service Quality: what is it?

„ A higher quality of care!‟

This was mentioned by Dutch patients as the most important improvement in health care regarding the future (Newcom Research & Consultancy, 2009).

Evidence in both manufacturing and service industries indicate that quality is a key

determinant of market share and return on investment as well as cost reduction (Parasuraman, Zeithaml, & Berry, 1985). Lewis (1993) stated that although establishing service quality may require a lot of time and effort, it provides an effective way of achieving success among competing services (Ekinci, 2002). The benefits stated above stress the importance of service quality, though what is service quality?

Service quality is considered a complex construct which has led to the development of several multidimensional models (Ekinci, 2002). To date many interpretations of service quality are given by various researchers but there is still no consensus of what it really is. Therefore, we first have to determine the approach how to define quality. According to Garvin (1983) quality can be defined from a product-based approach, a manufacturing-based approach, and from a user/ consumer based approach. Inthis thesis we focus on the user/ consumer

perspective. In evaluating quality Feigenbaum (1951) and Juran (1962) consider the

consumer‟s perspective as the perspective that defines quality. This understanding is evident within service quality literature (Schembri & Sandberg, 2002) and in preliminary results from health care report cards (Tucker III & Adams, 2001).

The user perspective of Garvin is somewhat similar to the term perceived quality. According to Zeithaml (1988) perceived quality is the consumer‟s judgment about an entity‟s overall excellence or superiority which is not equivalent to objective quality as it cannot be measured in terms of technical superiority or adherence to physical standards. Lewis (1989) defined perceived quality as a consumer judgment (a form of attitude) which results from the comparison made by consumers of their expectations of a specific service with their perceptions of actual performance on that service. This is quite similar to the definition of Bitner & Hubbert (1994) which suggests that perceived service quality is an overall judgment (attitude) about the relative superiority or inferiority of the company‟s services.

The definitions above refer to the expectation disconfirmation theory original proposed by Oliver (1980) which is predominant in both the perceived quality and client satisfaction literature (Liljander, Veronica, Strandvik, & Tore, 1993). The expectation confirmation theory suggests that feelings of satisfaction arise when consumers compare their perceptions of a product‟s performance to their expectations. If perceived performance exceeds a

consumer‟s expectations (a positive disconfirmation), the consumer is satisfied.

(14)

13

But if perceived performance falls short of his or her expectations (a negative disconfirmation), the consumer is dissatisfied (Oliver, 1980). According to Spreng,

MacKenzie, & Olshavsky (1996) there is considerable amount of empirical evidence which confirms the hypothesized impact of the expectation disconfirmation model on satisfaction and perceived quality.

2.2 Service quality in health care

Health care is a complex service which differs from other service industries and may even vary within the industry (Dean, 1999). In this thesis the focus is on hospital care, but even within hospital care different functions or service areas should be considered separately (Carmen, 1990; Reidenbach & Sandifer-Smallwood, 1990). In 2007 Berry & Bendapudi listed some important dissimilarities between health care and other services which may affect service quality perceptions. These dissimilarities are:

 Consumers are sick  Consumers are reluctant  Consumers relinquish privacy

 Consumers need „whole‟ person service  Consumers are at risk

 Clinicians are stressed

For a more detailed description of these dissimilarities we refer to Appendix A.

Beside the dissimilarities there are some similarities between health care and other service industries. First, health care is intangible, heterogenic and inseparable. Second, health care is perishable. Third, health care is a credence service which means that technical quality is more difficult to evaluate even after the service is performed (Berry & Bendapudi, 2007).

Functional and technical quality

Health care is a rare service that people need but not necessarily want and is the most personal and important service consumers buy (Berry & Bendapudi, 2007).

According to the classic work of Grönroos (1984) health service quality can be broken down into a dimension of functional quality and a dimension of technical quality. Functional quality refers to the manner how the health service is delivered to the patient. Technical quality is defined as what the patient receives as the outcome of the process in which the resources are used, i.e. the outcome quality of the process (Kang & James, 2004).

Health care quality measurement has typically been driven by technical quality with competence being a major determinant. An example of a model used to describe technical quality is the model developed by Donebian (Lam, 1997). The model of Donebedian (1988) assesses quality in healthcare with the aid of three categories: structure, process, and outcome. In this model Structure refers to resource issues such as availability of the facilities, staff, equipment, and expertise to deliver care appropriately (Ward, Rolland, & Patterson, 2005). Process measures what has been done for the patient, such as the medical tests conducted or the prescriptions issued (Ward, Rolland, & Patterson, 2005). Outcome denotes the effects of care on the health status of patients and populations (Donabedian, 1997).

(15)

14

Grönroos (1984) was one of the first authors to theorize that service quality consists of the related but distinct components technical and functional quality. He denoted the importance of the functional quality which is measured by the perception of the patient (Bowers & Kiefe, 2002). For example, Gill & White (2009) noted that improvements in functional quality will result in better health outcomes. According to Grönroos (1984) functional quality refers to the manner in which the health care service is delivered to the patient.

Important functional quality determinants of functional quality are provider reliability and responsiveness to patient needs (Bowers & Kiefe, 2002). We will elaborate on Grönroos‟s model in section 4.1.

Perspective

The perception of service quality in health care quality frequently differs between physicians and patients. This gap may have detrimental consequences for patient satisfaction and the success of the practice (McAlexander, Kaldenberg, & Koenig, 1994). Patients consider functional quality attributes such as tangibles, courtesy, and caring as to be more important whereas physicians put more emphasis on technical quality attributes such as competence, and patient outcomes (Jun, Peterson, & Zsidisin, 1998).

The main reason for this is the fact patients often cannot distinguish technical health quality from functional health quality as they find it hard to evaluate technical quality in health care simply because patients lack the expertise to gauge the clinical (technical)aspects (Oswald, Turner, Snipes, & Butler, 1998). For example, the service provider‟s technical competence, as well as the immediate results from treatments, may be difficult for a patient to evaluate (Kang & James, 2004). Because of the inability to assess technical quality patients base their quality judgment primarily on functional quality items such as attitudes towards caregivers and the facility. For example, in health care patients often rely on items such as reliability and empathy to assess quality (Kang & James, 2004). That is why Bloom & Reeve (1990) qualify health care as a credence good which means that the patient will never be able to evaluate the service completely owing to a lack of medical knowledge.

Because of the statements by Feigenbaum (1951) and Juran (1962) in section 2.1 our focus is on the consumer perception of quality. This is supported by various researchers including Van Leeuwen (2007) who states that consumer experiences define service quality.

Gabbott & Hogg (1995) stated that a major thrust in terms of exploring and improving service quality entails focusing on the consumer rather than the organization. Finally, Ward, Rolland, & Patterson (2005) stated that the mainstream quality management literature and practice use the consumer as the primary quality evaluator. To define the patient‟s perspective on service quality in health care we adopt the definition of Lee, Delene, & Bunda (2000) which suggest that quality is ultimately attained when a physician properly helps his or her patient to reach an achievable level of health, and they enjoy a healthier life.

(16)

15

2.3 The relationship between service quality, satisfaction, service value, and behavioral intentions

Service quality, satisfaction, and more recent service value have dominated service literature the past decades. They have been acknowledged by numerous scientists as the most important success factors of business competition for either manufacturers or service providers (Wang, Lo, & Yang, 2004). In figure 2 we depicted the constructs and there interrelationship as identified by Cronin, Brady & Hult (2000).

Figure 2: Perceived service quality, service value, satisfaction, and behavioral intentions

*Sacrifice was found to be not significant in the research of Cronin, Brady & Hult (2000).

Source: Cronin, Brady & Hult, 2000

As shown in figure 2 Cronin, Brady & Hult, (2000) found a direct effect between perceived service quality and both satisfaction and service value. This is in line with the opinion in literature that favorable service quality perceptions lead to improved satisfaction and value attributions and that, in turn, positive value directly influences satisfaction (Cronin Jr, Brady, & Hult, 2000).

Findings on the relationship between perceived service quality on behavioral intentions (direct, or indirect through satisfaction) led to mixed results (Dagger, Sweeney, & Johnson, 2007). Cronin, Brady & Hult, (2000) found both a direct effect of service quality on behavioral intentions and an indirect effect through satisfaction or service value. Adversely authors like Cronin & Taylor (1992) argue that satisfaction is an intervening variable that mediates between perceived service quality and behavioral intentions, and discard the direct effect of perceived service quality on behavioral intentions. Researchers like Taylor & Baker (1994) suggest that satisfaction is a moderating variable between perceived service quality and behavioral indentations. However, many researchers support the direct link between perceived service quality and behavioral intentions emphasizing the significant effect of health care service perceptions on behavioral intentions (Headley & Miller, 1993).

Perceived service quality

Service value

Satisfaction

(17)

16 Service quality and satisfaction

In the first section of this chapter we discussed the construct service quality which was the first wave of conceptual research in service marketing literature (Cronin Jr, Brady, & Hult, 2000). The second wave was consumer satisfaction which is described as „an evaluation of an emotion‟ (Hunt, 1977). In their research Rust & Oliver (1994) suggest that satisfaction

reflects the degree to which a consumer believes that the possession and/ or use of a service evokes positive feelings.

In relation to perceived service quality Zeithaml, Parasuraman, & Berry (1988) support the proposition from Olshavsky (1985) that service quality is an overall evaluation similar to attitude. Furthermore they found that satisfaction and perceived service quality are related. This is supported by a number of researchers who found that quality is positively related with satisfaction (Tucker III & Adams, 2001). In their exploratory research Zeithaml, Parasuraman & Berry (1985) also found that there were several occasions respondents were satisfied with a specific service but did not feel the firm was of high quality. This finding is supported by Liljander, Veronica , Strandvik & Tore (1993) who conclude that as perceived quality is constant over time and satisfaction is transactional in nature, it is possible that a consumer evaluates quality as good even though a specific situation is unsatisfactory (vice versa). Cronin & Taylor (1992) support this by suggesting that service quality perceptions should be considered as long-term attitudes and satisfaction should be referred to as a short-term, encounter-specific, consumer judgment. According to Taylor (1993) there is consensus in literature perceived service quality and consumer satisfaction should be uniquely

conceptualized and operationalized. Recently this statement was confirmed by Vinagre & Neves (2008), and Saha & Theingi (2009).

To date patient satisfaction continues to be used by many researchers to assess service quality in health care. In this study however we use the construct perceived service quality to assess the quality of a health care organization. We justify this decision because our focus is on the most appropriate items and dimensions reflecting hospital quality during a patient‟s

hospitalization. Therefore it is not recommended to use satisfaction as it can be influenced by a number of cognitive and affective processes (such as cost/ benefit analysis and emotions), while the antecedents of perceived service quality are limited (Oliver, 1993). Oliver (1993) mentioned that evaluations of perceived service quality specifically focuses on dimensions of service, while satisfaction is determined by a more variety of conceptual cues, such as regret, which doesn‟t need to be quality related.

In addition Crow, Gage, & Hampson (2002) conducted an extensive literature review

regarding the methodological issues and determinants of patient satisfaction in health care, in which they concluded that patient satisfaction is an unpredictable construct because it lacks a definitive conceptualization while the understanding of the process by which a patient

(18)

17

We use the relationship between the separate and unique constructs satisfaction and perceived quality to justify the choice to measure perceived service quality instead of satisfaction.

1 Perceived quality Satisfaction

The first direction shows a relation in which satisfaction depends on perceived quality. This relationship is supported by, among others, Parasuraman, Zeithaml & Berry (1985) and Cronin & Taylor (1992).

2 Satisfaction Perceived quality

This second direction shows a relation in which perceived quality depends on satisfaction. Bahia, Paulin & Perrien (2000) suggest that the causal link between satisfaction and perceived quality can go both ways depending on the factors shown in table 1.

Table 1: Factors that could influence the direction between satisfaction and perceived quality

Factors Satisfaction  Perceived quality Perceived quality  Satisfaction

1: The clients personality (propensity to be relational vs. transactional)

Client‟s transactional propensity. Reasoning one transaction at a time.

Client‟s relational propensity. Reasoning with regard to the whole relation rather than a specific transaction. 2: The nature of the service Commodity services. Services offered on a periodic and

continuous basis. Important services. Complex services.

Service with a high degree of involvement.

3: Corporate image Corporate image/ identity which is not strong enough.

Strong corporate image/ identity. 4: Stage in the client/ company

relation

At the beginning of the relation, reference to expectation and/ or satisfaction to evaluate perceived quality

Not applicable (more we advance in the relationship, more the satisfaction and perceived quality converge)

Source: Bahia, Paulin & Perrien, 2000

Of all service sectors no other sector affects the quality of life more than health care does. So, it is fair to say that health care it is a very important service.

In addition health care organizations often demand a high degree of involvement from the patient. Patients are often co-producer of the service, a consumer role identified by Legnick-Hall, Claycomb & Inks (2000) which requires a high degree of involvement.

Finally, health care is considered a complex service as patients are often not able to evaluate medical competence. So, we adopt the direction which suggests that perceived service quality is the antecedent of satisfaction. Empirical studies of Cronin & Taylor (1992), Woodside & Robert (1989), and Choi, Cho, Lee, Lee & Kim (2004) support this causal link between consumer perception of health care quality and satisfaction (Qin & Prybutok, 2008). Service quality and service value

The third and final wave on conceptual research in service marketing literature is

(19)

18

This means that service consumers place greater importance on the quality of a service than they do on the costs associated with its acquisition (Cronin Jr, Brady, & Hult, 2000). In 2006 Gupta & Zeithaml conducted a research into consumer metrics and their impact on financial performance in which they reviewed some important perceptual consumer metrics. They found that service value is considered an ambiguous and idiosyncratic consumer metric which is difficult to operationalize and therefore hard to measure. Therefore they excluded service value from their research.

Service quality and behavioral intentions

According to Fishbein & Ajzen (1975) behavioral intention is a function of both attitudes toward a behavior and subjective norms toward that behavior, which has been found to predict actual behavior. The effect of overall perceived service quality on behavioral intentions was empirically verified by Boulding, Kalra, Staelin & Zeithaml (1993). They stated that the greater consumers‟ perceptions of a firm‟s overall service quality the more likely the consumers are to engage in behaviors beneficiary to the strategic health of the firm (e.g. positive word of mouth, recommendation). In 1996 Zeithaml, Berry, & Parasuraman also established the effect of perceived service quality on behavioral intentions.

Headley & Miller (1993) as well as Reidenbach & Sandifer‐Smallwood (1990) found a significant effect of perceived service quality on patient behavioral intentions in a health care environment.

A well documented and important favorable behavioral intention is consumer loyalty, while an important unfavorable behavioral intention includes complaining (Zeithaml, Berry, & Parasuraman, 1996). According to Zeithaml, Berry & Parasuraman (1996) consumer loyalty is indicated by an intention to perform a diverse set of behaviors that signal a motivation to maintain a relationship with a firm, including positive word of mouth (Gupta & Zeithaml, 2006). A widely used metric to measure loyalty is the willingness to recommend the organization. In a large series of studies a positive and significant relationship was found between perceived service quality and willingness to recommend the hospital (Gupta & Zeithaml, 2006). The metric is often used in SERVQUAL.

Service quality and the Net Promotor Score

(20)

19

Due to its simplicity and ease of measurement the Net Promotor Score has gained popularity with many companies such as General Electric, & Symantec (USA), and Philips, KPN, & Rabobank (Holland). In addition the metric is used by comparison sites to compare companies within industries. For example the NPS metric is used since 2007 by independer.nl to

compare Dutch hospitals.

2.4 Service Quality: how to measure it?

In the previous section the constructs service quality, satisfaction, service value, and behavioral intentions were clarified and their interrelationship was indicated. This section discusses perceived service quality measurement.

Accurate measurement of service quality is as important as understanding the service delivery system because without a valid measure it would be difficult to establish and implement appropriate tactics or strategies for service quality management (Lee, Delene, Bunda, & Kim, 2000). However, measuring service quality is a difficult task. In their work Turner & Pol (1995) stated that quality of care is difficult to define but even harder to measure. For instance, health care patients themselves often have difficulty in evaluating medical

competence as mentioned in section 2.2. To overcome this problem several researchers have proposed service quality models which offer a framework for understanding what service quality is and how to measure it (Martinez & Martinez, 2010). In 2005 Seth & Deshmuk reviewed most of the accepted service quality models to date. They concluded that there is still no well accepted conceptual definition and model of service quality nor a generally accepted definition of how to measure service quality. Their linkage of service quality models can be found in Appendix B.

When measuring service quality we are interested in the relationship between the construct (perceived service quality) and its measures (items). In literature this is called a measurement model (Diamantopoulos, Riefler, & Roth, 2008). A structural model on the other hand only assesses the assumed causation among a set of dependent and independent constructs (Gefen, Straub, & Boudreau, 2000). According to Petter, Straub & Rai (2007) most researchers show commitment to solely justify and prove theoretical links between constructs (structural model examination), and thereby often ignore the relationship between measurement items and constructs (measurement model examination). It is however important to focus on both the structural model and the measurement model to fully consider the relationship between measures and their relevant latent constructs to avoid improperly specified constructs which can create measurement error which in turn affects the structural model (Petter, Straub, & Rai, 2007). This is confirmed by Anderson & Gerbing (1988) who states that a proper

specification of the measurement model is necessary to assign meaningful relationships in the structural model.

(21)

20

We speak of a multidimensional construct when the construct has more than one dimension, and each dimension is measured by items. On the contrary a unidimensional construct has only one dimension which means that all the measurement items are measuring the same aspect of the latent construct (Petter, Straub, & Rai, 2007).

Figure 3: Unidimensional model and multidimensional model

Figure 3 shows a unidimensional model and a multidimensional model. The lines show the relationships between the construct, dimensions (in the multidimensional model) and items. The relationship between the construct (in this thesis perceived hospital quality) and the dimensions (for example: tangibles and reliability) is an important way to classify service quality models (Martinez & Martinez, 2010).

According to Diamantopolous, Riefer, & Roth (2008) the direction of the relationship is either from the construct to the measures (reflective model) or from the measures to the construct (formative model). Before we correctly indentify our construct of perceived service quality as formative or reflective we first describe both constructs. We describe the constructs based on the assumption of unidimensionality. This is done because it is easier to interpret and because the basic assumptions of reflective and formative relationships apply for both the

unidimensional model and the multidimensional model. The difference is that in a

multidimensional model not only the relationship between the construct and the items can be reflective or formative but also the relationship between the overall construct and the

dimensions. A multidimensional model can be formative and reflective at the same time. This is not possible with unidimensional models which are either reflective or formative.

(22)

21  Reflective model

The reflective model accounts for more than 95% of the constructs measured with multiple items (Coltman, Devinney, Midgley, & Veniak, 2008). A key theoretical consideration is that reflective models assume that causality flows from the construct to the item (figure 4). In the case of service quality it assumes that the proposed dimensions are different forms manifested by perceived service quality (Martinez & Martinez, 2010).

Figure 4: Reflective model

Source: Original adapted from Coltman, Devinney, Midgley & Veniak, 2008

An example of a reflective model demonstrated by several researchers is „ease of use‟ (Petter, Straub, & Rai, 2007). In this unidimensional example variation in the construct „ease of use‟ causes variation in its measures such as: „easy to use software‟, „learning to use software is easy for me‟, and „interaction with software is clear and understandable‟.

 Formative model

In formative models causality flows from the items to the construct (see figure 5 on the next page). A change in items results in a change in the construct (Coltman, Devinney, Midgley, & Veniak, 2008). In the case of service quality the overall construct (perceived hospital quality) is formed by its dimensions and does not exist separately from its dimensions (Martinez & Martinez, 2010). In addition the construct (perceived hospital quality) is sensitive to the number and types of items representing the construct since the indicators define the construct (Coltman, Devinney, Midgley, & Veniak, 2008). An example of an unidimensional formative model is „sellers performance‟. In this example the construct is formed by its measures such as: „competitive pricing‟, „timeliness of product delivery‟, and „high quality products‟ (Pavlou & Gefen, 2005).

Construct

Item

(23)

22 Figure 5: Formative model

Source: Original adapted from Coltman, Devinney, Midgley & Veniak, 2008

In a study of marketing literature conducted by Jarvis, Mackenzie, & Podsakoff (2003) it was found that 29% of the studies published in the top four journals during a 24-year period improperly specified formative and reflective constructs (Petter, Straub, & Rai, 2007). This misspecification can create measurement error which in turn affects the structural model (Jarvis, Mackenzie, & Podsakoff, 2003). Furthermore, misspecification of the construct prohibits researchers from meaningful testing the theory due to improper results (Edwards & Bagozzi, 2000). It is therefore important to specify the proper construct. Jarvis, Mackenzie, & Podsakoff et al.(2003) provide a set of decision rules to identify the construct as formative or reflective. These rules and the implications for our research are listed in Appendix C.

In this study we are interested in the items, and consequently the dimensions, that influence perceived service (hospital) quality to enable an organization (hospital) to identify the items important to the patient. This implies that the causality flows from the indicators to the construct. A formative model is thus the most suitable to use. This is confirmed by the four rules listed in Appendix C. We therefore propose that perceived service quality in Dutch inpatient hospital care is a formative construct. This is supported by, among others, Rossiter (2002) who stated that future studies need to explore the service quality construct as a formative construct rather than a reflective judgment.

Because the construct of perceived service quality consists of a number of interrelated dimensions which can be conceptualized under an overall abstraction the model can be viewed as a multidimensional construct (Diamantopoulos, Riefler, & Roth, 2008). This is confirmed by Schembri & Sandberg (2002) who stated that there is consensus in literature that service quality is a multidimensional attitude held by consumers, with each dimension comprising of a number of items or service aspects. Jarvis et al. (2003) and Mackenzie, Podsakoff, & Burke (2005) stated that for a model to be multidimensional it is necessary to distinguish between (at least) two levels of analysis, namely:

 A level relating manifest items to (first-order) dimensions

 A level relating the individual dimensions to the (second-order) latent construct Construct

Item

(24)

23

In our thesis we are interested in perceived hospital quality (the latent construct) which is related in a formative manner to the several quality dimensions which in turn are measured by formative items. This is in line with Rossiter (2002) who claims that service quality is a second-order formed attribute in that its components, for example tangibles and reliability, are also formed attributes. Babakus & Boller (1992) and Buttle (1996) also noted that there is consensus among researchers that service quality is a second-order construct.

So, based on the information described in this section a multidimensional second-order formative model is at first sight most appropriate for this thesis.

Dimensions and items

As mentioned service quality is a multidimensional construct in which each dimension consists of several items that influence consumer service quality perceptions.

It can be seen as an umbrella construct with distinct dimensions (Babakus & Boller, 1992). One of the questions that continues to interest researchers is the number of dimensions of service quality in particular industries (Dean, 1999). Brady & Cronin (2001) stated that perceptions of service quality are based on multiple dimensions but that there is no general agreement as to the nature or content of the dimensions. The general consensus is that the number of dimensions is not generic and thus industry specific (Asubonteng et al., 1996). In their research Babakus & Boller (1992) stated that the number of service quality

dimensions is mostly dependent on the consumer involvement and the particular service being offered.

In this thesis the items and dimensions that influence perceived service quality in inpatient hospital care are very important because health care managers need to understand how consumers evaluate health services. Selecting inappropriate items and dimensions will undeniably lead to poor explanatory power of the scales. But if health care providers understand which items patients use to judge health care quality steps may be taken to

monitor and enhance the performance of those items (Bowers, Swan, & Koehler, 1994). This was acknowledged by Asubonteng et al. (1996) who conducted an extensive literature review and concluded that most studies support the assumption that service attributes are

determinants of behavior such as willingness to return and willingness to recommend.

Results can be expected to be higher levels of perceived quality and satisfaction on the part of the consumer (Jun, Peterson, & Zsidisin, 1998).

(25)

24

As our goal is to develop a model which measures perceived service quality in Dutch inpatient hospital care we first conducted an extensive literature review to identify the dimensions and items which are most important to the patient. We identified three stream in service quality literature: the American perspective, the Nordic perspective, and the

Hierarchical perspective.

The American perspective from Parasuraman, Zeithaml, & Berry (1985,1988) defines perceived service quality as the difference between expectations and perceptions along five quality dimensions (Seth & Deshmukh, 2005). Their measurement instrument called SERVQUAL and its modifications for health care will be discussed in chapter 3.

The Nordic perspective from Grönroos (1982, 1984) defines perceived service quality as the result of the consumer‟s view of a bundle of service dimensions, some which are technical and some which are functional in nature (Martinez & Martinez, 2010).The GM model of Grönroos will be discussed in chapter 4.

The Hierarchical perspective from Brady & Cronin (2001) suggests a hierarchical and multidimensional model which is based on a third-order factor model formed by three primary dimensions. The multidimensional and hierarchical model will be also discussed in chapter 4.

(26)

Chapter 3: SERVQUAL and SERVPERF

25

Chapter 3: SERVQUAL and SERVPERF

3.1 SERVQUAL: past and present

Parasuraman, Zeithaml, & Berry (1985) have been acknowledged as one of the founders of the American perspective on service quality measurement. They focus on the functional quality attributes of service quality such as empathy and reliability. In 1988 Parasuraman et al. developed SERVQUAL, one of the most cited and accepted measurement models of service quality to date.

The SERVQUAL (service quality) instrument has been widely cited in marketing literature and its use in industry has become widespread (Brown, Churchill, & Peter, 1993). The first SERVQUAL model of Parasuraman, Zeithaml, & Berry out of 1988 is based on the their exploratory research which led to their well-known Gap model three years before. The Gap model proposes that service quality is a function of the difference between expectation and performance along the quality dimensions (Seth & Deshmukh, 2005). It is based on insights obtained by in-depth interviews of executives in four nationally recognized service firms and a set of focus group interviews of consumers. One of the main results was the discovery of 10 service-quality dimensions consumers use in forming expectations about (and perceptions of) services they use (Parasuraman, Zeithaml, & Berry, 1985). In their research Parasuraman, Zeithaml, & Berry (1985) found that consumers used basically similar criteria in evaluating service quality regardless the type of service. The 10 dimensions, called „service quality determinants‟ are shown in table 2.

Table 2: Determinants of service quality

No. Service qualiy determinant Description

1 Tangibles The physical evidence of the service

2 Reliability Consistency of performance and dependability

3 Responsiveness Willingness or readiness of employees to provide service 4 Courtesy Politeness, respect, consideration, friendliness of contact

personnel

5 Understanding/ Knowing Making the effort to understand the consumer‟s needs 6 Competence Possession of the required skill and knowledge to perform

the service

7 Security Freedom from danger, risk, or doubt

8 Credibility Trustworthiness, believability, honesty, consumers best interest at heart

9 Communication Keeping consumers informed in language they can understand and listen to them

10 Access Approachability and ease of contact Source: Parasuraman, Zeithaml, & Berry, 1985

(27)

26

In addition to the determinants of service quality Parasuraman, Zeithaml, & Berry (1985) also identified four possible gaps in the service delivery process affecting a consumer‟s evaluation of the service experience. These four gaps will result in gap five, called the perceived service quality gap which indicates the difference between expected and perceived service quality by consumers. The original conceptual GAP model of service quality (SERVQUAL) can be found in Appendix D.

The Gap model identifies 10 (potentially overlapping) dimensions representing 97 items. As mentioned, three years later in 1988 Parasuraman, Zeithaml & Berry (1988) conducted a research in which they eliminated a number of items to refine and construct a multiple-item scale to measure service quality. This multiple-item scale for measuring consumer perceptions of service quality became the model known as SERVQUAL. The SERVQUAL model

represents a 22-item instrument along five dimensions for assessing consumer perceptions of service quality in service and retailing organizations (Parasuraman, Zeithaml, & Berry, 1988). Each of the 22 items in the SERVQUAL model was converted into: a) a statement to measure expectations about firms in general within the service category being investigated, and b) a statement to measure perceptions about the particular firm whose service quality was being assessed (Parasuraman, Zeithaml, & Berry, 1988). In accordance with recommended procedures half of the statements were worded positively, and half of the statements

negatively (Parasuraman, Zeithaml, & Berry, 1988). Consumers‟ responses were obtained by the use of a 7-point Likert scale ranging from „strongly agree‟ (7) to „strongly disagree‟ (1). The higher the perception minus expectation score, the higher is perceived to be the level of service quality (Jain & Gupta, 2004). The equation of SERVQUAL is shown in figure 6. Figure 6: Equation of the SERVQUAL model

SQ = Perceived service quality of individual i K = Number of service items

P = Perception of individual „i‟ with respect to performance of a service firm attribute E = Service quality expectation for item „j‟ that is the relevant norm for individual „i‟ Source: Jain & Gupta, 2004

The original SERVQUAL model was revised in 1991 and again in 1994 by the same authors. By then the model represented a 21-item instrument divided into the five dimensions listed in table 3.

Table 3: SERVQUAL dimensions

No. Service qualiy dimension Description

1 Tangibles Physical facilities, equipment, and appearance of personnel 2 Reliability Ability to perform the expected service dependably and

accurately

3 Responsiveness Willingness to help consumers and provide prompt service 4 Assurance Courtesy and knowledge of staff and their ability to inspire

trust and confidence

(28)

27

The SERVQUAL model is based on the expectation disconfirmation paradigm. This

paradigm maintains that satisfaction is related to the size and direction of the disconfirmation experience where disconfirmation is related to the person‟s initial expectations (Churchill & Suprenant, 1982). Because the causality flows from the indicators to the construct

SERVQUAL is classified as a formative model (see figure 7). Figure 7: SERVQUAL model

Source: Martinez & Martinez, 2010

The SERVQUAL model identifies the reasons for any gaps between consumer expectations and perceptions.

The authors of the SERVQUAL model (Parasuraman, Zeithaml & Berry) considered their model to be a valid and reliable measurement model of service quality and successfully tested their model in five service industries: retail banking, long-distance telephone company, credit card, security brokerage, product repair and maintenance. However, over time research has shown that SERVQUAL is not a generic model. For instance Babakus & Boller (1992) stated that the dimensions of SERVQUAL are dependent on the particular service being offered, while Carmen (1990) found that despite the fact that the stability of the SERVQUAL dimensions is impressive evidence suggests that the dimension are not completely generic. Most of the disagreements regarding SERVQUAL as a tool to measure service quality boils down to 3 major issues (Asubonteng, McCleary, & Swan, 1996):

1 disagreement concerning the proposed linkage between satisfaction and quality 2 the number of basic dimensions that comprise service quality

3 omission of the outcome dimension

For the first disagreement we refer to section 2.3. The second disagreement concerns the number of basic dimensions that comprise service quality. The original SERVQUAL model consists of the five dimensions listed in table 3. As mentioned Babakus & Boller (1992) stated that the proposed dimensionality of SERVQUAL is however problematic. This was

(29)

28

According Lytle & Mokwa (1992) effective outcomes could moderate individual perceptions of service quality. This is supported by Grönroos (1990) and Babakus & Mangold (1991) who concluded that SERVQUAL only measures functional quality and ignores technical quality dimensions such as outcome.

Still, despite its limitations much of the research to date has focused on measuring service quality using the SERVQUAL instrument. Not least because there is convergence of opinion that service quality comprises attributes that are both measurable and variable (Asubonteng, McCleary, & Swan, 1996).

3.2 SERVPERF

During the past decades researchers proposed several other models for measuring service quality to overcome the limitations of the SERVQUAL model. The most important

modification of the SERVQUAL model is the model which measures only perception scores. The perceptions only measurement after the service experience is known as the SERVPERF (service performance) model.

The model introduced in 1992 by Cronin & Taylor questioned the conceptual basis of the SERVQUAL model which they found confusing with service satisfaction (Jain & Gupta, 2004). This was supported by Brown, Churchill & Peter (1993), Babakus & Boller (1992), Bolton & Drew (1991), and Carmen (1990). They stressed the need for a methodological more precise scale (Jain & Gupta, 2004). Based on theoretical arguments and empirical evidence Cronin & Taylor (1992) developed the SERVPERF model in which the expectation component of SERVQUAL is removed and instead the performance alone component is used (Jain & Gupta, 2004). They removed the expectation component because patients‟

expectations failed to significantly predict their satisfaction or disconfirmation (Cronin & Taylor, 1994). This was supported by McAlexander, Kaldenberg, & Koenig (1994) who stated that models that measure service quality as performance only are superior to models that measure service quality as a function of performance and expectations. The considerable support from researchers for SERVPERF is based on two arguments (Jain & Gupta, 2004). First, because it is a performance only scale the SERVPERF model is more efficient as it consists of 22 items instead of the 44 used in the original SERVQUAL model. Second, SERVPERF is empirically found superior to SERVQUAL as it is able to explain greater variance in the overall service quality construct measured through single items.

To overcome some limitations of their SERVQUAL model Parasuraman et al. (1991)

(30)

29 Figure 8: Equation of the SERVPERF model

SQ = Perceived service quality of individual i K = Number of service items

P = Perception of individual „i‟ with respect to performance of a service firm on item „j‟

Source: Jain & Gupta, 2004

When we compare SERVQUAL and SERVPERF we can say that the latter is more

appropriate when one is interested in simply assessing the overall service quality of a firm or making quality comparisons across service industries because of its psychometric soundness and instrument parsimoniousness. The SERVQUAL model is more appropriate when one is interested in identifying a firm‟s service shortfall because of its superior diagnostic power (Jain & Gupta, 2004). In 2007 Carrillat, Jaramillo & Mulki investigated the difference

between SERVQUAL and SERVPERF in terms of predictive validity of service quality. They concluded that both instruments are equally valid predictors of overall service quality.

However, due to the fact that the SERVPERF only uses perception scores it is much more efficient.

Finally we state that SERVPERF is also a formative model because the construct is defined in terms of its measures (Martinez & Martinez, 2010). Figure 9 shows the SERVPERF model. Figure 9: SERVPERF model

(31)

30

3.3 SERVQUAL & SERVPERF and modifications for health care

In this study our focus is on perceived service quality by consumers in a health care setting. Back in 1992 Babakus & Mangold examined the usefulness of the SERVQUAL scale for assessing patients‟ perceptions of service quality in a hospital environment. They concluded that their modified SERVQUAL model is reliable and valid in the hospital environment. Unlike the original SERVQUAL model, which measures expectations and perceptions, they used a perception only scale (SERVPERF) to measure perceived service quality.

Their conclusion was justified by Cronin & Taylor (1992), Babakus & Boller (1991), and Carmen (1990 who also found that the effort to measure pre-encounter expectations and post-encounter perceptions outcome was less useful than measuring perceptions only after the service experience. They all found it far more practical because the use of a pre- and post-encounter measure in a medical setting is extremely difficult to implement. (Headley & Miller, 1993)

The most important modifications made by Bakakus & Mangold (1992) were the use of a five-point Likert scale („strongly agree‟ (5) to „strongly disagree‟ (1)) instead of the original seven-point Likert scale, and the use of 15 instead of 22 items. Besides they stated that significant wording changes may be necessary for individual service settings. This was supported by Carmen (1990), Reidenbach & Sandifer-Smallwood (1990).

(32)

Chapter 4: Other important models for measuring service quality

31

Chapter 4: Other important models for measuring service quality

4.1 The Nordic perspective

Unlike the American perspective the Nordic perspective, also known as the European perspective, not solely focuses on functional quality attributes.

The Nordic perspective considers two more components, technical quality and image (Kang & James, 2004). One of the founders of the Nordic perspective is Grönroos (1982,1984). His model of technical and functional quality, known as the GM model, was the first model measuring service quality. Like SERVQUAL, Grönroos adopts the disconfirmation paradigm to propose that quality of the service in dependent on expected and perceived service

(Martinez & Martinez, 2010).

In the Nordic perspective two forms of quality are relevant to service-providing organizations: technical quality (the what), and functional quality (the how). Technical quality in a health care environment is defined primarily on the basis of technical accuracy of the diagnoses and procedures (Babakus & Mangold, 1992). Functional quality refers to the manner in which the health care service is delivered to the patient. Since patients are often unable to assess the technical quality of a health care service, functional quality is usually the primary determinant of patients‟ quality perceptions (Donabedian, 1980). For example, the immediate results from treatment as well as technical competence of health care providers is very difficult to evaluate by a consumer (Kang & James, 2004). Therefore patients rely on how quality attributes such as empathy and reliability to assess quality. Technical and functional quality together mainly determine the image of the service provider in the mind of the consumer. Other factors influencing the image are factors as tradition, ideology, word of mouth, pricing, and public relations (Seth & Deshmukh, 2005).

In Grönroos‟s service quality model (figure 10 on the next page) image is very important. Kang & James (2004) conclude that image can be viewed as a filter in terms of consumer‟s perception of quality. The importance of corporate image in the experience of service quality is supported by Lehtinen & Lehtinen (1982). They both stated that a favorable and well-known image is an asset for any firm because image has an impact on consumer perception of the communication and operation of the firm in many respects (Kang & James, 2004). For example, if an organization has a positive image in the mind of the consumer minor mistakes will be forgiven. The opposite is true when an organization has a negative image, any mistake than will often be magnified in the mind of the patient.

(33)

32 Figure 10: Grönroos‟s service quality model

Source: Grönroos, 1984

According to Grönroos (1982) and Rust & Oliver (1994) marketing scholars have yet to identify items that define technical quality dimensions, although it is widely accepted that technical quality significantly affects consumers perceptions of quality (Kang & James, 2004). Therefore most quality dimensions are functional (process) dimensions often measured through SERVQUAL. There is however one exception as mentioned in section 3.1, namely outcome. By mentioning that functional quality is often measured through SERVQUAL we have pointed out the major weakness of the Grönroos model, namely that it does not offer an explanation on how to measure functional and technical quality. Nevertheless, the Nordic perspective has contributed greatly in the understanding of the components of service quality, in particular by recognizing the existence of outcome and image as a quality dimension (Ekinci, 2002).

4.2 The Hierarchical perspective

The Hierarchical perspective was proposed by Brady & Cronin in 2001 based on qualitative research and literature review (Martinez & Martinez, 2010). The model was proposed as an effort to integrate the Nordic and American schools of thought which the authors feel do not completely capture the construct of service quality (Brady & Cronin Jr., 2001). They included the structure-process-outcome model of the Nordic school along with the functional

dimensions of the American school in a third order performance only measurement model. They described a model in which service quality is formed by three primary quality

dimensions: personal interaction, physical environment, and outcome. Each of these

dimensions is formed by sub-dimensions as shown in the figure on the next page (figure 11).

(34)

33 Figure 11: Hierarchical model

The personal interaction dimension consists of three sub-dimensions: attitude, behavior, and expertise.

The dimension physical environment consists of the sub-dimensions ambient conditions, design, and social factors. Ambient conditions pertains to non visual aspects such as

temperature, scent, and music (Bitner, 1992). Design refers to the layout or architecture of the environment, and according to Brady & Cronin it can be either functional (practical) or aesthetic (visually pleasing). Social factors refer to the number and type of people evident in the service setting as well as their behavior (Aubert-Gamet & Cova, 1999).

The outcome dimension includes waiting time, tangibles (there were no scars after the operation), and valence. Valence reflects the degree to which the service outcome itself is perceived as good or bad (Brady & Cronin Jr., 2001). According to Brady & Cronin Jr. (2001) valence captures attributes that control whether consumers believe the service outcome is good or bad, regardless of their evaluation of any other aspect of the experience.

A challenging aspect of valence is that, in most cases, it is uncontrollable (Brady, Clay, Voorhees, Cronin Jr., & Bourdeau, 2006). For example, a patient may have a positive

perception of every service quality dimension but a negative valence of the outcome because the doctors could not prevent the patient from losing his arm. The doctors were not to blame because there was nothing they could do, but still the patient may have an unfavorable service experience due to the negative valence outcome. In their research Brady et al. (2006) found that the effect of valence on satisfaction was stronger than the relative effects of functional quality and service environment quality. Yet, they also noted that in service industries where the outcome can be controlled largely through effectively managing the process valence is less significant. In a hospital the outcome is understandably much more controllable than the outcome of a sporting event, but still valence may have quite an influence on the service experience. Valence should not be confused with service failure. The main difference between a service failure and valence is that service failure is the result of poor service were valence could be a possible outcome of a very good service (Brady, Clay, Voorhees, Cronin Jr., & Bourdeau, 2006). Overall Service Quality Personal interaction

Attitude Behavior Expertise

Physical environment Ambient conditions Design Social factors Outcome Waiting