Master-thesis H.J. Berns S1256556 12-10-2009
Supervisor: dr. K.J. Alsem Second supervisor: drs. J. Berger Faculty of economics and business University of Groningen
Abstract
The purpose of this research is to find out if quality in healthcare provides satisfaction and which elements compose quality. Furthermore, it is examined if there are differences between the different stakeholders’ assessment of patients’ perception of the consumed services, and the between the perceptions of the patients themselves. This dissimilarity is also examined for the differences among the relation between quality and satisfaction. The different stakeholders groups that are involved are, besides the patients, employees and patient organizations. This research has been applied at the Centrum voor Revalidatie-UMCG, which is the rehabilitation clinic of the university medical centre in Groningen. The reason for exploring the differences among the CvRs’ stakeholders is that a lack of congruence between patients and other stakeholders prevents an overall image of providing the highest possible care quality.
In order to measure the level of quality, the idea of SERVQUAL (Parasuram, Zeithaml & Berry, 1988), has been used. This standardized questionnaire enables to compare the results between the different stakeholders. Furthermore, some items were added to the questionnaire in order to measure satisfaction, and patients also received additional questions concerning their age, gender, and if they were in- or outpatients. The data is obtained by distributing the questionnaires among randomly chosen in- and outpatients and by emailing employees and patient organizations. Based on the initial data, the quality dimensions ‘assurance & reliability (both tangible as intangible)’, ‘personal help’, and ‘knowledge of needs’ have been formed, as the original findings of SERVQUAL are not completely relevant for the healthcare sector.
With these new dimensions, the scores of the stakeholders could be compared to the scores given by the patients, both on the three quality dimensions as on satisfaction. Besides it, the relationship between the quality dimensions and satisfaction has been measured, as well as the differences inside the patient group. Results are that the quality dimension ‘assurance & reliability’ does have a positive influence on patients’ satisfaction. Next, concerning the dimensions ‘assurance & reliability’ and ‘knowledge of needs’, employees have a more negative image on patients’ perception then patients themselves have, as well as that the quality dimension ‘assurance & reliability is more important for patient satisfaction then employees think it is. Differences inside the patient groups could not be found.
Table of contents 1. Introduction 1 1.1 Prologue 1 1.2 Conditions 1 1.3 Questions 2 1.3.1 Management problem 2 1.3.2 Problem statement 2
2. A review of healthcare quality literature 3
2.1 Totality of care quality 3
2.2 Quality measurements test 4
2.3 SERVQUAL and her modifications to healthcare and hospitals 4 2.4 Results of quality 8
2.5 Evaluation and conclusion 9
2.6 The conceptual model 10
3. Research design 13
3.1 Methodology 13
3.1.1 Used scale and items 13
1. Introduction 1.1 Prologue
January 2006, a new healthcare cost system was introduced in the Netherlands. Along with the new system came change that forced the healthcare market from a supply-driven to a demand-oriented care system. One of the many consequences of this change is that the demands of the patients and others involved will become more important then the process created by the service provider. Transparency will become more important and healthcare providers have to start showing figures of their production process of their provided service. This should lead to more available information for the consumer. This information is needed because of an increasing level of freedom of choice. Within two or three years, the consumer can choose his healthcare provider for up to 70% of the total offered treatments (NRC, January 10th 2009). This means that healthcare providers need to realise that competition will rise.
1.2 Conditions
The healthcare market includes many different sectors. However, in this research there will be chosen to focus on hospitals. This section of healthcare has the largest financial currencies, and accommodates the most employees.
Hospitals start to realise that change is needed. The old situation, in which an internal focus was adequate, shall be adapted in the near future, or has been changed already. When consumers will be able to choose between hospitals, these hospitals need to start realising that the consumer can reject their offered service. When consumers can choose between different
alternatives, a list of preferences in the mind of the consumer will be evaluated, and the hospital with the best matching image will be chosen. Besides own preferences, other factors
also can have their influence. Past experiences and word-of-mouth communication from the environment can influence the image forming as well. One of the most important aspects of a hospital’s image is quality. When choosing for a hospital, the perceived quality will have a strong influence. Therefore, for a hospital it is essential to know how important quality is to the hospital itself, what level of quality they want to provide, and what kind of level of quality is asked by their consumers.
Besides consumers, which in the case of a hospital are patients, there are other stakeholders who are important for a hospital. The most important stakeholders are the health insurers, referrers, patient organizations, and the employees. For all the stakeholders, the obtainment of a desired level of quality by the consumer is substantial. This is because the consumer as patient is a consumer for the other stakeholders as well. Therefore, because all these other stakeholders can have an influence on the choice of the consumer, their perception of quality is also important.
Concluding, when a hospital wants to distinguish itself from the competition in order to attract new or already known consumers, perceived quality is one of the most important features. When paying the right attention to quality, and when evaluating which level of quality is desired and how this must be proclaimed, this could be an answer to the changing market conditions.
Haren, a small village near Groningen. Together, both locations offer treatments for several issues, such as amputation, cerebral palsy, multiple sclerosis, or neurorehabilitation, in all 19 different treatments. Being the only academic rehabilitation centre in The Netherlands, the provided quality must be as high as possible, just as stated in the centre’s mission. Although the CvR-UMCG is a rehabilitation centre, this research will focus on the level of hospitals. Despite the fact that there are dissimilarities between the characteristics of a rehabilitation centre and a regular hospital, the similarities are in majority. One difference is the longer lasting relation between a consumer and the medical employees, caused by the longer duration of the service. Where most of the clinical hospital consumers stay for several days, clinical consumers in a rehabilitation centre can stay for several months. However, this should not have a noticeable influence on the type of delivered service. Therefore, the CvR-UMCG will be considered as regular hospital.
1.3 Questions
1.3.1 Management problem
The choice for visiting a specific hospital exists out of three parts. First, the consumer will often focus on hospitals which are located nearby. Secondly, the nearest hospitals must offer the treatment that is demanded. Due to present developments in the market conditions, the consumer is clarified that he has more choice of hospitals. However, even when the competition is reduced to a regional group of hospitals, the noticeable level of quality is the third determinant in consumer choice. Therefore, it has become more important to differentiate from competition. Consumers will have more freedom in choosing between hospitals, even when this is only regional, and if their preferences of quality do not match with the level of quality of the offered services, the consumers will not choose to visit the hospital anymore for consuming the offered service. If the offered service has the desired level of quality, the consumer will be satisfied and will choose the hospital instead of the competition. This behaviour will continue in the future, which leads to long term profits and a continuation of the conduct of business. Other stakeholders also have a perception of quality, and will base their decision of contacting the CvR-UMCG, or their advice to consumers to visit the CvR-UMCG, on the provided quality level. As mentioned, this could also have influence on the decision of the consumer if he will be visiting the CvR-UMCG.
1.3.2 Problem statement
Quality is one of the more important determinants of the decision whether or not to consume the offered service. When hospitals’ services are compared, quality will be a key-feature. Therefore, for the CvR-UMCG it is important to know what level of quality is desired by the consumer in order to be satisfied, and if this level of quality is delivered by the CvR-UMCG. For that reason, the main question of this research will be:
“Does quality in healthcare provide consumer satisfaction?” Sub questions will be:
“Which dimensions of quality will lead to satisfaction?”
2. A review of healthcare quality literature 2.1 Totality of care quality
Quality is a widely used concept with an even wider substance. From the quality of a small and daily used product to a concept such as the quality of life, the same term is used. However, the intention is different. Measuring quality when dealing with a product should not be too complicated. Several product characteristics can be measured. When measuring the quality of a service, this could become more complicated, especially when the quality of healthcare is measured. The concept of quality in healthcare can be explained in several manners. In the research of Bowers & Kiefe (2002), the differences of the concept of quality in healthcare is clarified. For healthcare providers, quality means the physical quality of a patient. Has a medical interference a positive outcome? And is the disease fully cured? Quality in the case of healthcare insurers means the reduction of diseases. This could be measured in a hospital, or in a whole regional area. Healthcare quality for patients is a function of their individual personal interaction with their healthcare providers, as well as in terms of outcomes.
As mentioned earlier, the concept of quality in healthcare in this research will focus on the view of the patient, as he is the consumer. Quality can be measured best by measuring the level of satisfaction, which is an attitude and related to quality (Chassin & Galvin, 1998; Jun, Peterson & Zsidisin, 1998). A high level of quality will lead to a high level of satisfaction. For this reason, quality can be an important issue for strategic marketing. This will be discussed later on. Contrary, satisfaction is based on the perception of quality. Considering the fact that in this case quality is a medical subject, the medical or physical quality is combined with the consumer’s perception of quality to form his level of satisfaction (Bowers & Kiefe, 2002).
The different forms of quality for a consumer can be explained by the division of quality by Grönroos (1984). Quality can be divided in two components. The first component is technical or outcome quality. In this case, this could be seen as medical quality. It is based on the performance of the service. The second component is functional or process quality. Functional quality is measured by the context and means of the delivery of service, as perceived by the consumer. In healthcare, consumers are hardly capable of judging the technical quality. Therefore, consumers find it easier to make a judgement based on the atmosphere and personal interaction.
Together, functional and technical quality will form the image of the service in the minds of consumers (Jun, Peterson & Zsidisin, 1998) (see figure 1).
Attention should be paid to the revealed halo-effect of Wirtz (2003). Process attributes could influence the observed core of the service quality. This effect is based on the concept of functional quality. Because consumers are not capable of evaluating the whole service, they create an image of the service, based on easy accessible data. The more dimensions a
Quality
Technical Functional
Image of Service
consumer can base his evaluation on, the less the halo-effect is. Furthermore, a higher level of consumer involvement will decrease the halo-effect. In order to measure the perceived functional service quality, several tests are created. As functional quality will be seen through the eyes of the consumer, it is important to know that this form of quality is a perceived quality. Quality itself is not as important as it is encountered by the consumer, especially in healthcare. When a hospital does not evaluate functional quality through the eyes of the consumer, important aspects of quality will be over-seen (Kenagy, Berwick & Shore, 1999). 2.2 Quality measurement tests
In order to standardize the measurement of functional quality in healthcare, several tests are developed. Some tests are developed for non medical purposes. However, these tests could be adapted to the circumstances of healthcare. More specific tests are developed to measure functional quality in hospitals. Nelson et al. (1992) are using a measurement method called ‘Hospital Quality Trends’ (HQT). This system is developed by healthcare professionals and researchers. One of the main conclusions from this research is that Nelson et al. (1992) found another form of the halo-effect, which creates an effect that the respondents rate the functional service level generally high or low on all factors, without any noticeable differences between the different dimensions. Fornell et al. (1996) are using the American Customer Satisfaction Index (ACSI), which compares the customer expectations with the perceived functional service quality. This score is an indicator for customer loyalty. The researchers used 250 firms from all kind of different service delivering sectors, including hospitals. The results indicate the overall level of satisfaction nation wide (USA). Furthermore, the index can indicate the satisfaction level per sector. One of the most used tests in functional service quality measurement is the SERVQUAL model (Parasuraman, Zeithaml & Berry, 1988). Just as the research of Fornell et al. (1996), SERVQUAL measures the gap between the expectations and perceptions of functional quality. In the next paragraph the SERVQUAL model will be explained, completed by several modifications in order to adjust the model to healthcare and hospitals.
2.3 SERVQUAL and her modifications to healthcare and hospitals
The SERVQUAL model measures five different functional quality dimensions: • Reliability; this dimension measures if the service provider acts as promised. • Responsiveness; which indicates if the service provider acts promptly. • Assurance; means that the service must be safe, secure and comfortable. • Empathy; this measures if consumers are treated with respect.
• Tangibles; which indicate the perception of the physical environment.
These five functional quality dimensions were formed out of 10 identified criteria (see figure 2). Because in subsequent research a high degree of correlations between some of the variables were found, Parasuraman, Zeithaml & Berry (1988) consolidated these variables into the five broader dimensions (Lovelock & Wirtz, 2007).
test. Nevertheless, all of the following researches are based on the SERVQUAL model, although modified.
Headley & Miller (1993) came up with a modification specified to healthcare. They found that the pattern of items leading to the five dimensions in SERVQUAL did not match. In this research, they found six new dimensions, with none being identical. The found dimensions are: • Tangibles • Reliability • Responsiveness • Empathy • Dependability • Presentation
The dimension ‘assurance’ was eliminated from the original SERVQUAL scale, and the new dimensions ‘dependability’ and ‘presentation’ where added to the measurement. One of the items in the dimension dependability was: “The provider is dependable” and an item in the factor presentation was: “Employees are well dressed and neat”. Furthermore, Headley & Miller (1993) found that the dimensions of reliability, dependability, and empathy where most predictive for the behaviour of the consumer, both with positive and negative effects. When the dimensions had a positive score, this led to fewer complaints, more compliments, and increasing repeated purchases and less switching of care providers. This behaviour is an indicator for the hospital’s long term profitability. An addition to this conclusion will be explained further on.
Bowers, Swan & Koehler (1994) also came up with a modification of SERVQUAL. The researchers found five patient-based determinants of quality, which should lead to a higher level of satisfaction:
Original criteria: Tangibles Reliability Responsiveness Competence Courtesy Credibility Security Access Communication Understanding the customer Dimensions: Tangibles Reliability Responsiveness Assurance Empathy
• Empathy • Reliability • Responsiveness • Communication • Caring
These dimensions are based on the SERVQUAL measurement, except that the adaptations are based on a qualitative research. All are found to be significant. The first three are SERVQUAL dimensions; the last two are added in order to complete the healthcare functional quality dimensions. In SERVQUAL, communication and caring were covered by empathy (Lovelock & Wirtz, 2007). However, Bowers, Swan & Koehler (1994) found that these dimensions where too important in healthcare in be covered by other dimensions. The SERVQUAL dimension tangibles was found to be non significant.
Another modification was developed by Babakus & Mangold (1992). These researchers have modified the SERVQUAL scale in order to make a more practical and user-related scale which could be used in hospitals. Most important adaptation is the use of 15 items instead of 22, for the reason that in this research some items where found to be redundant and led to a lower response rate. Besides, the questions could be answered in a 5 points Likert scale instead of 7 points, because past experiences learned that a 7 point Likert scale would cause more frustration among the respondents. In the results, only one factor was found instead of five. However, it still contained five observable variables in accordance with the five original factors of service quality. The results were that patients do have a clear image of the desired level of service attributes. However, there is a lack of knowledge in evaluating the overall service performance. Another result of the research is that by using SERVQUAL, lower scores in the answers could indicate that there are deeper underlying problems in the organization of the hospital.
A notable adjustment to the measurement of SERVQUAL is unfolded by Wirtz & Matilla (2001). They revealed that whenever there is a situation in which there is a market with less freedom of choice, and the consumer has a perception which lies marginally above the level of expectations, this does not mean that this is a positive outcome. Therefore, the wants and needs of consumers can be measured instead of the expectations. If the wants and needs are exceeded, this will mean that there is a positive outcome. However, the system of pre- and post-measurement can be substituted by one especially influential modification of Cronin & Taylor (1994). These researchers found out that using perception only scores provides superior quality measurement. Besides, it cuts the number of questions in half. Although this research was not developed in or for the healthcare sector, it solved an important difficulty. Measuring expectations of patients that will be visiting a hospital in the future is difficult. As the after-measurement only can provide an adequate quality of data, it solves the mentioned complexity.
After this important research, many researchers used the perception only measurement. Kilbourne et al. (2004) used perception only measurement, for the reason that measuring perceptions only led to a higher convergent and predictive validity. Although Kilbourne et al. (2004) used nursing homes in a cross-national test; they found that they could delete the fourth item of the original SERVQUAL, which consist of tangible elements, because of the low reliability of the scores. This is equal to the findings in the research of Bowers, Swan & Koehler (1994).
The research of Parasuraman, Zeithaml & Berry (1988), which have created the SERVQUAL model, and the quality dimensions of Bowers, Swan & Koehler (1994), are compared with quality dimensions which are created by the research of Jun, Peterson & Zsidisin (1998). A comparison of the found dimensions can be found in figure 3.
These dimensions were found by employing focus groups interviews. Participants of those groups were patients and hospital employees. The found topics were the most mentioned ones, which was revealed by word coding qualitative data. Almost all the dimensions were mentioned earlier in the SERVQUAL model or the modification of Bowers, Swan & Koehler (1994). Parasuraman, Zeithaml & Berry (1988) uncovered tangibles, reliability, responsiveness, competence, courtesy, communication, access, caring, and understanding the consumer. Tangibles, reliability and responsiveness were mentioned directly in SERVQUAL, and the other dimensions were covered by assurance and empathy. The research of Bowers, Swan & Koehler (1994) covered directly reliability, responsiveness, communication, empathy and caring. Some of the found items by Jun, Peterson & Zsidisin (1998) were found non-significant by Bowers, Swan & Koehler (1994), such as tangibles, competence, courtesy, and patient outcomes. Jun, Peterson & Zsidisin (1998) revealed a new dimension, collaboration (see figure 3). Furthermore, this research verified the differences between importance of dimensions indicated by different stakeholders, such as patients, medical administrators and physicians. Physicians did not mention courtesy and rated responsiveness and caring as not important, contrary the answers of the patients. Patients did not mention patients’ outcomes as a dimension of functional quality, and medical administrators did not mention caring. These outcomes have revealed that medical employees focused more on technical quality, while patients concentrated more on functional quality, just as mentioned by O’Conner, Trinh & Shewchuk (2001). However, Jun, Peterson & Zsidisin (1998) did not perform a statistical testing of the answers, which Parasuraman,
Dimensions of Jun, Peterson & Zsidisin (1998)
Mentioned by
Parasuraman, Zeithaml & Berry(1988)
Mentioned by Bowers, Swan & Koehler (1994) Tangibles Reliability Responsiveness Competence Courtesy Communication Access Understanding patients Caring Patient outcomes Collaboration Tangibles Reliability Responsiveness Assurance Empathy Not mentioned Not mentioned Not significant Reliability Responsiveness Not significant Not significant Communication
Empathy (access + understanding patient)
Zeithaml & Berry (1988) and Bowers, Swan & Koehler (1994) did accomplish. O’Conner, Trinh & Shewchuk (2001) and Young et al. (2009) found that medical employees have their own perception on the delivered functional quality. However, employees must be aware of the quality perception of patients in order to create the desired level of patient satisfaction. Medical employees tend to underestimate the importance of perception for most of the quality dimension mentioned by the SERVQUAL model. Tangibles are the only dimension of which the importance for the patient is overestimated. The rest of the patient perception is not valuated as it should be. Jun, Peterson & Zsidisin (1998) emphasize that employees are stakeholders as well. Therefore, their perception of quality is important too. This is because employees will reveal their perception of quality to the patients.
2.4 Results of quality
According to Bowers & Kiefe (2002), the determinants of service quality form the consumers’ perception of the overall quality, which has a strong influence on the level of satisfaction. Eventually, a higher level of satisfaction will lead to positive intentions and the corresponding behaviour to revisit the hospital in the future, or to create a positive worth of mouth communication to relatives and friends. Finally, this action shall create an amplified financial performance. This assertion is also confirmed by the research of Choi et al. (2004). Nelson et al. (1992) are even more concrete. They endorse the concept that quality can lead to the formation of the affective dimension of satisfaction. Because satisfaction will enforce the behaviour of returning to the hospital for future treatment, this behaviour will influence the financial performances of the hospital by 17% to 27%.
Next, Fornell et al. (1996) and Anderson, Fornell & Lehmann (1994) also found that perceived quality has a positive effect on overall consumer satisfaction. According to the researchers, this level of satisfaction will lead to long term profitability. Choi et al. (2004) found both the connection between the level of perceived service, satisfaction and behavioural intentions, as the direct connection between the level of perceived service quality and behavioural intentions. In order to loop back to healthcare, there are more influences on quality. Besides the perception of functional quality, the assessments of physical outcomes also have influence. Together, the overall satisfaction is formed (Bowers & Kiefe, 2002). However, Rust, Zahorik & Keingham (1994) found that although satisfaction is influenced by situational and personal factors beyond the reach of the service, quality appears to be the most influential determinant of service satisfaction.
Once more, it is important to realise that several hospitals’ stakeholders have an influence on the quality perception and the final behaviour of the consumer. Therefore, all the relevant parties need to be satisfied with the delivered service. Bowers & Kiefe (2002) and Headley & Miller (1993) state that if this alignment is proper, the future behavioural interactions between the stakeholders, the consumer and the hospital will be positive.
2.5 Evaluation and conclusion
Several service quality tests have been discussed. Most of the found literature of service quality in healthcare has been based on the SERVQUAL model. Many adaptations have been made. Most of these adaptations are based on the 22 items of SERVQUAL, and differed in the statistical calculations. Where Parasuraman, Zeithaml & Berry (1988) found five quality dimensions, other researchers found other dimensions, most after interpreting the factor analyses. However, regardless the outcomes of the computations, most of the researchers used the 22 items for their research. Only a few researchers did not use questionnaires and obtained answers by focus group registration.
The most useful recapitulation of the literature can be used to summarize the findings. Babakus & Mangold (1992) used a 5 point Likert scale instead of a 7, based of previous experiences. Cronin & Taylor (1994), Choi et al. (2004), and Kilbourne et al. (2004) used perception only measurement, instead of comparing expectations and perceptions, without a loss of useful data. Furthermore, Babakus & Mangold (1992) stated that complexity of the questionnaire, which could lead to a lower response rate, should be reduced as much as possible. Most of the found researches were executed in North America, with an exception of Kilbourne et al. (2004), who made a comparison between British and American nursing homes, and Choi et al. (2004), which performed a South Korean study. An explanation could be that the North American healthcare market has the highest degree in freedom of choice since decades. This situation made it necessary to perform research to the organizational experiences of stakeholders. Since the most of the European healthcare markets are just starting to liberate, the incentive to perform the gross of research to this subject has only started recently. The found testing techniques to measure functional quality and satisfaction can be applied in European markets, such as the Dutch, without the need for drastic adaptations. However, it could not be presumed that the found answers also are suitable in the European situations. Due to the differences in culture between North America and Europe, or more specific The Netherlands, it can be assumed that the answers of the functional quality measuring tests also differ. Therefore, this research will use for the main part the generic measurement of the SERVQUAL test, in order to reveal the quality dimensions that are relevant in The Netherlands.
Certain comment can be made when all these researches are examined. First, the researches that developed a functional quality evaluation tests, such as Nelson et al. (1992) and Fornell et al. (1996), are not commonly used in further research, in contrast with the SERVQUAL model, nor is it commonly used in current hospitals. Another comment to the research of Fornell et al. (1996) is that this measurement does not indicate satisfaction at firm level, and only creates an industry comparison. Therefore, practical implications for organizations are not possible.
Next, as shown, several researches have adaptations of SERVQUAL for application in hospitals. However, not all the findings are generally applicable, or confirmed by other researchers. Headley & Miller (1993) found two new dimensions in the SERVQUAL model: dependability and presentation. Nevertheless, more recent research in healthcare and hospitals did not found the same results as these new dimensions. Choi et al. (2004) found even more diverse dimensions, which were not used in further researches. Most remarkable in their research is that they found that tangibles are usable as a dimension.
stakeholders should be measured as well. The level of their quality perception has influence on the long term profits too. If a general practitioner experienced a non pleasant experience with a hospital, or he observed negative experiences of his patients, he could refer or suggest his patients to other hospitals. Other stakeholders such as insurers and patients organizations could have the same influence on patients, or consumers, by influencing their choice.
As mentioned, every stakeholder has his own way of forming his perception of quality. However, patients are the only ones who perceive the actual functional quality. Employees or general practitioners are more focussed on the technical aspect of quality, and it is reasonable to state that the insurers or patient organizations do not have a personal experience with visiting a hospital as representative of their organization. Since the functional quality is measured, other stakeholders are asked how they assess what the perception of their patients or clients is. Recapitulating, other stakeholders are asked how they asses the patients’ perception and satisfaction.
Also mentioned before, the only stakeholders that are used for measuring the level of functional quality in the gross of the literature are consumers and employees. However, in a hospital setting, there are more influential stakeholders. Besides patients and physicians, this research will include general practitioners, health insurers, and patient organizations as well. This will create a comparison between the results of the measurement of different groups of stakeholders of how the patients perceive functional quality.
Another comparison can be made. The perception of functional quality is an indicator for the level of satisfaction. In order to expose this relationship, satisfaction will be measured directly, besides the items of SERVQUAL. Oliver (1993) developed a four-item scale in order to measure satisfaction. The overall level of satisfaction can be compared to the dimensions of functional quality, as measured by the SERVQUAL test.
Finally, the patients will be asked for some personal aspects. For example the age, the gender, and if they were in- or outpatients. These questions are only relevant for patients, and will not be asked to other stakeholders, as the respondents of other stakeholders are representing an institution.
2.6 The conceptual model
Based on the found literature, several common themes can be noticed. Headley & Miller (1993) found that future behaviour is influenced by satisfaction, and that this satisfaction is derived from perceptions of quality. This is in accordance with Choi et al. (2004), who found that perceived service quality will lead to satisfaction. That quality will lead to satisfaction is also found in the researches of Babakus & Mangold (1992), Anderson, Fornell & Lehmann (1994), Fornell et al. (1996) and Bower & Kiefe (2002). In addition, Bowers, Swah & Koehler (1994) found that several of their found quality dimensions have a positive influence on quality. For the reason that all these researches examined if the perceived quality did have an influence on the level of satisfaction, this research also will start to verify this. Therefore, the first hypothesis is:
H1: Quality dimensions have a positive influence on satisfaction.
Another common theme is that medical employees tend to asses the patient perception different from the actual patient perception. Jun et al. (1998), O’Connor, Trihn & Shewchuck (2001) and Young et al. (2009) found that medical employees asses that the patients have a lower perception of the provided organizational service quality then that the service truly is perceived by the patients themselves. Since medical employees have the same education and background as general practitioners, the same effect can be expected for this group of stakeholders. The employees of the patient organizations and health insurers do not have a medical education, neither have they been a patient. This first part means that it is not expected that these stakeholder groups underestimate the patient’s perception. As the greater part of these groups has not been a patient, they did not experience the service quality, and therefore missed the small details that help to form a more precise image of the perception. In order to examine if this under- or overestimation also can be found in this research, the second hypothesis is:
Patient’s dimensions of quality Patient’s satisfaction Dimensions of quality Satisfaction Insurers General Practitioner Patient Employees Patient Organizations
Figure 4: Conceptual Model
As viewed by: As viewed by:
H2: Employees (H2a) and general practitioners (H2b) have a more negative image on patients’ perception then patients selves have.
H2: Patient organizations (H2c) and health insurers (H2d) have a more positive image on patients’ perception then patients selves have.
Besides the underestimation of the patients’ perceptions by several stakeholder groups, this same effect can be expected for the relation between the perceptions of the service quality and the level of satisfaction. Jun et al. (1998) found that functional quality (see figure 1) is the most important part for patients, where technical quality is most important for employees. Due to the fact that this research focuses on functional quality, the mentioned quality will be less important for the employees, which results in lower scores given by the employees. Concluding, employees assess the patients’ relationship between the perceived quality and satisfaction at a lower level then the actual relationship of the patients. Again, the general practitioners share the same background with the employees. For this reason, the same effect can be expected for this group. Thus, the third hypothesis is:
H3: Quality will have a larger effect on satisfaction for patients then when this is assessed by employees (H3a) or general practitioners (H3b).
Finally, Rust, Zahorik & Keingham (1994) and Bowers and Kiefe (2002) examined if personal factors did influence the perception of quality. Although Rust, Zahorik & Keingham (1994) expected this effect, it was not found. However, Bowers & Kiefe (2002) did not want to exclude this influence, and emphasized personal factors may influence the patients’ level of satisfaction. Elements of personal factors are the age of the patient. In this research will also be examined if the patient is treated internal or external. It can be expected that when age rises, the wants for quality will increase, accompanied by the change concerning the influence of quality on satisfaction. Another difference can be expected between in- and outpatients. As outpatients have a better mobility, it is easier for this group to change from healthcare provider. Therefore, this group will be more critical, and quality dimensions will have a stronger effect on satisfaction. This can be summarized in the fourth hypothesis:
H4a: When age rises, the quality will have more influence on satisfaction. H4b: Quality will be more important for outpatients then for inpatients
3. Research design 3.1 Methodology
The purpose of this research is to measure the perceived quality of the different stakeholders of the CvR-UMCG, which is stated in the main and sub questions, as well as the hypotheses. The perceived quality is measured by a questionnaire, which is based on the SERVQUAL (Parasuraman, Zeithaml & Berry, 1988) questionnaire. This questionnaire is distributed among all the stakeholders. When all the stakeholders have completed the same questionnaire, a comparison between perceptions and the relation between these perceptions and satisfaction can be made. The questionnaire is composed out of three elements. The first measures perceptions, the second satisfaction, and the third part is meant for patients only and asks for specific patient related characteristics. The results will be shown in chapter four. 3.1.1 Used scale and items
The questionnaire that will be used contains 22 questions, instead of the 44 question in the survey used in SERVQUAL. As stated by Cronin & Taylor (1994), perception only measurement provides superior measurement quality. Therefore, this approach will be applied in this research and only perceptions will be measured.
The implemented questionnaire is using a Likert scale, the following scaling steps are predetermined. Although a Likert scale is formally an ordinal scale, the data can be used as an interval scale. This will enrich the usability of the data. As the respondents can indicate the level of agreement about the statement, the answers can be categorized as an itemized rating scale. The five grades of possible answers have a range from ‘strongly disagree’ to ‘strongly agree’. As mentioned, all the 22 items that are used in the questionnaire are given by Parasuraman, Zetihaml & Berry (1988), and the remaining four items are given by Oliver (1993). When the items in the questionnaire are changed, the answers of the research can not be compared with organizations in the same or in other industries.
In the original SERVQUAL questionnaire, the 22 items form five different dimensions of the perception of quality. Every item has one question. Examples of questions can be seen in table 1.
Dimension Example of question Items per
dimension Tangibles The CvR has up-to-date equipment 4
Reliability When the CvR promises to do something by 5 a certain time, it does so
Responsiveness The CvR does not tell customers exactly when 4 services will be performed
Assurance You can trust employees of the CvR 4 Empathy The CvR does not give you personal 5 attention
As several questions are stated negative, just as the original questionnaire of SERVQUAL, the scores of these questions are reversed in the dataset. Only 22 questions will be asked to measure the 22 quality items, instead of 44. Although the original SERVQUAL test uses 44 questions, literature has proven that only post-measuring the perceptions reduces the complexity of the questionnaire and increases the response rate, without a loss of quality of the dataset.
Just as the questions that measure the perception of quality, satisfaction also is measured by one question per item, which forms four items in total. Again, the answers can be given in a five step Likert-scale. The questions asked in order to measure satisfaction can be seen in table 2. At last, some patient related questions are asked, which can reveal differences among patients themselves. These questions can be found in table 3.
Example of question
1 My experience at the CvR was good 2 I am happy that I decided to go to the CvR
3 My visit to the CvR worked out as well as I thought it would 4 I am sure it was the right thing to go to the CvR
As the research is performed in The Netherlands, the questions are stated in Dutch. In order to prevent misinterpretation caused by translation, the questions have been translated in Dutch and retranslated into English. The Dutch translation is adapted if the retranslated and original questions did not match. The original questionnaire, in Dutch, can be found in appendix 2.
Example of question
1 My age is: 18-39 / 40-59 / 60≥ 2 My gender is: male / female
3 I am treated as a: outpatient / inpatient
3.1.2 Setting
The research is performed on the most important stakeholders of the CvR-UMCG. The perceived quality and satisfaction of patients is compared to how other stakeholders asses the perception of the patients. The other stakeholders are the employees of the CvR-UMCG, general practitioners, health insurers, and patient organizations.
Patients received the questionnaire while they were still under therapy. Both inpatients as outpatients were approached, from both the locations of the hospital, Groningen and Beatrixoord. 60 patients were approached in total, as this is an adequate amount compared to the annual patient population. This is accomplished by approaching every second patient. 30 outpatients were approached while they stayed in waiting rooms or in restaurants of both locations. If patients were willing to cooperate, they could leave the completed questionnaire Table 2: Examples of satisfaction questions
are gathered in the family room of the ward. They could also leave the completed questionnaires in specially assigned boxes. The respondents received a paper version of the questionnaire, which is preceded by a cover page. Besides the purpose of the research, this page also explained some terms that are used, and how the respondents could use the answer options. Outside of the 22 quality perception questions and the four satisfaction questions, the questionnaire meant for the patients also included some patient related questions, such as age, gender, and if patients are treated as in- or outpatient. Other stakeholders were not asked for this information, since this is not relevant as these stakeholders represent an institution. The answer options for age were divided by three categories, between 18 and 39, between 40 and 59, or 60 and older. This division is about the same as the stages of life. Respondents had to be older then 17. Above this age, the patients are appropriate to choose their own medical choices. Furthermore, the research is only focussed on directly involved persons or institutions. Therefore, partners or relatives of respondents were excluded.
Just like the group of patients, employees from the CvR-UMCG were addressed to complete the questionnaire. 65 employees were addressed by email, in order to approach an equal amount of potential respondents, compared to the patients. This amount is also adequate compared to the population of employees. The email addresses were selected as every 7th person on the organizational email list was picked. These persons received an email, which is used as a cover page. In the email, the purpose of the research and the role of the employee in this research were explained. The questionnaire was added as an attachment. On the first page of the questionnaire, instructions were given how to complete the questionnaire and explanation of some used terms. Furthermore, the employees were explained that they must complete the questionnaire as how they asses the patients’ perceived service quality, just as the other non-patient stakeholders. The employees could return the completed questionnaire by email, or it could be printed out and dropped in one of the boxes which were also used to collect the questionnaires of the patients. Email was used to accelerate the process, which should increase the level of response. When the deadline for answering came near, an email was send as a reminder.
Before general practitioners were addressed by mail, they are asked by phone if they were willing to cooperate. If so, the general practitioners would receive a letter which contained the cover page of the questionnaire. On this cover page, the purpose of the research and the role of the general practitioner in this research were explained. Adjacent to it, the mail contained the questionnaire and a stamped and addressed envelope which can be used for returning the completed questionnaire. The questionnaire was introduced on the first page, on which instructions were given how to complete the questionnaire and explanation of some used terms. It was the intention to approach all of the known large general practitioners’ group practices in the province Groningen and the North of the province Drente with the question if these general practitioners wanted to corporate in the research.
As there is only one insurance company that contracts the CvR-UMCG, only one respondent would represent this group of stakeholders. The person who is responsible for the contact between the insurer and the CvR-UMCG is approached to fulfil the questionnaire by letter. The cover page included a short introduction about the CvR-UMCG and described the goal of the research. Just like the other non-patient stakeholders, the first page contained an explanation of the used terms in the questionnaire. Next, it stated that the questionnaire should be completed as how the insurer views how their clients perceive the service of the CvR-UMCG. When completed, the questionnaire could be returned by mail.
excluded from the research, as stated earlier. From other organizations, the CvR-UMCG does not have a formal contact address. Just as the employees, the email itself contained a cover page. This mail included a short introduction of the CvR-UMCG and the purpose of the research. Again, it emphasized that the questionnaire should be fulfilled as how these organizations asses that their members perceive the service of the CvR-UMCG. The questionnaire could be returned by email.
3.1.3 Data analysis
In total, 77 respondents were involved in this research. The overall response rate is 57%, as 59 of the 136 addressed respondents did not answer. The respondents were divided in earlier mentioned groups.
60 Patients have been approached and 46 have answered, which leads to a response ratio of 77%. This group of patients consists out of 22 male (48%) and 24 female (52%). 7 of the patients have an age between 18 and 39 years old (15%), 17 patients are between 40 and 59 years old (37%), and 22 patients are 60 years or older (48%). This division is conforming the expectation that as the age of a person increases, the possible need for revalidation also increases. 23 Are outpatients (50%), and the other 23 are inpatients (50%). In the group of employees, 24 employees responded while 65 were approached. This results in a response rate of 37%. The 11 relevant patient organizations involved in the healthcare process of the CvR-UMCG were approached. 7 Of them were capable of answering, which is a response rate of 64%. A summary of these data is given in table 4.
As none of the approached general practitioners were willing to cooperate, this whole group is excluded from the research. Other manners of approaching in order to reinforce cooperation have been tried. However, none of these attempts led to success. Furthermore, there is only one health insurance company which is capable of purchasing care, and this company only has sufficient knowledge to be able to answer the questionnaire. Nonetheless, only one respondent representing a whole group is not adequate. For this reason, this group also is excluded.
Group Participated Total Patients (46) Participated
Number / % Number / % All 77 / 57 136 Male 22 / 48 Patients 46 / 77 60 Female 24 / 52 Employees 24 / 37 65 18-39 7 / 15 Patient- 7 / 64 11 40-59 17 / 37 organizations 60+ 22 / 48 Inpatients 23 / 50 Outpatients 23 / 50 3.2 Quality dimensions
In the second chapter of this research is mentioned that the original SERVQUAL dimensions are not adequate enough for healthcare. For this reason, a factor analysis is used
to reveal if the original dimensions are sufficient, or if new healthcare quality dimensions emerge. A factor analysis is used to find the dimensions of quality, per group of stakeholders.
First of all, in the dataset with the response from the group of patients, two questions were removed due to the fact that these two questions created multicollinearity. Multicollinearity is a strong relation between two different independent variables. Multicollinearity can be recognised by the correlation matrix, which includes all the 22 items. When any item has an average significance over .05, or has individual correlation coefficients above .9 or when one of the coefficients is zero, this is an indication for multicollinearity. Initially, both items loaded in the first factor. When the two items were maintained, they would have an inappropriate influence on the correlation coefficient of the whole dimension. After removing (“When you have problems, the CvR is sympathetic and reassuring” and “Employees of the CvR are polite”), a new factor analysis has been run. Furthermore, it has been verified if the amount of collected data was valid. This has been confirmed by a Kaiser- Meyer-Olkin Measure of Sampling Adequacy which is above .7, the Bartlett’s Test of Sphericity which is significant, and all of the communalitiets are above .5.
Three new quality dimensions have been found, which can be found in table 5. The tenth question (“The CvR does not tell patients exactly when services will be performed”) loaded as its own factor, and for this reason this question is not discussed any further. This question also is an example of a question where the scores of the respondents have been reversed to change the scores from negative into positive. The exact choice for, originally, four dimensions can be found in appendix 3, just as the clarification why the names of the quality dimensions are chosen. After the fourth dimension, the initial eigenvalues are less then one, and over 70 % of the cumulative loadings are already explained. The first quality dimension is ‘assurance and reliability’, which is both tangible as intangible. The Cronbach’s α for this dimension is .93.
The second quality dimension is ‘personal help’, and the Cronbach’s α for this dimension is .90.
The third and last quality dimension is ‘knowledge of needs’. This quality dimension only contains two items (“The appearance of the physical facilities of the CvR is in keeping with the type of services provided” and “Employees of the CvR do not know what your needs are”). Therefore, the Pearson correlation is .40 (p<.01).
Finally, all the four questions meant to measure satisfaction can be used. The Cronbach’s α for patient satisfaction is .91. Results of both the factor analysis of the patient dataset can be found in appendix 1.
Quality dimensions patients Cronbach’s α Pearson correlation Assurance & reliability (tangible & intangible) .93
Personal help .90
Knowledge of needs .40**
3.3 Theoretical design of the questionnaire
As the obtained data is used for a quantitative analysis, the study is a conclusive research. Furthermore, the research can be defined as descriptive, since it determines the perception of the provided service of the CvR-UMCG. The primary data is obtained by a
4. Results
The results that are found in this research are split in two parts. In the first part, defined in paragraph 4.1, the results of the quality dimensions scores, given by the stakeholders, are shown. These will form the perceptions. In the second part, paragraph 4.2, the relationships between the quality dimensions and satisfaction is given.
4.1 Descriptive results
As a result of the new formed quality dimensions, the mean scores per dimension, per group of stakeholders can be calculated. These mean scores form an illustration of the perception of the stakeholders, as it reveals how they observe the CvR-UMCG. The scores can be seen in table 6. Also, based on the mean scores, the differences per stakeholders view on patient perception can be compared to the perception of the patients’ selves.
Focussing on the first quality dimension ‘assurance & reliability’, it can be seen that employees (M= 3.71, SD= .35) and patient organizations (M= 3.91, SD= .47) give lower scores on their assessment concerning the patients’ perception then the patients’ own score (M= 3.97, SD= .85). However, only the difference between the employees’ assessment of patient perception and the patients’ own score is significant (p<.05). At the second quality dimension ‘personal help’, the patients’ perception (M= 3.84, SD= 1.02) is lower then the scores given by employees (M= 3.95, SD= .46) or the patient organizations (M= 4.16, SD= .63). Nevertheless, the differences of these scores are not significant. Concerning the third dimension ‘knowledge of needs’, the patients’ score (M= 3.95, SD= .91) is higher then the score given by employees (M= 3.52, SD= .50), which is a significant difference (p<.01). Patient organizations have the highest score on this dimension (M= 4.00, SD= .58). However, this difference is only significant towards the employees group (p<.05). The difference towards the patients is not significant. Concerning satisfaction, the patient group (M= 4.13, SD= .97) have a higher score then the employees (M= 4.08, SD= .34), although this effect is not significant. The difference between the employees and the patient organizations (M= 4.43, SD= .49) is significant (p<.05). Therefore, when assessed by the patient organizations, patients have a higher level of satisfaction compared to the assessment of the employees.
These scores lead to the answer to the second hypothesis. Employees have a more negative image on patients’ perception then patients selves have (H2a), when focussing on the first and third quality dimension, ‘assurance & reliability’ and ‘knowledge of needs’. For the second dimension, ‘personal help’, there is no significant difference between the groups of stakeholders. Although there is found significant differences between employees and patient organizations, the significance lacks when patients and patient organizations are compared (H2c). As general practitioners (H2b) and health insurers (H2d) did not participate in the research, the concerning hypotheses could not be tested. These findings can be summarized in the following listing:
Variable Patients1 Employees2 Patient Organisations3
Means (standard deviation in parenthesis)
Assurance & reliability 3.97 (.85)a 3.71 (.35)a 3.91 (.47) (tangible & intangible)
Personal help 3.84 (1.02) 3.95 (.46) 4.16 (.63) Knowledge of needs 3.95 (.91)b 3.52 (.50)b, c 4.00 (.58)c Satisfaction 4.13 (.97) 4.08 (.34)d 4.43 (.49)d
Means (standard deviation in parenthesis)
Assurance & reliability (tangible & intangible)
Dependable organization 4.07 (1.14) 4.17 (.57) 4.29 (.49)
Service at time 4.00 (1.08) 3.38 (.65) 3.57 (1.13)
Employees well dressed 4.39 (.86) 4.04 (.46) 4.14 (.69)
Keeping promise 3.87 (1.11) 3.42 (.83) 3.86 (.90) Trust employees * 4.09 (1.03) 4.17 (.38) 4.43 (.79) Visually appealing 3.85 (1.19) 3.29 (.86) 3.71 (.49) Safe feeling 4.07 (1.12) 4.25 (.44) 4.29 (.76) Accurate records 3.83 (1.12) 3.58 (.65) 3.71 (.76) CvR support employees ** 3.72 (1.12) 3.62 (.65) 3.43 (.54) Up-to-date equipment 3.85 (1.15) 3.13 (.68) 3.71 (.49) Personal help Personal attention 4.11 (1.16) 4.71 (.46) 4.57 (.54) Individual attention 3.91 (1.33) 4.29 (.69) 4.43 (.79) Willing to help 3.85 (1.40) 4.29 (.55) 4.29 (.76)
Convenient operating hours 3.67 (1.25 3.21 (.93) 3.86 (.90)
Providing prompt services 3.63 (1.34) 3.38 (1.14) 3.86 (1.07)
Best interest at heart * 4.35 (1.16) 4.38 (.71) 4.71 (.49)
Responding patient request ** 3.39 (1.36) 3.38 (.88) 3.43 (1.40)
Knowledge of needs
Facilities equal to service ** 3.93 (1.00) 3.00 (.66) 3.71 (.76)
Employees know needs * 3.96 (1.17) 4.04 (.55) 4.29 (.49)
Satisfaction
Happy to go to CvR 4.13 (1.00) 4.21 (.42) 4.43 (.54)
Sure it was right 4.33 (1.10) 4.12 (.45) 4.57 (.54)
Experience was good ** 3.89 (1.08) 3.88 (.34) 3.86 (1.07)
Visit worked out well * 4.17 (1.18) 4.12 (.45) 4.86 (.39)
Table 6: Mean and standard deviation of the patients’ quality dimensions, per group of stakeholder, and per item. Items ranked according to factor analysis.
1
n=46 2 n=24 3 n=7
a= Significant difference between groups (p<.05) b= Significant difference between groups (p<.01) c= Significant difference between groups (p<.05 d= Significant difference between groups (p<.05) * Highest overall score of quality dimension (or of satisfaction) of all the stakeholders
Table 6 can be summarized as:
• Patients have a higher score then employees on ‘assurance & reliability ’and ‘personal help’.
• Patient organizations have a higher score then employees on ‘knowledge of needs’ and satisfaction.
• In ‘assurance & reliability’, “trust employees” has the highest overall score, and “CvR support employees” the lowest.
• In ‘personal help’, “best interest at heart” has the highest overall score, and “responding patient request” the lowest.
• In ‘knowledge of needs’, “employees know needs” has the highest overall score, and “facilities equal to service” the lowest.
The second part of table 6 shows the individual item scores. It is ordered by the quality dimension which it is part of, and ranked by the factor loading (see appendix 1). It reveals the individual share of the item towards the dimension. When focussing on ‘assurance & reliability’, Q14 (You can trust employees of the CvR) has the highest overall score (M= 4.14, SD= .85), measured for all the stakeholders groups. Q17 (Employees get adequate support from the CvR to do their jobs well)has the lowest overall score (M= 3.66, SD= .87). For ‘personal help’ Q21 (recoded into: The CvR does have your best interest at heart) has the highest overall score (M= 4.39, SD= .99), and Q13 (recoded into: Employees of the CvR are not to busy to respond to patient requests promptly) has the lowest overall score (M= 3.39, SD= 1.22). ‘Knowledge of needs’ only exists out of two items, where Q20 (recoded into; Employees of the CvR know what your needs are) has a higher overall score (M= 4.01, SD= .97) then Q04 (The appearance of the physical facilities of the CvR is in keeping with the type of services provided) (M= 3.62, SD= .97).
Besides the overall scores, it is even more interesting to see which items have the highest scores, divided per stakeholder group. Concerning the patient group in the first dimension ‘assurance & reliability’, it can be seen that Q03 (The CvR’s employees are well dressed and appear neat) has the highest score (M= 4.39, SD= .86), and Q17 (Employees get adequate support from the CvR to do their jobs well) has the lowest score (M= 3.72, SD= 1.12). In the dimension ‘personal help’, Q21 (recoded into: The CvR has your best interest at heart) has the highest score (M=4.35, SD= 1.16) and Q13 (recoded into: Employees of the CvR are not to busy to respond to patient requests promptly) has the lowest score (M= 3.39, SD= 1.36). In the last dimension ‘knowledge of needs’ the deviation is equal to the overall highest and lowest score of this dimension, where Q20 (Recoded into:”Employees of the CvR know what your needs are) has the highest score, and Q04 (“The appearance of the physical facilities is in keeping with the type of service provided) the lowest. Concerning satisfaction, patients gave the highest score to Q02 ((I am happy that I decided to go to the CvR) (M= 4.33, SD= 1.10) and the lowest score to Q03 (I am sure it was the right thing to go to the CvR)(M= 3.89, SD= 1.08), which is equal to the overall lowest score for the items that measure satisfaction. A summary of the other highest and lowest scores per group of stakeholder can be found in the second part of table 6.
when the patients become older. Again, these are not significant differences. Finally, besides the difference between out- and inpatients concerning ‘personal help’, the inpatients have higher scores on the quality dimensions then the outpatients, as well as on satisfaction. Nevertheless, due to the lack of significance of the differences, there can not be made any distinctions between the groups. The individual item scores are almost all in accordance with the overall item score of the patient group. The groups that have items that are not equal to the lowest overall patient item scores of ‘assurance & reliability, which is the item “CvR supports employees, are females, the age group between 18 and 39, above 60, and the inpatients, as they have different items with the lowest score in this dimension. The items “Employees know needs” and “Facilities equal to service”, which form ‘knowledge of needs’, differ in highest or lowest score between the different patient groups. “Employees know needs” has the highest scores for the groups containing males, age 18 to 39 and 40 to 59, and outpatients, where “Facilities equal to service” has the highest scores for the other groups. Contrary it is the case for the lowest scores. These and complementary results can be found in appendix 4.
4.2 Explorative results
Besides the respondents’ mean scores, the relationship between several stakeholders’ quality dimensions and their view on patient satisfaction can be calculated. This relationship consists out of the correlation and the regression. The level of correlation indicates the strength and direction of the relationship between the individual quality dimensions and satisfaction of different stakeholders. The regression explains how the typical value of satisfaction changes when one of the quality dimensions is changed, while the other quality dimensions are fixed.
Concerning patients, two dimensions have a significant correlation with satisfaction. ‘Assurance & reliability’ (r= .84, p<.01) and ‘knowledge of needs’ (r= .29, p<.05) both have this significant relation. Within the group of employees, both ‘assurance & reliability’ (r= .61, p= <.01) and ‘personal help’ (r= .56, p<.05) have a significant correlation with satisfaction. The patient organizations’ quality dimensions have no significant correlation with satisfaction. These findings, and those of the regression analyses, can be found in table 7.
Dimension Patients1 Employees2 Patient
Organisations3 Correlations
Assurance & reliability .84** .61** -.10 (tangible & intangible)
Personal help .04 .56** .43
Knowledge of needs .29* .25 .22 Regression
Assurance & reliability .87**a,b .43a,d -.85b,d (tangible & intangible)
Personal help -.06c .24e .99 c,e Knowledge of needs -.05 .11 .07
Table 7: Correlation and regression of the patients’ quality dimensions, per group of stakeholder.
Table 7 can be summarized as:
• In ‘assurance & reliability’, the patients have a stronger correlation to satisfaction then the employees.
• ‘Assurance & reliability’ has the only significant regression.
• Patients have a higher regression score on ‘assurance & reliability’ then employees or patient organizations.
• There is a low degree of congruence between patient organizations, employees, and patients on both ‘assurance & reliability’ and ‘personal help’.
Subsequently, the regression analyses have revealed the degree of influence of the dimensions on satisfaction. The only positive and significant influence within the patient group is the dimension ‘assurance & reliability’ (b= .87, p<.01). None of the other stakeholders groups’ quality dimensions have a significant influence (p>.05). Despite the lack of significant regressions, there are several significant differences between the regressions of the diverse groups, which also can be found in table 7, indicated by the characters. Based on these found significant differences, it can be claimed that ‘assurance & reliability’ has more effect on satisfaction for patients, then for employees’ assessment on the patients’ perception of this effect (p<.05). The difference between the effects of patients’ and patient organizations’ quality dimension ‘assurance & reliability’ also is significant (p <.01). For this reason it can be said that ‘assurance & reliability’ has more effect for patients then for patient organizations. Next, concerning the difference between the regressions of the quality dimension ‘personal help’, there is a significant difference between patients and patient organizations (p<.01). For this reason, it can be said that ‘personal help’ has more effect when assessed by patient organizations then the effect of patients themselves. Other significant differences between employees and patient organizations can be found in the dimension ‘assurance & reliability’, where the effect assessed by employees is greater then the effect assessed by patient organizations (p<.01), and in the dimension ‘personal help’, where the effect assessed by patient organizations is greater then the effect assessed by employees (p<.05).
With these findings, an answer to the first and third hypothesis can be formed. For the reason that the dimension ‘assurance & reliability’ has a positive and significant correlation and regression, it can be said that the quality dimension ‘assurance & reliability’ has a positive influence on satisfaction (H1). Furthermore, as mentioned, the effect, exhibited by the regression, of ‘assurance & reliability’ is greater for patients then how employees asses this effect for patients. This means that as the functional quality dimension ‘assurance & reliability’ rises, this has a more positive effect on patients’ satisfaction then what employees think that the effect for patients is on satisfaction. This proofs that the quality dimension ‘assurance & reliability’ is more important for patients then for employees (H3a). Just as in the second hypothesis, the general practitioners have not participated, so an answer to the second part of the hypothesis (H3b) could not be formed. Briefly, the hypotheses are answered as:
• H1: partially supported • H3a: partially supported • H3b: not supported
inpatients (r= .88, p<.01) emphasize the dominating influence of this quality dimension. ‘Personal help’ only has a significant correlation for the age group between 18 and 39 (r= .68, p<.05), and ‘knowledge of needs’ only has a significant correlation for males (r= . 45, p<.05) and for the age group between 40 and 59 (r= .55, p<.05). When focusing on the effects of the relationships between the three quality dimensions and satisfaction, again the quality dimension ‘assurance & reliability’ attracts the attention. This dimension has a significant effect on males (b= .73, p<.01), females (b= 1.06, p<.01), age 40-59 (b= .73, p<.01), age 60 and older (b= 1.01, p<.01), outpatients (b= .82, p<.01) and inpatients (b= .91, p<.01). Only the youngest age group (18-39) has no significant regression with satisfaction. The quality dimension ‘personal help’ only has one significant regression with satisfaction within the group of female patients (r= .18, p<.05), and the dimension ‘knowledge of needs’ has two significant effects on satisfaction, for females (b= -.33, p<.05) and for patients that are 60 years or older (b= -.27, p<.05). However, as the last three mentioned regressions do not have a significant correlation, it is not known what the direction of this effect is. These results can be found in appendix 4. Based on these found regressions, some conclusions can be made. Females have a higher score on regression towards ‘assurance & reliability’ then males. For this same dimension, the oldest patient group has a higher score on regression then the middle aged group. Next, inpatients have a higher score on regression for ‘assurance & reliability’ then outpatients. However, none of these groups have a significant difference between the scores. Therefore, it can not be said that, for example, the effect of ‘assurance & reliability’ on satisfaction is stronger for females then for males. Furthermore, this means that several groups within the patient group can not be compared on the effect of the quality dimensions towards the level of satisfaction. As both parts of the fourth hypothesis (H4a and H4b) are based on the comparison between the different patient groups, these hypotheses can not be supported.
• H4a: not supported • H4b: not supported
Table 8: Means (M) of patient scores, regression (b) and correlation (r) per item in relationship to patient satisfaction, divided per quality dimension. Items ranked by factor loading.
*p<.05 **p<.01
Dimension M r b
Assurance & reliability (tangible & intangible)
Dependable organization 4.07 .80** .19
Service at time 4.00 .68** .22
Employees well dressed 4.39 .68** .37*
Keeping promise 3.87 .69** -.14 Trust employees 4.09 .67** -.11 Visually appealing 3.85 .54** -.20 Safe feeling 4.07 .60** -.18 Accurate records 3.83 .78** .08 CvR support employees 3.72 .66** -.08 Up-to-date equipment 3.85 .71** .51** Personal help Personal attention 4.11 .30* .36* Individual attention 3.91 .15 -.05 Willing to help 3.85 .11 -.18
Convenient operating hours 3.67 -.10 .01
Providing prompt services 3.63 .16 -.14
Best interest at heart 4.35 -.10 -.01
Responding patient request 3.39 -.08 .02
Knowledge of needs
