Factors affecting use of word-of-mouth by dental patients

Abstract

Objectives: Word-of-mouth (WOM) refers to communication among consumers, which greatly influences the marketing strategies of dental clinics. This study aimed to explore factors that affect use of WOM by dental patients and to analyse their pathways. Methods: The participants were 520 outpatients from four private dental clinics. Data were obtained from a survey using self-reported questionnaires, which included questions regarding seven latent variables: five exogenous variables, including medical service quality (physical environment, customer service, patient relationship quality) and individual characteristic variables (opinion leader tendency, social hub tendency); and two endogenous variables (intention to recommend, WOM experience). Statistical analysis was performed using structural equation modelling. Results: Significant associations were found in the pathways between relationship quality and intention to recommend, intention to recommend and WOM, and opinion leader tendency and WOM (P < 0.001). Higher patient relationship quality and higher intention to recommend were related to positive WOM, as was higher opinion leader tendency. Conclusions: Improving patient relationship quality can promote positive WOM for dental clinics. Strategies are needed to promote a positive perception of dental clinics by effectively responding to the views of patients with strong opinion leader tendencies.

Introduction

When dental patients are selecting a clinic for treatment, it is difficult for them to evaluate service quality objectively in advance because dental services are experiential. They are only able to evaluate the quality of dental services within a limited range1., 2., and thus rely heavily on information from other people [i.e. word-of-mouth (WOM) from family, friends and acquaintances] when selecting a dental clinic3., 4..

The term WOM, first introduced by Whyte in 1954⁵, refers to communication that occurs between consumers, during which information regarding a product or service is passed from one to another. In the broad sense, WOM also includes peer-to-peer communication on social networks. WOM marketing thus refers to professional marketing approaches used deliberately to influence consumer communications⁶.

The American Marketing Association defines marketing as the activity, combination of systems and processes for the creation, communication, delivery and exchange of behaviours that offer value to clients, which occur in a broad scope, ranging from the activities of not-for-profit organisations to those of corporations⁷. This dynamic, broad marketing concept can help improve the management of dental clinics by identifying the needs of patients, thus allowing them to provide high-quality patient-oriented services and to maintain and manage loyal customers⁸. Furthermore, that study noted that a positive WOM recommendation for a medical clinic can be a more powerful marketing tool than repeat patronisation⁸. In addition, Chaniotakis et al.⁹ found that the ‘empathy’ dimension of service quality in the health-care industry is directly affected by WOM, whereas tangible dimensions, such as equipment, appearance of dental staff and facilities, are indirectly affected. Therefore, to promote positive and prevent negative WOM in relation to medical facilities, which is more influential than early positive WOM, marketing specifically targeted to such settings may be necessary¹⁰.

The dental-care industry is rapidly changing and increasingly transcending borders. Given the unpredictable nature of the overall health-care industry, which is seeing the opening of markets and increasing entry of large corporations, dental clinics are striving to develop marketing strategies using a variety of methods, such as service quality improvement and relationship marketing, including WOM marketing11., 12.. However, research regarding WOM marketing in the dental clinic industry is scarce. In particular, few studies have been conducted to determine variables that influence the use of WOM by dental patients. Based on this gap, this study was designed to explore factors that affect the use of WOM by dental patients, as well as to analyse interrelationships among factors such as medical service quality (physical environment, customer service, patient relationship quality), individual characteristics (opinion leader tendency, social hub tendency), intention to recommend and experience with WOM, using structural equation modelling.

Methods

Study participants

This study was conducted by surveying patients who visited four different private dental clinics providing general dental care in a metropolitan community in the south-east area of Korea. Of 562 total questionnaires obtained, 42 were excluded because of lack of completion. Thus, the final sample size was 520 (i.e. 103, 104, 160, and 153 questionnaires from the four dental clinics).

One dental hygienist at each dental clinic was trained in the guidelines of the self-reported questionnaire before the start of the survey period. Then, in a separate room after completion of dental treatment, that hygienist then briefly explained to the study participants how to complete the self-reported questionnaire and encouraged them to answer all the questions.

This study was conducted in full accordance with the World Medical Association Declaration of Helsinki and was approved by the Institutional Review Board of Kyungpook National University Hospital (IRB- KNUH 2015-08-031). All participants were at least 19 years old and provided written informed consent before enrolment.

Survey instrument

This study utilised questions regarding 25 observed variables and seven latent (five exogenous, two endogenous) variables. The exogenous variables included three medical service quality variables (physical environment, customer service, patient relationship quality) and two individual characteristic variables (opinion leader tendency, social hub tendency), while the endogenous variables included intention to recommend and WOM experience. Intention to recommend was assessed using the Korea Net Promoter Score (KNPS)¹³.

In detail, medical service quality was measured using a 2-factor solution based on five dimensions (tangibility, reliability, responsiveness, assurance, empathy) of the Service Quality scales14., 15.. The physical environment was assessed by four observed variables, namely accessibility (two items), cleanliness (three items), aesthetic appeal (three items) and convenience (three items), while customer service was assessed by three observed variables, namely expertise (three items), reliability (three items) and responsiveness (three items). Patient relationship quality was assessed by three observed variables, including patient satisfaction (two items), patient trust (two items) and patient involvement (two items), and opinion leader tendency and social hub tendency were assessed by adopting the six observed variables from a questionnaire developed by Childers¹⁶. All questionnaire items described above were measured using a 5-point Likert scale.

Intention to recommend was assessed using the question ‘How likely are you to recommend our dental clinic to others?’ and measured on a 7-point scale according to the KNPS¹³, with higher scores indicating a greater intention to recommend. WOM experience was assessed using the question ‘Have you ever discussed our hospital with others?’ and measured on a 3-point scale (1 = said positive things about it, 2 = said negative things about it, 3 = never said anything about it).

Statistical analysis

Statistical analysis was performed according to a structural equation modelling procedure. Collected data underwent the following processes: higher-order factor analysis, confirmatory factor analysis, test of goodness of fit of the model, modification of the model, and determination of the final model. In detail, first, higher-order factor analysis was performed to build a model in which the higher-order factors included lower-order factors for a number of items for the exogenous variables. To determine the validity and reliability of the measurement instruments, confirmatory factor analysis was performed, and convergent validity and discriminant validity were assessed. Next, a test of goodness of fit was performed based on absolute and incremental fit indices; then, to improve the model fit, the initial research model was modified using a modification index. Statistical analysis was performed using AMOS 20.0 (SPSS Inc., Chicago, IL, USA).

Results

Table 1 shows the sociodemographic characteristics of the study participants, including gender, age, education level, monthly income and referral path, and highlights that the cohort was composed of a majority of women (53.5%, n = 278). Tables 2 and 3 show results of analyses of convergent validity and discriminant validity. The results regarding convergent validity of the items showed that in five constructs (lower-order concepts of physical environment, customer service, patient relationship quality, opinion leader tendency, social hub tendency), the factor loading, critical ratio (estimate/standard error, indicating the statistical significance of the path coefficient), average variance extracted (AVE) and construct reliability (CR) values were 0.5, 1.965, 0.5 and ≥0.7, respectively, suggesting high convergent validity (Table 2). In addition, the results for discriminant validity showed that in all combinations of the two constructs, the CR value was lower than the AVE value for each construct, suggesting acceptable discriminant validity (Table 3).

Table 1.

Demographic characteristics of subjects

Characteristic	n (total = 520)	%
Gender
Male	242	46.5
Female	278	53.5
Age (years)
≤29	90	17.3
30–39	103	19.8
40–49	116	22.3
50–59	123	23.7
≥60	88	16.9
Education level
Middle school or lower	44	16.2
High school	152	30.6
College	114	23.1
University or higher	210	22.0
Monthly income (USD to SKW)
≤899	84	16.2
900–1,899	159	30.6
1,900–2,899	120	23.1
2,900–3,899	71	13.7
3,900–4,899	43	8.3
≥4,900	43	8.3
Path of visit to clinic
Recommendation	426	81.9
Internet	36	6.9
Advertisement	13	2.5
Others	45	8.7

SKW, South Korean Won; USD, US dollars.

Table 2.

Convergent validity of survey instruments

Variable	Category	Factor loading	CritR	AVE	CR
Physical environment	Convenience	0.836	21.495	0.796	0.939
	Aesthetic appeal	0.800	25.620
	Cleanliness	0.901	19.651
	Accessibility	0.752	Fix.
Customer service	Responsiveness	0.895	35.052	0.912	0.969
	Reliability	0.940	30.739
	Expertise	0.891	Fix.
Patient relationship quality	Patient involvement	0.839	27.620	0.884	0.958
	Patient trust	0.912	27.957
	Patient satisfaction	0.918	Fix.
Opinion leader tendency	1. I am the first person my friends and neighbours talk to about dental care	0.756	16.512	0.597	0.912
	2. I start conversations about dental care	0.731	15.887
	3. I frequently talk to other people about dental care	0.705	13.319
	4. I inform other people about dental care	0.599	16.296
	5. When I talk about dental care, other people agree with my opinion	0.722	16.196
	6. I convince other people to visit the dental clinic that I use	0.718	15.622
	7. People to whom I have recommended a dental clinic are now treated at that clinic	0.694	Fix.
Social hub tendency	1. I very much enjoy learning new things about different people	0.768	20.347	0.719	0.939
	2. I very much enjoy meeting new people	0.842	19.527
	3. Compared with other people, I am very friendly	0.813	17.838
	4. I go out of my way to introduce people to each other	0.753	18.673
	5. I enjoy participating in small talk	0.783	19.245
	6. Compared with other people, I have friends and acquaintances	0.803	Fix.

AVE, average variance extracted; CR, construct reliability; CritR, critical ratio.

Table 3.

Discriminant validity of survey instruments

Constructs	Physical environment	Customer service	Patient relationship quality	Opinion leader tendency	Social hub tendency	AVE	CR
Physical environment	1					0.796	0.939
Customer service	0.802	1				0.912	0.969
Patient relationship quality	0.800	0.922	1			0.884	0.958
Opinion leader tendency	0.452	0.377	0.516	1		0.597	0.912
Social hub tendency	0.379	0.482	0.375	0.706	1	0.719	0.939

AVE, average variance extracted; CR, construct reliability.

Figure 1k and Table 4 show the final structural equation model and goodness of fit. For goodness-of-fit testing using absolute fit indices [chi-square (χ²), goodness-of-fit index (GFI), adjusted GFI (AGFI), root mean square residual (RMR)] and incremental fit indices [Tucker–Lewis index (TLI), normed fit index (NFI), comparative fit index (CFI)], good results were obtained for all except χ² and AGFI (Table 4). As an alternative to χ², the NC (normed χ² = χ²/df) value was calculated and was found to be 2.47, thus securing the model goodness of fit¹⁷. Therefore, we determined that the study model was appropriate for the sampled data.

Figure 1.

Path diagram illustrating the factors affecting use of word-of-mouth (WOM) by dental patients. Rectangles represent observed variables and ellipses represent latent variables. Intention to recommend and WOM are endogenous variables; all others are exogenous variables. The number e1–e25 within circles indicate measurement error of the corresponding observed variables, and D1 and D2 within circles indicate structural error of the corresponding endogenous variable. Single-head arrows between exogenous and endogenous variables indicate hypothesised causal directions. KNPS, Korea Net Promoter Score.

Table 4.

Comparisons between the study model and the optimal (modified) model

	χ2	χ2/df	GFI	AGFI	CFI	RMR	RMSEA	TLI	NFI
Standard index	P > 0.05	<3	≥0.9	≥0.9	≥0.9	≥0.05	≥0.08	≥0.9	≥0.9
Study model	1,012.47 (P = 0.000)	3.847	0.859	0.824	0.924	0.039	0.074	0.913	0.901
Modified model	701.65 (P = 0.000)	2.741	0.902	0.876	0.955	0.035	0.058	0.947	0.931

χ², chi-square statistic; χ²/df, normed χ²; AGFI, adjusted goodness of fit index; CFI, comparative fit index; GFI, goodness of fit index; NFI, normed fit index; RMR, root mean square residual; RMSEA, root mean square error of approximation; TLI, Tucker–Lewis index.

Table 5 shows results for the path coefficients between the variables in the structural equation model. Significant associations were identified in the paths between patient relationship quality and intention to recommend, between intention to recommend and WOM, and between opinion leader tendency and WOM (P < 0.000). Those standard coefficients were 0.774, 0.360 and 0.225, respectively, which indicates that higher patient relationship quality leads to higher intention to recommend, which in turn promotes positive WOM. Moreover, a higher opinion leader tendency score was also shown to promote positive WOM.

Table 5.

Path analysis values for final models

	Path			Standardised regression weights	Regression weights	SE	CR	P
H1	Physical environment	→	Intention to recommend	0.000	0.001	0.111	0.005	0.996
H2	Customer service	→	Intention to recommend	0.048	0.092	0.199	0.464	0.643
H3	Patient relationship quality	→	Intention to recommend	0.774	1.424	0.204	6.965	<0.001
H4	Opinion leader tendency	→	Intention to recommend	−0.003	−0.007	0.105	−0.062	0.950
H5	Social hub tendency	→	Intention to recommend	−0.020	−0.033	0.079	−0.415	0.678
H6	Intention to recommend	→	WOM	0.360	0.137	0.016	8.348	<0.001
H7	Opinion leader tendency	→	WOM	0.225	0.159	0.034	4.736	<0.001

CR, construct reliability; SE, standard error; WOM, word of mouth.

Discussion

The results presented indicate that the physical environment and customer service of dental clinics do not significantly influence intention to recommend, whereas patient relationship quality has significant influence. These findings are consistent with those reported by Anderson and Mittal¹⁸, which indicated that physical environment and customer service have indirect effects on patient satisfaction and customer loyalty, and they also concur with the results presented in the study by Dayal et al.¹⁹, which suggested that patient relationship quality influences WOM and intention to recommend. In other words, regarding WOM marketing, our findings, as well as those of others, suggest that relationship-oriented marketing can enhance customer intention to recommend by influencing their levels of satisfaction, trust and commitment.

Intention to recommend and opinion leader tendency had significant correlations with WOM activities. In addition, as shown in Reichheld’s study²⁰, intention to recommend is not only highly associated with actual recommendation behaviour by customers, but is also useful as an indicator of positive WOM, as it influences present and future purchase and WOM behaviours. It has also been shown that people with opinion leader tendencies often spread WOM because they have a tendency to pass information voluntarily to others, actively participate in the decision making of others and have excellent communication skills⁶.

Structural equation modelling entails testing causality and correlations among variables and has recently been widely used in various fields, including business, advertising and education. First of all, to test the validity of the survey questionnaire, construct validity, which consists of convergent validity and discriminant validity, was tested. All indices of the AVE were ≥0.5 and all the coefficients of CR were ≥0.7, suggesting a high level of convergent validity (Table 2). Discriminant validity was also confirmed, as the correlations among the variables were low (Table 3). Next, the goodness of fit of the research model was tested with the absolute fit and incremental fit indices. The first model was revised because the χ², GFI and AGFI values were low. Following a total of five revisions, the model that did not violate the selection principles and showed the highest goodness of fit was utilised as the final version (Table 4). The study results were interpreted based on the final model (Figure 1) and they appeared to be significant.

The limitations of this study include potential bias in the results because the participants were limited to patients from four dental clinics in a single city. The private dental clinics surveyed in this study are similar to the major types of dental clinics in Korea. However, it is difficult to generalise them as being representative for all the dental clinics of Korea. In addition, the flow of online information may not have represented the influence of WOM. Furthermore, various situational variables were not considered, including whether an initiator started the WOM voluntarily or in response to a recipient’s query during a WOM trigger. Further research is thus needed to compare the flow of online WOM, which spreads quickly, and that of offline WOM, as well as to apply various responsive variables that can influence WOM effects, including WOM triggers.

Despite these limitations, to the best of our knowledge, this study is the first to present a structural equation model for the path of WOM based on data from Korean dental patients. Our results verified that patient relationship and individuals with opinion leader tendency are key marketing targets for the dental industry. In addition, they suggest that a good relationship between patients and the dental health-care team plays an important role in determining patient satisfaction and positive WOM. According to a previous study, dentists and other oral professionals should attempt to encourage and support patients to experience the treatment procedures as a person, not as a patient²¹. Furthermore, given that patient satisfaction in this study was assessed in terms of patient subjective satisfaction, trust and commitment, we suggest that a patient-oriented dental treatment system with good quality control of treatment, as well as a dedicated attitude towards patients, will increase the numbers of positive WOM recommendations.

Conclusion

Our results showed that intention to recommend can be a powerful tool to predict WOM, and that patient relationship quality and opinion leader tendency were the primary determinants of WOM. Positive WOM for dental clinics can be expected from efforts to improve patient relationship quality, which includes patient satisfaction, trust and commitment, as well as strategies to promote a positive perception of dental clinics by effectively responding to the views of patients with strong opinion leader tendencies.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2017R1A2B4009648).

Competing interests

The authors declare that there are no competing interests in relation to this study.