Behavioral Intention In Revisiting Hospital Under The Effect Of Expertise, Reputation And Service Quality

Behavioral Intention In Revisiting Hospital Under The Effect Of Expertise, Reputation And Service Quality Pham Bao Duy International University, Vietnam National University HCMC, Vietnam Nguyen Tan Loi Eastern International University, Vietnam Ho Nhut Quang International University, Vietnam National University HCMC, Vietnam Abstract The well-noted extensive solution for strengthening the hospital’s obstacles from various perspectives are seeking in Vietnam context Considering the foothold of medical industry formulated by government, the medical industry grants as sustainable fundamental development observed through FDI and governmental equity with shaky restriction As strikingly demands in healthcare services, private and public hospitals in Vietnam, however, divulge the noticeable missing pieces are service quality, trust and satisfaction in arousing rehash visitors whom reveal the disavowal through pursuing highly experience expectation in comprehensive alternatives Consequently, the deeply understanding in the rehash patients’ behavioral intention, especially, with further with appealing unfamiliar one is envisaged vital pharmaceutical to intensify specifically hospitals’ aspects Methodically, quantitative research spread out the study in advance, the sample size accounted at 316 processing in Explanatory Factors Analysis and deeply exploiting by Structural Equation Modeling method The outcomes spotlight the significant initial effect of independent dimensions in reputation and expertise toward trust, service quality toward satisfaction as well as towards behavioral intention in revisiting the hospital Undoubtedly, there are some recommendations toughen up specifically the retained problems for hospitals in Vietnam Keywords: Vietnam Healthcare, Service Quality, Expertise, Behavioral Intention in Revisiting, Hospital in Vietnam, Sustainable Development Introduction Healthcare has long been the initial service for human, which provides significant purposes in examination, prediction, treatment and control the health According to the former researches, it is considered as the high level of association services (Hogg, Laing, & Newholm, 2004) In some emergency situation, the hospital is not only considered as the place for health improvement and examination but also become the crucial place for saving a life Therefore, the high level in emotional vulnerable following with the hazard is not deniable (Jadad, 1998) According to General Statistic Office in Vietnam, there are 1,101 hospitals in 2012 and higher 36% compared with the statistics in 2007 In addition, the number and scale of corporation and organization operating in medical and health care field such as Hoan My, VinMec and TMMC has been increased in recent years It 621 provides the proof that the importance and attention of medical industry in Vietnam However, with the interesting figures demonstrated in the recent report, the fundamental right – basic healthcare would be not considered as the right in developing countries included Vietnam where people from rural areas, even the government spend hundreds of million dollars in medical industries which mostly focus on facilities’ expansion and improvement in suburban In other sides, according to Vietnam 2035 general report has been posted by World Bank and Ministry of Planning and Investment in Vietnam, it shows directly that the proportion of GDP shares 6% for medical and health care over decades from 2015 to 2035 While GDP and the growth rate have been reported outstanding positively increasing whereas the out-of-pocket spending in medical still appeared with 49%, it shows the importance of medical fields among people even the total expenditure must be paid by their money According to WHO, the bed occupancy rate should not be over 80% of the hospitals’ capacity Vietnam, however, witnessed a high occupancy rate especially in large cities A research of “Study on Current Situation of Overcrowding, Under-Crowding in Hospitals at Levels and Recommended Solutions for Improvement” of Ministry of Health in 2011 demonstrates that central and provincial hospitals always in overcrowding situation It is in the ranged from 120-150% of the hospital beds used and some special cases over 200% in big central hospital in Ho Chi Minh city and Ha Noi capital According to the research of PricewaterhouseCoopers (Vietnam) Ltd Company called “The Vietnamese healthcare industry: moving to next level”, it mentioned the overload services in large and popular hospitals in Vietnam while remote area hospital and regional polyclinic in suburban is lack of patient Moreover, it agreed that Vietnamese tends to approach national hospital instead of provincial hospital due to the mindset in lack of quality in medical staff and medical equipment Following the article “The poor still miss out on healthcare in Vietnam” published in 2015 on Joint Learning Network – The global community of health systems practitioners and policymakers in 27 countries including Vietnam, it stated that “Health Care Fund for the Poor” is signed in 2002 under the Decision 139 signed by Prime Minister of Vietnam in order to support the healthcare service Vietnam especially people live in communes However, this program was not appreciated by World Bank, the research “Health Insurance for the Poor: Initial Impacts of Vietnam’s Health Care Fund for the Poor” found that the reduction of out-of-pocket spending has not happened According to a recent survey by the Ministry of Health, every year around 40,000 Vietnamese go abroad for health treatment purpose and spend $1 billion in 2010 and nearly $2 billion for total expenditures in early 2016 It raises the problem that whether hospitals in Vietnam still not meet the needs of domestic demand in health care purpose even 6% of GDP in Vietnam spend for medical as mentioned above In addition, in Vietnam, the average people make the out-of-pocket payment for health care is almost accounted for 75% the total spending in health care (Knowles et al 2005) According to the research of Gludner and Rifkin in 1993, private healthcare provider appealing the demand of people who have intend in healthcare service as the public service is deficient and imperfect, the case of Vietnam and Uganda Literature Reviews 2.1 Service Quality Service quality is a measurement on the matching between services delivered and customer’s expectation Service quality must be delivered that match with customer expectation on a reliable basic (Lewis and Booms, 1983) Unfortunately, the evaluation for service quality in the scale that built-up through previous study for healthcare industry has been on the rocks Instead of the value comes from the result of health care, patients not completely evaluate the comprehensive problem in service quality through their perspective In addition, some sectors found many difficulties to determine whether it will be added on the service quality 622 assessment or not such as the emergency affected to the probability in survival or vegetable existence, the question is not able to figure out people who responsible to assess the evaluation In addition, the lack of skill also expertise to define the service whether it was conducted following the process or not (Newcome, 1997; Williams, 1994) As a result, hospitals take their advantages in the evaluation of patient with the misleading in technical service quality aspect (Bowers et al., 1994; Ettinger, 1998; Donabedian, 1988), focus on the interaction between patients and physicians and approach with potential customer with the misleading of former evaluation According to Bowers et al in 1994, they suggest that the scale of service quality in patient’s determinant take important role in their satisfaction through SERVQUAL model Before the Bower and his partners’ findings, a former study also used SERVQUAL in implication in the antecedent of service quality in satisfying patient (Reidenbach and Sandifer-Smallwood, 1990) It is explained from another study in health care service industry that “As a construct, customer satisfaction has been noted as a special form of consumer attitude; it is a post-purchase phenomenon reflecting how much the consumer likes or dislikes the service after experiencing it” (Woodside, AG Frey, LL and Daly, RT., 1989) It comes to the first hypothesis: Hypothesis A service quality of hospital is positively related to patient’s (customer’s) satisfaction In health care research, SERVQUAL scale is the precursor model for evaluating the outcomes behavioral intention comes from service quality (Reidenbach and Sandifer-Smallwood, 1990), and other variant model with the same result, for example, Headley and Miller developed 6-dimensional based on primitive SERVQUAL model in 1990 It can be seen obviously that the service quality is a significant dimension not only satisfy customer but also attract customer in repurchasing service or product Hypothesis A service quality of hospital is positively related to patient’s (customer’s) behavioral intention in revisiting hospital 2.2 Satisfaction In 1980, Oliver built the definition that “In brief, customer satisfaction is a summary cognitive and affective reaction to a service incident (or sometimes to a long-term service relationship) Satisfaction (or dissatisfaction) results from experiencing a service quality encounter and comparing that encounter with what was expected” To analyze the level of satisfaction that customer measured based on service, product that provided by an organization through figures based on questionnaires and feedback from the frontline staff It could be the positive judgments’ outcome from using a product or service from customer perspectives (Westbrook, 1980) Related to the definition, it suggests that satisfaction is the emotional evaluation, it is a chain of individuals’ assessment rather than an individual perspective (Cronin and Taylor, 1994; Hunt, 1977) The scale of satisfaction is defined from dissatisfying to satisfying where other arguments implicate that the customer satisfaction assessment proves a comprehensive evaluation than the specific outcome of a transaction According to Singh and Sirdeshrnukh (2000), customer’s experiences is defined as the directly evaluation on some cues which included satisfaction Based on implicit and explicit cues, customer can gradually formulate the trustworthiness with firm (Doney and Cannon 1997) If build up a strongly satisfaction from customer, customer may have more confidence with the firm, which is the basic for increasing their trust on service provider Thus, Hypothesis A patient’s (customer’s) satisfaction is positively related to their trust in hospital Satisfaction is the factor that combine many antecedent elements, when customer’s satisfaction increased, it leads to the last variable, repurchasing intention or it can be considered as sub-dimensions of customer loyalty (Kitapci, Akdogan, & Dortyol, 2014) In medical industry, there are varied study that mentioned this relationship which shows the impact of satisfaction on behavioral intention (Anderson and Sullivan, 1993; Bitner, 1990; Reichheld, 1996; Woodside and Shinn, 1988; Woodside et al., 1989) Considering customer’s satisfaction as the intermediate variable, majority of studies suggests that there were the indirect influences 623 between the behavioral intention and service quality where using value and satisfaction as the mediate factor (e.g., Anderson and Sullivan, 1993; Gotlieb, Grewal, and Brown, 1994; Patterson and Spreng, 1997; Roest and Pieters, 1997; Taylor, 1997) Hence, Hypothesis A patient (customer’s) satisfaction is positively related to their behavioral intention in revisiting hospital 2.3 Trust Trust comes from the belief of a party’s promise or sentence is reliable and the obligation that party need to be fulfilled in vice versa for relationship purpose (Schurr and Ozanne, 1985) Based on the trust, the interaction of a buyer’s perception future and service provider (seller) is anticipated (Doney and Cannon, 1997) It creates a long-term orientation of a relationship B2C in positive ways (Ganesan, 1994) The trust’s advantages which create strong relationship in business has been researched in the literature review of Morgan and Hunt in 1994 The individual experience is considered as the trustworthy source rather than the referral from relatives or friends which is explained as the second-hand trust referral or the popular Building trust efforts is core value of all business in general and hospital service in specific, the results from this long journey is the substantial development where customer loyalty and attraction are not deniable There are some evidences show the behavioral intention in repurchasing services, products are the origin of trust (Morgan and Hunt, 1994; Chaudhuri and Holbrook, 2001) As trust shows confidence in looking for new customer as the reliability and integrity has been prepared, it is the main component for long-term relationship orientation as it moves the focus in present to continuity and future conditions (Doney and Cannon, 1997; Ganesan, 1994) Therefore, it results in a hypothesis that: Hypothesis A patient (customer’s) trust is positively related to their behavioral intention in revisiting hospital 2.4 Expertise Knowledge and experience of service providers in the main services are two terms that typically measure in expertise (Crosby el at., 1990) In Medicine and Surgery perspectives, the expertise requires a mastery in relevant skills also the diversity of knowledge in many aspects Unlike other fields, physicians require the diverse knowledge such as biology, chemistry, physics as the basement and up-to-date their specialization that they pursuit from the beginning Besides, ethics, cognitive and motor must be consistent interpersonally according to their leaning in behavior and responsibility Moreover, clinicians require higher level in their enormous knowledge not only in their specialization but also conduct the relevant field from pharmacist to the surgeon Considering medical diagnosis is the general skill of the physicians, the expertise of the doctors is defined through the accuracy of medical diagnosis because the combination of higher experience and knowledge are deeply and varied (Feltovich et al., 1984; Neufeld et al., 1981) A study found that the source of credibility and trustworthiness is the results of individual’s perception on level of expertise, it implicates a positively effects on trust (Busch and Wilson, 1976) In other words, the level of experts creates the trust’s foundation According to the research of Crosby, Evans and Cowles in 1990, trust signal was founded from the expertise’s perception of customer It can be related to the trustworthy company where the appearance of relationship between expertise and trust create positively effects (Newell and Goldsmith, 2001) In specific of hospital service, the expertise is the undeniable role which contribute to the decision and recommendation on customer’s health The enhancements in trust are depending in the major of the expertise which provide the skilled-set learned from the perennial experience and qualifications or highly achievements in their professional career Therefore, Hypothesis A worker’s expertise in hospital is positively related to patient’s (customer’s) trust 624 2.5 Reputation The customer’s belief and trust that the firm is truthful and equitable is defined as firm reputation (Doney and Cannon, 1997) In widely views, it is a general overview measurement of a corporate or a firm in level whether it is “good” or “bad” (Weiss, Anderson, & MacInnis, 1999; Roberts & Dowling, 2002) In the sense of reliability, reputation is defined as the collective opinions which evaluate positively the trustworthiness and it results in the individual’s perspective in what they believed or positive said about the firm’s character (Freeman, 1979) Hospital’s reputation could be directly affected by concrete financial problems, even the professional pride is highly attracted by a motivating factor In specific, the sponsor and investment from corporate and individuals are founded as the huge amount to maintain the operation Especially in human health service sector, it is necessary to concentrate on the corporate reputation due to the dense of customer relation which is the most problematic affected customer perspectives to the hospital (Chase, 1978) In previous research, it implicated that the customer’s evaluation on reputation of a service provider will positively impact on the acknowledgement on firm’s trustworthiness through information transference process (Doney and Canon, 1997) A study of Devon Johnson and Kent Grayson in 2005 suggested that firm reputation is the antecedent of both affective and cognitive trust, “customer who is not yet sufficiently familiar with a service provider may extrapolate his/her opinions directly from the reputation of the firm” Hence, the hypothesis is built, Hypothesis A hospital’s reputation is positively related to patient’s (customer’s) trust 2.6 Behavioral Intention The decision that intend to perform in a specific way is considered as intention (Fishbein and Ajzen, 1975) A person who have their subjective perception ability that he or she will enjoy in a given behavior is defined as behavioral intention (Committee on Communication for Behavior Change in the 21st Century, 2002) In other way, it can be the level that a person has built self-conscious intention to engage or not engage with some specified future behavior, it is a signal about the customer future’s behaviors (Venkatesh et.al., 2003; Lai and Chen, 2011) Through previous research, it was definite to believe the important role of conceptual framework in study The model was prompted and changed by related empirical studies in a health care service provider which can apply in Vietnam context 625 Service Quality H1(+) H5(+) Satisfaction Behavioral Intention H6(+) H4(+) H2(+) Expertise Trust H7(+) H3(+) Reputation Figure 6: Conceptual Model Methodology 3.1 Research Method Qualitative and quantitative methods are considered two main method in processing the research for various purpose, especially in achieving knowledge from the study (Ritchie and O’Connor, 2004) The quantitative research is supported from the statistics where it can retrieve from the primary or the secondary data such as survey, questionnaires or previous data Meanwhile, the qualitive research method based on the evaluation of themes which is retrieved from the observation or interview In other words, it is the unmathematical method In 2001, Soguno suggested that the objectives of the study were able to express clearly which could bring back to the society the enhancement in general views in many aspects also the effect of each other In this study, quantitative research was selected to go further With the same goals that qualitative research delivering, the goals focus on the solution, recommendation based on society problems, concerns or the supporting in further research for other developing quantitative potential approach Besides, the quantitative methods delivered a deeper insight or different aspects the problem that the study concerns which support for sociologist or the experts in various industry In other words, it could not be rejected that it provided the comprehensive conclusion and recommendation for social problems or concerns especially Meanwhile, it went further with other research that give a deeper knowledge in the phenomena following the research of Strauss and Corbin; Lundahl and Skärvad in 1998 and 1999 respectively It can be pointed out the common collecting data in quantitative method such as deliver survey through paper form, online form, telephone interview or face-to-face interview Moreover, the collection can be assessed through email, pop-ups website ads In other words, there are various ways to conduct the data for quantitative research On other views, due to the various ways in collecting data, it tended to apply popular with a significant sample size at the short period comparing to the qualitative research method 626 3.2 Subject The data was collected through paper research in site location, no any online form is accepted on this research due to integrity and reliability of the research also the characteristics of the dependent variables which focus on the patients have already visited one of among four hospitals mentioned below Vinmec Central Park International Hospital which is accounted for 178 operating beds with the mission of “Delivering first-class healthcare service”, Vinmec is considered as the hospital have the best hospital’s service quality in Vietnam Secondly, Hoan My hospital which is belonged to Hoan My Medical Corporation In Vietnam, Hoan My is leading the private healthcare network where 13 hospitals and clinics has been established since 1999, there are 808 doctors and 3918 full-time employees are working in among these hospitals Also, it accounted to 2399 beds are in operation and the limitation over 3407 beds In Ho Chi Minh city, Hoan My Hospital was allocated in Phu Nhuan district which serves daily over 2,000 patients with the capacity up to 261 beds Thirdly, HCMC University Medical Clinic which was one of the most trustful destination of the patient where it serves over million of patient annually With centers were allocated in around the city, they proudly to be one of the largest operating beds capacity in Ho Chi Minh city at over 1000 operating beds Moreover, the hospital is well-known as the medical center having more than 100 professor and vice professor involved in medical examination and treatment Finally, 175 Military Hospital which is considered as the big central hospital in Vietnam that apply modern technique in healthcare process, 175 Military Hospital is highly appreciated from the respect of doctor whom most of them begin their career in military where ethics and behaviors is strictly applied In 2018, the hospital is expected to upgrade the operating bed scales up to 1,500 after the newest building is delivered in operation There are 350 questionnaires had been delivered to the patients, and there are 316 valid respondents accounted at 90.3% The sample size included age range from 18 to over 55 where the outcome placed mostly in the 26-35 age group Over 316 valid respondents, 52.85% of them are female accounted for 167 people Patients in the final sample focused on the income level from 10 to 20 million VND at 42.72% 3.3 Sample Size According to the research of Gorsuch; Hatcher in 1983 and 1994 respectively, the ratio should be allocated in 1:5 where items is answered by respondents in EFA analysis which is also tested in this study for further validating In other words, with 40 items, the samples size is required to approach at 200 units Moreover, supporting from the research of Comfrey and Lee in 1992, he found that by assessing the sample size higher than 300 units, it can result in the great research outcomes, meanwhile, the lower one is not quite appreciated Hence, it comes to the decision that the valid sample size must be 316 Some data was eliminated due to the invalid data, the paper survey must be higher than the number mentioned above 3.4 Measures In scaled question, it cannot be denied that the Likert Scale is the most appropriate method to apply for conducting the survey purpose Which is raised and promoted through the research of Rensis Likert in 1932 However, the Liker Scale provided the various of measurement scale from 2-10, it results in the complicated decision to determine which scale is better for this research Although, most researches apply 5-point scales for assessing their data through questioner, there are some evidences show that 7-point measurement scales which could provide stronger correlations between one another items in t-test outcomes (Lewis, 1993) Applying the Theory Planned of Behavior (TPB) in 2002, Ajzen not only provided the instruction in applying the behavioral intention item in the study but also point out that 7-point scales are preferable when constructing the factors around behavioral intention Besides, the demographic data giving the general information for the receiver to have a comprehensive evaluation about the sample 627 Construct Sub-scale Service Quality of The Process Concerns (SQCP) Service Quality Service Quality of The Hospital Concerns (SQHC) Service Quality of The Doctor Concerns (SQDC) Service Quality of The Tangible Concerns (SQT) Item Measurement SQCP1 There were many signboards in hospital SQCP3 The process for taking queue number for health examination was quick and simple I did not have to wait long for medical examination from physician SQCP4 The lab test was done in a prompt way SQCP5 The payment procedure was quick and simple SQHC1 The hospital’s employee, nurses and doctors are friendly SQHC2 They are willing to help me as much as they could SQHC3 They explained medication process well SQHC4 They were really cared about my health SQCP2 SQDC2 The doctor gave an explanation sufficiently of my problem, lab test’s result and treatment process The doctor was willing to answer many questions, enough to understand everything SQDC3 The doctor made me feel comfortable SQT1 The waiting areas for examination and treatment were wide and pleasant SQT2 It was easy to access amenities (e.g., canteen, ATM) SQDC1 EXP5 The parking lots were always available for stakeholders (e.g., hospital’s employee, nurses, doctors, patients, relatives) The hospital was equipped with the latest care equipment and facilities (e.g lifts, air conditioners) The physicians were knowledgeable and highly educated The physician instructed and explained fully clear and understandable about concerns The hospital applied latest research, techniques and methods in medical examination and treatment The hospital’s employee and nurses were welleducated in responding any situation The hospital’s employee, nurses and physicians were well-known about his/ her responsibilities and obligations REP1 The hospital was highly regarded in Vietnam SQT3 SQT4 EXP1 EXP2 Expertise (Exp) EXP3 EXP4 REP4 The hospital was known as the one of the most capable hospital in Vietnam The hospital had positive posts and comments on media (e.g journals, news, television, scientific conference, social networks) My friends, families and relatives positively know about this hospital TRS1 I had no reason to doubt about physician’s advice REP2 REP REP3 TRS TRS2 TRS3 I absolutely believed in the lab test and examination’s result I would feel a sense of improper result if I used treatment and examinations of other hospitals 628 SAS2 I feel the physician understand and respond caringly and specifically my condition I feel more comfortable and safe when I was taken care of nurses in this hospital I satisfied about facilities and equipment in the hospital I satisfied about the hospital’s staff behavior (e.g doctors, nurses, employee) SAS3 I satisfied about the hospital’s expertise SAS4 I satisfied with my experience in treatment and examination that I received in the hospital SAS5 I satisfied with my decision to choose the hospital TRS4 TRS5 SAS1 SAS BEI1 BEI2 BEI BEI3 BEI4 BEI5 I will recommend friends; family members and relatives should use service in this hospital If I needed medical services in the future, I would consider this hospital as my first choice I will tell other people good things about this hospital I not care about the distance between my home and this hospital if I needed medical services in the future Even price increased, I still choose this hospital as the primary medical services Table 16: Measurement Scales 3.5 Process of Data Analysis In this study, AMOS (version 20.0) and SPSS (version 21.0) are two tool-kit has been applied for extracted the raw data collected from the respondents through the survey on-site Particularly, SPSS support in figure out the descriptive statistics not only for demographic data but also for the scaled questions For ensuring the validity and reliability of the study also the model, this research is verified from SPSS to AMOS In specific, the reliability of this research would be tested through the Cronbach’s Alpha and Corrected Item – Total Correlation Before moving the model and data to AMOS, the testing in EFA is required for validation purpose where it test the validation of items and group of items called factors in Promax methods, this output was used as the default model purpose in Pattern Matrix Model Builder of AMOS’s plugin The default model is built to understand whether the validation and reliability is accepted on this first testing in the AMOS – called CFA, also checking the application of model in population between factors including observed and latent After conducting the CFA, the model was formulated in SEM where the test is conducted for checking the complicated relationship among unobserved and latent variables Data Analysis and Results 4.1 Sample Characteristics Following the methodology, the survey has been delivered on-site among hospitals mentioned above As online survey has not been used, the paper surveys have been transferred to Excel where allow writer to analyze the data Unfortunately, there are 316 among 350 surveys have been conducted is valid Invalid questionnaires are come from mostly people misunderstanding in scaling question also the routine in not 629 following the instruction has been implemented before conducting In short, 90.3% data is analyzed in deeper testing FREQUENCY 316 34 VALID INVALID PROPORTION 90.3% 9.7% Table 17: Response Rate It can be observed directly that the proportion of Female is higher than Male at nearly 52.85% while another computed at 47.15% Particularly, it can be seen from the pie chart that women tend to concern about these problems rather than men Continually, it can be seen that people in the two ages group: 26-35 and 36-45 accounted for more than 55% of the respondents It means that people in those age quite concern about their health rather than other age group People with the higher age group is lower as they are unwilling to conduct the test while their relatives are responsible to handle it In general, about the income level, two those income levels group below 10 million and 10 – 20 million accounted more than 83% where half of them belongs to each other It can be explained that the vary of income level approach this study Evaluation Gender Age Income Level Criteria Male Female 18 – 25 26 – 35 36 – 45 46 – 55 Over 55 Below 10 million 10 – 20 million Over 20 million FREQUENCY 149 167 60 91 88 24 53 131 135 50 PROPORTION 47.15% 52.85% 18.99% 28.80% 27.85% 7.59% 16.77% 41.46% 42.72% 15.82% Table 18: Demographic Analysis 4.2 Preliminary Analysis It can be seen from any research that the reliability test takes an important role to evaluate whether the research can be trustable, consistent and reliable or not By considering through internal consistency measured by the Cronbach’s Alpha which is formulated and implemented by Lee Cronbach’s in 1951 Basically, most of researchers have the agreement that Cronbach’s Alpha index of the data must at least in the range between 0.6 and 0.7 which is examined as an acceptable point, also it could be pointed out through the study of Slater (1995), Peterson (1994) also Nunnally (1978) Besides, the reliability test including another tool that support in the increasing of the Cronbach’s Alpha and the reliability of the research Based on the valued in Corrected Items – Total Correlation, it can be seen clearly that whether factors affect inversely to the reliability test or not According to the research of DeVellis in 1991, the figure Corrected Item – Total Correlation have the point which is lower than 0.3 would be removed in order to increase the reliability In other meanings, Corrected Item – Total Correlation tools would be a helpful tool to erase the item that downgrade the Cronbach’s Alpha of the factor for increasing the reliability in purpose According to the output, the corrected items – total correlation of all items in variables have the lowest point at 0.565; therefore, it created a huge gap between the risk point and the research when one item among these meet the corrected items – total correlation at 0.3 From the highest to the lowest point with Cronbach’s Alpha, the Cronbach’s Alpha of Reputation, Trust, Satisfaction, Behavioral Intention, Service Quality in Convenience Concern, Expertise, Service Quality in Health Care's Provider Concern, Service Quality in Tangibles and Service Quality in Doctor’s Concern are 630 Figure 2: Measurement Standardized Modelling 632 4.3.1 Model Fit Checking In CFA analysis for model fit purpose, several figures are considered Firstly, it can be mentioned to Chisquare/df (CMIN/df) or 𝑋 /df, this evaluation support for assessing the conceptual model in detail whether it fit with the data sample that implemented before If the ratio of the degree of freedom ratio under Chi-square is placed in the range from to 1, which would be considered as the adequate fit between the collected data and the conceptual model (Carmines and McIver, 1981) By providing other updates, the ratio which is lower than and higher than is highly recommended (Carmines and Hocevar, 1985) P-value as mentioned above as the Sig is not allowed higher than 0.05 to prove the correlation among variables In addition, standardized root means square residual (SRMR) is the differences between the residual collection of data and the covariance model (Hooper et al., 2008) According to the research of Byrne, Diamantopoulos and Siguaw in 1998 and 2000 respectively, the SRMR ratio is allowed from 1.0 to and the model would be affordable model is appreciated below 0.5 Notwithstanding, the ratio as high as 0.8 would be acceptable (Hu and Bentler, 1999) Besides, the Root Mean Squared Error of Approximation or RMSEA also aims to measure how well model optimal fit to the population instead of sample sizes (Byrne, 1998) By collecting from the previous study, it can be founded that RMSEA assessment varies the evaluation based on which output is belonged to the ranges In 1996, MacCallum, Browne and Sugawara suggested that the output lower than 0.01 is considered the excellent fit while close fit is the range between 0.01 and 0.05 If any didn’t achieve higher than 0.08, it is considered as the good fit while output lower than 0.10 is briefed as mediocre fit Continuing with MacCallum and his partners’ research, other studies suggest that it would be cut-off if RMSEA is higher than 0.10 PCLOSE or close-fitting model is set-up for retest the RMSEA purpose, the assumption of null hypothesis is the RMSEA is not higher than 0.05 In other words, RMSEA must be at least containing the output with close fit Therefore, it called PCLOSE where it must be higher than 0.05 to prove that RMSEA is lower or equal 0.05 Non-normed Fit Index (NNFI) or known as Turkey-Lewis Index (TLI) in the Amos application, this is an upgrade version of NFI based on the disadvantage of NFI where it implements the model fit based on Chi-squared and df The value of NNFI are recommended higher than 0.9 for model fitting purpose (Hair et al, 1998) In addition, the Comparative Fit Index is considered as the variant of NNFI, unlike NFI it was developed for applying the model fit with the small sample size without corruption (Tabachnick and Fidell, 2007) According to Hu and his partner in 1999, they recommended CFI indicating the good model fit if CFI is higher or equal 0.95 Meanwhile, the index which is higher than 0.9 is also review as the acceptable model fit Likewise, GFI was created as an alternative for Chi-squared, based on the variances in model fit support for the population covariances It would be allowed for adjusting the GFI based on degrees of freedom and observed variables It was recommended the GFI are equal or higher than 0.95 is more appropriate for the model while 0.9 was considered as good and some cases can be extent for acceptable purpose with the index higher than 0.8 (Miles and Shevlin, 1998) while AGFI required to be higher than 80% is good fit Measurement Thresholds Current Indices Chi-square/DF (CMIN/DF) < Good; < Acceptable > 0.95: Great 0.95 – 0.9: Good 0.9 – 0.8: Sometimes Acceptable > 0.95: Great 0.95 – 0.9: Good 1.450 (Good Result) CFI (Comparative Fit Index) GFI (Goodness-of-Fit Index) TLI (Tucker Lewis Index) AGFI SRMR 0.9 – 0.8: Sometimes Acceptable ≥ 0.9 > 0.8 < 0.09 < 0.01: excellent fit 633 0.961 (Great Result) 0.864 (Acceptable Result) 0.957 (Good Result) 0.841 (Good Result) 0.0428 (Good Result) 0.038 (Close Fit) 0.01 – 0.05: close fit 0.05 – 0.80: good fit RMSEA (Root Mean Squared Error of Approximation) 0.08 – 0.10: mediocre fit >= 0.1: poor fit Table 19: Model Fit Assessment Checking the Standardized Regression Weight also extract the Average Variance Extracted (AVE) in order to build-up the foundation support the convergent validity measurement afford with the theoretical model also applying for the discriminant validity assessment in seeking the both AVEs while provide the r squared in correlation where both AVEs in the construct is suggested to be higher than the squared of the correlation estimate In addition, evaluating the Standardized Regression Weight where it is preferred to be higher than 0.5 for checking convergent validity purpose (Hair et al, 2006) 4.3.2 Reliability Checking Besides, Composite Reliability (CR) is reevaluated for CFA model fits checking reliability purpose, comparing to the Cronbach’s Alpha which is tool for assessing the reliability in EFA From AMOS results, it shows the estimation in r squared for computing the CR in the established formula that has been reminded in Fornell and Larcker in 1981 The output of CR must be equivalent to the assessment criteria of Cronbach’s Alpha where the ratio higher than 0.7 is preferable The formula below shows the instruction in calculating CR also the AVE followed the former research of Joreskog in 1971 and Fornell and Larcker in 1981 respectively Composite Reliability (CR) Formula Equation 1: Composite reliability (Joreskog 1971) 𝑝 𝐶𝑅 = (∑𝑖=1 λ)2 𝑝 𝑝 (∑𝑖=1 λ)2 + ∑𝑖=1(1 − λ2 ) Where: λ: is corresponding factor loadings on the Standardized Regression Weight 1- λ2 : is the variance’s error in the ith indicator of construct p: is the number of indicators Average Variance Extracted (AVE) Formula Equation 2: Extracted variance–VE (Fornell& Larcker 1981) ∑𝑝𝑖=1 𝜆2 𝐴𝑉𝐸 = 𝑝 ∑𝑖=1 𝜆 + ∑𝑝𝑖=1(1 − 𝜆2 ) Where: λ: is corresponding factor loadings on the Standardized Regression Weight 1- λ2 : is the variance’s error in the ith indicator of construct p: is the number of indicators The Average Variance Extracted (AVE) and the Composite Reliability (CR) has been extracted from the output of AMOS assessment through the Standardized Regression Weight table Based on the result of AVEs, there are factors have variance extracted result over the assessment criteria (>0.5) where it ranged from 0.624 to 0.741 Additionally, the second reliability assessment – composite reliability results in the good output which most of the factors having CR over 0.8 while it is considered as preferable when CR larger or equal 0.5 As can be seen from the evaluation, it cannot be denied that the model qualifying with the reliability test 4.3.3 Convergent Validity Checking 634 Estimate SAS1 SAS2 SAS3 SAS4 SAS5 SQCP1 SQCP2 SQCP3 SQCP4 SQCP5 BEI1 BEI2 BEI3 BEI4 BEI5 EXP1 EXP2 EXP3 EXP4 EXP5 REP1 REP2 REP3 REP4 TRS1 TRS2 TRS3 TRS4 TRS5 SQHC1 SQHC2 SQHC3 SQHC4 SQT1 SQT2 SQT3 SQT4 SQDC1 SQDC2 SQDC3 < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < - SAS SAS SAS SAS SAS SQCP SQCP SQCP SQCP SQCP BEI BEI BEI BEI BEI EXP EXP EXP EXP EXP REP REP REP REP TRS TRS TRS TRS TRS SQHC SQHC SQHC SQHC SQT SQT SQT SQT SQDC SQDC SQDC 0.97 0.978 0.985 0.956 0.911 0.886 1.065 0.82 0.97 0.992 0.944 0.977 0.967 0.927 0.954 1.038 0.939 1.045 0.942 0.942 0.925 1.01 0.971 0.973 1.016 0.934 0.963 0.909 0.671 1.035 0.964 S.E C.R P 0.062 0.061 0.063 0.063 15.659 16.09 15.66 15.092 *** *** *** *** 0.061 0.061 0.065 0.058 15.005 14.588 16.432 14.081 *** *** *** *** 0.06 0.061 0.061 0.061 16.102 16.293 15.516 16.071 *** *** *** *** 0.069 0.067 0.062 0.065 13.929 13.794 15.347 15.878 *** *** *** *** 0.048 0.049 0.048 19.475 21.321 19.589 *** *** *** 0.056 0.055 0.058 0.054 16.841 16.902 17.402 17.976 *** *** *** *** 0.064 0.066 0.064 15.125 15.482 14.67 *** *** *** 0.05 0.048 0.06 19.316 19.125 11.246 *** *** *** 0.078 0.075 13.301 12.856 *** *** Table 20: Regression Weights: (Group Number-Default model) 635 Label Standardized Regression Weights SAS1 SAS2 SAS3 SAS4 SAS5 SQCP1 SQCP2 SQCP3 SQCP4 SQCP5 BEI1 BEI2 BEI3 BEI4 BEI5 EXP1 EXP2 EXP3 EXP4 EXP5 REP1 REP2 REP3 REP4 TRS1 TRS2 TRS3 TRS4 TRS5 SQHC1 SQHC2 SQHC3 SQHC4 SQT1 SQT2 SQT3 SQT4 SQDC1 SQDC2 SQDC3 < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < - SAS SAS SAS SAS SAS SQCP SQCP SQCP SQCP SQCP BEI BEI BEI BEI BEI EXP EXP EXP EXP EXP REP REP REP REP TRS TRS TRS TRS TRS SQHC SQHC SQHC SQHC SQT SQT SQT SQT SQDC SQDC SQDC Estimate 0.793 0.806 0.823 0.806 0.783 0.792 0.791 0.773 0.853 0.751 0.819 0.798 0.804 0.776 0.797 0.784 0.754 0.748 0.818 0.843 0.874 0.841 0.884 0.843 0.834 0.804 0.806 0.822 0.840 0.819 0.794 0.810 0.774 0.864 0.867 0.860 0.591 0.791 0.815 0.764 Table 21: Standardized Regression Weights (Group number 1-Default model) 636 Considering data from the table 23 the p-value displays as the encryption *** which represents for the number that lower than 0.01 Hence, it must be lower than the requirement at 0.05 In addition, by focusing on the estimation of Standardized Regression Weights, all values are higher than the threshold at 0.5 4.3.4 Discriminant Validity Checking According to the research of Fornell and his partner – Larcker in 1981, the discriminant validity is assessed by computing the Average Variance Extracted (AVE) of the variables in the construct, then comparing to the squared of correlation ration or r squared (r ), both AVEs of these must be higher than the r squared, or root squared of AVEs greater than the involved correlation ratio Estimate 𝒓𝟐 Both AVEs > 𝒓𝟐 SAS < > SQCP 0.361 0.130 Valid SAS < > BEI 0.776 0.602 Valid SAS < > EXP 0.045 0.002 Valid SAS < > REP 0.464 0.215 Valid SAS < > TRS 0.614 0.377 Valid SAS < > SQHC -0.217 0.047 Valid SAS < > SQT 0.483 0.233 Valid SAS < > SQDC 0.042 0.002 Valid SQCP < > BEI 0.442 0.195 Valid SQCP < > EXP -0.080 0.006 Valid SQCP < > REP 0.289 0.084 Valid SQCP < > TRS 0.372 0.138 Valid SQCP < > SQHC -0.185 0.034 Valid SQCP < > SQT 0.277 0.077 Valid SQCP < > SQDC 0.239 0.057 Valid BEI < > EXP -0.065 0.004 Valid BEI < > REP 0.483 0.233 Valid BEI < > TRS 0.676 0.457 Valid BEI < > SQHC -0.067 0.004 Valid BEI < > SQT 0.478 0.228 Valid BEI < > SQDC 0.071 0.005 Valid EXP < > REP 0.021 0.000 Valid EXP < > TRS -0.015 0.000 Valid EXP < > SQHC -0.024 0.001 Valid EXP < > SQT 0.116 0.013 Valid EXP < > SQDC -0.079 0.006 Valid REP < > TRS 0.571 0.326 Valid REP < > SQHC -0.169 0.029 Valid REP < > SQT 0.439 0.193 Valid REP < > SQDC -0.045 0.002 Valid TRS < > SQHC -0.067 0.004 Valid TRS < > SQT 0.458 0.210 Valid TRS < > SQDC 0.197 0.039 Valid SQHC < > SQT -0.012 0.000 Valid 637 SQHC < > SQDC -0.255 0.065 Valid SQT < > SQDC 0.125 0.016 Valid Table 22: Correlations (CFA) By applying the criteria which AVEs > 𝑟 , it is founded that the outputs show the AVEs of both constructs is higher than 𝑟 in any correlation Hence, it can not be deniable that the model is valid with discriminant validity assessment 4.4 Structural Equation Modeling (SEM) In the previous section, CFA was used in the AMOS application at the first step to evaluate the correlation between the observed variables and latent variables Then, it was upgraded in SEM with the framework model where it is the extension of general linear model and allowed researcher It is considered as the powerful and argil technique to analyze the framework, also pointed out particularly the relationship between unobserved and observed items The outcomes of first SEM have been showed in the diagram below Figure 3: First Structural Equation Model 4.4.1 First SEM Model Fit Similarly, to conduct the Model Fit assessment of CFA, the results illustrate straightforwardly comparing to the criteria have been implemented in the CFA section Chi-square over degree of freedom or CMIN/df at 1.499 where it demonstrates the great outcome where it need to be lower than Goodness-of-Fit Index (GFI) 638 and Adjusted-Goodness-of-Fit Index (AGFI) reach at 0.858 and 0.837 respectively where it leads to a good outcome following the thresholds Assessing the Comparative Fit Index (CFI), the result produce positively where it shows the great outcome at 0.956 In addition, TLI is considered as the confidently model fit which reaches at 0.952 The close fit to the population is confirmed where the outcome after evaluating RMSEA where it reaches at 0.040 SRMR is calculated through plugin at 0.0541 which is qualified in the requirement of theory (0.09) To sum-up, the information will be provided in short below: Measurement Chi-square/DF (CMIN/DF) Thresholds < Good; < Acceptable > 0.95: Great CFI (Comparative Fit Index) 0.95 – 0.9: Good 0.9 – 0.8: Sometimes Acceptable > 0.95: Great 0.95 – 0.9: Good 0.9 – 0.8: Sometimes Acceptable ≥ 0.9 GFI (Goodness-of-Fit Index) TLI (Tucker Lewis Index) AGFI SRMR RMSEA (Root Mean Squared Error of Approximation) Current Indices 1.499 (Good Result) 0.956 (Great Result) 0.858 (Acceptable Result) 0.952 (Good Result) 0.837 (Good Result) 0.0541 (Good Result) > 0.8 < 0.09 < 0.01: excellent fit 0.01 – 0.05: close fit 0.05 – 0.80: good fit 0.08 – 0.10: mediocre fit 0.040 (Close Fit) >= 0.1: poor fit Table 23: Model Fit Assessment in first round Correlation Observed Variables Testing Estimate S.E C.R P SQHC < > SQT -0.014 0.076 -0.186 0.853 SQHC < > SQDC -0.246 0.067 -3.693 *** SQCP < > SQHC -0.228 0.08 -2.842 0.004 REP < > SQHC -0.234 0.091 -2.588 0.01 EXP < > SQHC -0.031 0.078 -0.392 0.695 SQT < > SQDC 0.141 0.074 1.916 0.055 SQCP < > SQDC 0.277 0.079 3.517 *** SQCP < > SQT 0.396 0.095 4.184 *** REP < > SQT 0.745 0.115 6.497 *** EXP < > SQT 0.159 0.09 1.76 0.078 REP < > SQDC -0.044 0.086 -0.514 0.607 EXP < > SQDC -0.095 0.075 -1.258 0.208 SQCP < > REP 0.513 0.112 4.568 *** SQCP < > EXP -0.126 0.093 -1.355 0.176 EXP < > REP 0.036 0.105 0.345 0.73 Table 24: Covariances (Group number 1: Default model) (First Round) 639 Label According to the figure above, p-value promote weak correlations among relationship between Service Quality in Hospital Concerns and Service Quality in Tangible Concerns; Expertise and Service Quality in Hospital Concerns; Service Quality in Tangible Concerns and Service Quality in Doctor Concerns; Expertise and Service Quality in Tangible Concerns; Reputation and Service Quality in Doctor Concerns; Expertise and Service Quality in Doctor Concerns; Service Quality in Process Concerns and Expertise; Expertise and Reputation where p-value assessments are over 0.05 Hence, those correlations should be eliminated in the second SEM Hypothesis Testing According to the Business Statistic Textbook 7th Edition, p-value is defined as the assessment whether it is considered as the statistically significant when the value is lower 0.05 (p