Investigating the Medical Tourism Supply Chain in Shiraz, Iran: A Hybrid Grounded Theory and Rough DEMATEL Method

Background: This study sought to provide a comprehensive analysis of the medical tourism supply chain (MTSC) in Shiraz, Iran, to improve the city’s potential tourism market share in the post-COVID-19 future. In doing so, the study relied on a mixed research methodology. Primarily, interviews were conducted with 12 stakeholders involved in Shiraz’s MTSC, including general policymakers, managers of private/public healthcare providers, travel agency managers, and managers of medical tourism companies. The data were collected through semi-structured interviews and were then analyzed according to the systematic approach of grounded theory (GT). Results: The results helped to configure a model that analyzed Shiraz’s MTSC, which included 6 main dimensions, 17 sub-dimensions, and 48 criteria. To detect any interrelationships among the criteria, the model was further analyzed quantitatively through the rough Decision making trial and evaluation laboratory (DEMATEL) method. Conclusion: Proposing a novel methodology in medical tourism research, the study could practically contribute to different stakeholders in the medical tourism industry in Shiraz.


Introduction
Undeniably, 2020 was a tragic year for the tourism industry. Tourism, which was once one of the most lucrative industries in the service sector, underwent severe losses caused by the COVID-19 pandemic. Statistically speaking, around 1.5 billion international tourist arrivals occurred only in 2019, which generated a five-billion-dollar revenue for the tourism industry.
Moreover, the tourism industry directly contributes about 2.9 trillion dollars to the global GDP (Lock, 2020). However, according to the UNWTO, the number of international tourists decreased by 65% only in the first half of 2020 (UNWTO, 2020), and the tourism industry witnessed a 42% decrease in 2020 due to the COVID-19 outbreak (Lock, 2020). At the moment, several effective COVID-19 vaccines have been developed (Knoll and Wonodi,3 2021; Jones and Roy, 2021;Chagla, 2021), and the start of mass vaccination in various countries (McCarthy, 2021) could hopefully help to contain the ongoing pandemic.
Following that, as Gössling et al. (2021) predict, global tourism activities can be finally resumed.
Of course, the coronavirus could leave long-term health effects on infected individuals (del Rio et al., 2020;Yelin et al., 2020). As a result, among different tourism types, medical tourism is likely to gain momentum in the post-COVID-19 period. This situation could bring about an opportunity for medical tourism destinations (e.g., Iran), and could contribute to the recovery of tourism in destinations that offer medical tourism in the post-COVID-19 future (Abbaspour et al., 2020). Iran possesses a high potential for medical tourism, although the Iranian medical tourism industry encountered numerous obstacles even before the COVID-19 outbreak. Some of these obstacles are a lack of a comprehensive information management system for medical tourists, inadequate marketing, underdeveloped infrastructure, a shortage of skilled human resources, and a lack of effective training programmes in this field (Momeni et al. 2018). Despite the losses that this industry incurred during the COVID-19 crisis, the industry found the opportunity to re-examine, re-evaluate, and re-structure itself entirely. As such, the industry could improve the situation and overcome the barriers mentioned above, while preparing itself to benefit from the potential market after the COVID-19 pandemic is fairly contained.
The purpose of this study is to conduct a comprehensive analysis of medical tourism supply chain (MTSC) in Shiraz city, one of Iran's most significant medical tourism destinations. The study draws on a mixed research methodology. First, in a qualitative investigation, the study uses the grounded theory (GT) methodology to identify the factors that could help analyze the MTSC in Shiraz. Then, in a quantitative analysis, the interrelationships among the factors identified are determined through the rough DEMATEL technique. As a result, the study makes two major contributions: (a) it tries to provide practical findings that could help improve the situation for different stakeholders engaged in the medical tourism industry in Shiraz; and (b) it uses a novel methodology, which, to the best of the authors' knowledge, has not been employed in the literature on the tourism industry. The following sections in the paper are as follows: in the next section, the relevant literature is reviewed. Section 3 substantially explains the context in which the study in conducted and the methodology utilized. In section 4, the results and discussion are presented, and finally, the conclusions are reported in section 5. 4 2. Literature Review

Medical Tourism and its Supply Chain
Traveling for medical and health-related purposes does not represent a new phenomenon.
Historically speaking, people always traveled to find quality healthcare services (Reed, 2008). However, medical tourism became popular only in the late twentieth century (Connell, 2013), as a subset of health tourism, in addition to wellness tourism (Smith & Puczkó, 2008).
Most studies define medical tourism as the process of traveling abroad to receive medical services to save money or reduce waiting time (De la Hoz-Correa et al., 2018;Heung et al., 2010;Chuang et al., 2014). Of course, some definitions (see Hudson & Lee, 2012) have also highlighted the domestic nature of medical tourism.
Medical tourism has recently gained momentum, particularly in Asian territories (including India, Thailand, Singapore, and Malaysia) and in other countries worldwide, such as the United States, Canada, Brazil, South Africa, Indonesia, Mexico, Cuba, and the Philippines (Crooks et al., 2010). Like other types of tourism, medical tourism has been also highly affected by the COVID-19 pandemic. Although medical tourist arrivals almost entirely disappeared during the COVID-19 crisis, medical tourism can potentially rebound in the post-crisis future (Oğuz et al., 2020;Sharma et al., 2020;Abbaspour et al., 2020).
Supply chain management, as a technical term, was first used in the literature in the 1980s, and it gained popularity in the 1990s, when many scholars tried offer clear definitions of the notion (Ellram & Murfield, 2019). Since then, supply chain management has been a topical concept in management research. A supply chain (SC) consists of different participants who seek to (in)directly fulfil customers' demands. Such participants may include manufacturers/producers, suppliers, transporters, warehouse managers, retailers, and even customers (Chopra & Meindl, 2007). Alford (2005) was among the first scholars who investigated the SC in the tourism context, followed by other researchers such as Ke (2006), . However, Chen (2009) defined the tourism SC as a medium that connected all tourism activities performed through the flow of information, materials, and funds. According to Chen (2009), the tourism SC is meant to share resources, reduce costs, and generate customer value. In a more specific context, Lee and Fernando (2013) defined the MTSC for the first time as a complex network including at least five different sectors: accommodation, chemistry and pharmaceuticals, hospitals, transportation, and insurance.

Previous Research
A few studies have investigated the MTSC. Ferrer and Medhekar (2012) examined three main factors (cost, speed, and reliability) that could affect the global MTSC and the decisionmaking process for traveling to another country for treatment purposes. The results showed that low cost, no waiting time, and privacy (or reliability) of medical treatment could increase demand for medical tourism. Lee and Fernando (2015a) conducted one of the first investigations into the factors affecting the MTSC. As they explain, there are four effective factors in the MTSC: cooperation, coordination, information sharing, and integration. Their findings could raise practitioners' awareness in the medical tourism industry, especially in developing countries.
In another research, Lee and Fernando (2015b) developed a model analyzing the MTSC. The model was composed of three main dimensions: drivers, practices, and the MTSC performance. The drivers of the MTSC included trust, commitment, and mutual dependency, while practices encompassed collaboration, coordination, and information sharing. Similarly, MTSC performance was divided into financial and non-financial categories. The study then statistically investigated the relationships between the different elements of the MTSC. In the same vein, Rahman and Zailani (2017) examined the effectiveness and outcomes of Muslim-friendly medical tourism, determining the relationships among the determinants of the MTSC, including trust, commitment, mutual dependency, collaboration, coordination, information sharing, and performance. Chung and Chang (2017) constructed a framework to measure the sustainability of the MTSC. The framework consisted of four main criteria (financial, customer, internal process, and growth perspectives) and 16 sub-criteria. The framework was analyzed using the analytical network process (ANP) technique. The results demonstrated that the financial perspective was the most significant element for integrating and improving the MTSC. As such, a healthy financial status would be essential to achieve stability in such SCs.
Conducting an exploratory research, Kaewkitipong (2018) investigated the MTSC in Thailand, identified its stakeholders, and explored information flow in the chain. The results indicated that the lack of cooperation and integration among the SC stakeholders in the sector had led to limited information exchange. Thailand, using structural equation modeling (SEM). The results of this study suggested that mutual dependency, information sharing, and coordination were among the significant effective factors in improving the performance of the MTSC members. Furthermore, there 6 was a lack of commitment and trust among the members of the Thailand's MTSC. Karadayi-Usta and SerdarAsan (2020a) proposed a conceptual model of the MTSC to gain a clear understanding of its nature and business processes. The conceptual model identified seven business processes, namely service design, service recovery management, customer relationship management, supplier relationship management, demand management, capacity and resource management, and service delivery management. The model could help its users to internally shape their organization's SC. In addition, this model serves as the basis for SC collaboration decisions.
Similarly, Karadayi-Usta and SerdarAsan (2020b) built a collaborative framework for MTSC operations. More specifically, they concentrated on the collaboration between an assistance company and a medical institution, by developing a framework composed of steps, tools, and techniques, for SC operations in medical tourism services. Mekhum (2020) measured the impact of SC capabilities on health tourism performance by considering the mediating role of health care quality. The results revealed that SC capabilities, such as distribution channel, staff skills, and distribution time, positively affected the quality of health services and health tourism performance. The quality of health services also had a significant positive effect on health tourism and had a mediating effect on the distribution channel, time, and tourism performance.

Place of Study
Located in the southwest of Iran, Shiraz is the fifth most populous city in Iran and the capital of Fars Province. Geographically speaking, Shiraz is located in a region conveniently close to the Arab States of Persian Gulf, which considerably invest in the medical market. The city benefits from a wide range of clinical and paraclinical services, qualified and renowned physicians, and latest medical equipment (Jabbari et al., 2013). According to the data provided on the website of Shiraz University of Medical Sciences and Health Services (2018), Shiraz hosts 31 public/private specialized hospitals.
Another issue, as observed by Lovelock and Lovelock (2018), is the possibility of providing leisure elements in medical tourism destinations. Shiraz, as a historically and culturally rich city (Manoukian, 2012), can offer many leisure facilities. Therefore, every year a significant number of tourists travel to this city to receive medical treatment, which shows the high potential of this city in attracting medical tourists as a "hub" in the south of Iran. The present study investigated the status of Shiraz's medical tourism industry and explored previous research on the tourism SC. Following that, study identified different MTSC stakeholders in Shiraz, as illustrated in Figure 1.

Research Process
This study sought to conduct a comprehensive analysis of the MTSC in Shiraz, using a mixed research methodology. First, in a qualitative investigation, the GT was used to frame a process analysis model. To accomplish this, in-depth face-to-face interviews were conducted with a panel consisting of 12 experts in Shiraz city's medical tourism. The experts were selected through purposive sampling from different MTSC stakeholders including general policymakers, managers of private/public healthcare providers, managers of tourism agencies and medical tourism firms, hotel managers, facilitators, and academics.

9
The interviews were structured and carried out from April to August 2020. Given the social distancing policy imposed following the COVID-19 outbreak, the interviews were conducted via online platforms. After an interview protocol was first formulated, the quality of its questions were evaluated by two experts. Moreover, several factors were taken into account in the interview process. First, the interviewers ensured that the interviewees were prepared in advance. Next, each interviewee orally expressed his/her consent, and the interview was carried out in a friendly and stress-free online environment. The purpose of the interview was fully explained to each interviewee. The qualities of the data analyzed and the model proposed were evaluated according to the nine criteria suggested in Flint's (2001) framework (

Grounded Theory
As a research strategy used in social sciences, GT was first developed by Glaser andStrauss in 1967 (Kenny &Fourie, 2014). It can be simply defined as the process of constructing a theory from obtained data (Glaser and Strauss, 1967). To be more specific, according to Dunne (2011: p. 111): "in grounded theory, the researcher is not focused on testing hypotheses taken from existing theoretical frameworks, but rather develops a new theory grounded in empirical data collected in the field." Although Glaser and Strauss (1967) initially applied GT to nursing research, it has been successfully applied to numerous other contexts, such as SC risk management ( GT has undergone various changes throughout its development; however, currently there are three main streams using GT: The Straussian approach or the systematic approach (Strauss & Corbin, 1990), the Glaserian approach (Glaser, 1992), and the constructive approach (Charmaz, 2000). Researchers are advised to select one of these approaches depending on the purposes they pursue in their investigations (Heath and Cowley, 2004).
Combining these approaches, however, does not seem to be an ideal option (Van Niekerk & Roode, 2009). Given these issues, the present study relied on the Straussian approach to achieve its purposes because this approach provided more guidelines compared with the others (Heath & Cowley, 2004;Van Niekerk & Roode, 2009). According to Strauss and Corbin (1990), GT is accomplished through three core steps, namely open coding, axial coding, and selective coding (as elaborated on below).
The process of verbatim analysis of data for the purpose of discovering concepts, their specifications, and dimensions is called open coding. In this step, categories are constructed that encompass all objects, events, or actions/interactions related to the phenomenon under investigation (in this case, MTSC). When categories emerge, by comparing their subordinate members with each other, one learns the differences of such classified elements (objects, events, or actions/interaction) in terms of their properties and dimensions. Therefore, they are classified into sub-categories, after which this step in finalized.
"Axial coding" is conducted to construct a detailed description of the phenomenon.
Thus, axial coding links categories with their sub-categories based on their properties and dimensions. This process occurs around the axis of a category, and that is why it is called axial coding. All concepts identified revolve around a main phenomenon. According to Strauss & Corbin (1990), axial coding is carried out based on the following categories: The last step of GT is "selective coding." In this step, the theory is refined and integrated. According to Strauss and Corbin (1990), at the selective coding stage, no new properties, dimensions, or relationships emerge through the analysis. The researcher selects the main category that encapsulates the central theme of the study and then integrates all other categories. The objective of this step is accomplished by reviewing technical memos collected through data analysis and interviews.

Rough Set Theory
Rough set theory was first introduced by Pawlak as a mathematical approach (1982); following that, Zhai et al. (2008) introduced the rough number concept. Since then, rough set theory has been incorporated into many decision-making techniques, such as TOPSIS (Stevic et al. 2018, Shojaei & Bolvardizadeh., 2020, AHP (Pamučar et al., 2018), ANP (Li and Wang, 2018), DEMATEL , and BWM (Liu et al., 2020). By aggregating group information, rough set theory can overcome the vagueness and subjectivity that arise from diverging judgments in a group decision-making process (Mao et al., 2020). Zhu et al. (2015) mention the following definitions regarding rough numbers: Definition 1. Let U be the universe containing all objects and P be a random object of U, A be a set of n classes {A1, A2, …, An} that cover all the objects in U. Given that these classes are ordered as {A1 < A2 < …< An}, then ∀P ∈ U, Ak ∈ R, 1 ≤ k ≤ n point to the class to which the object belongs. The lower approximation, upper approximation, and boundary region of the class Ak are defined as: Definition 2. Ak can be shown as the rough number RN(Ak), which is determined by its corresponding lower limit and upper limit: where ML, MU are the numbers of objects that are contained in Apr, respectively.

Rough DEMATEL
Decision-making trial and evaluation laboratory (DEMATEL) is a useful technique to conceptualize the structure of cause-effect relationships among the elements in a complex system (Fontela and Gabus, 1976). DEMATEL offers an effective way of visualizing the structure of complex causal relationships, describing the relationships between different elements of a system (Song & Cao, 2017).
Step 1: Construct a group direct-relation matrix m experts make pairwise comparisons for n criteria according to the crisp DEMATEL scale, where 0 indicates "No Influence" and 4 shows "Very Strong Influence" (Wu, 2008 , k = 1, 2, …,m where r k ij is the crisp judgment of the expert k regarding the influence of the i-th criterion on the j-th criterion.
Step 2: Determine the rough group direct-relation matrix The crisp judgments are then converted into rough ones according to Definition 1. Then, the rough group direct-relation matrix R can be created through: Step 3: Create the rough total-relation matrix The linear scale transformation is used as a normalization formula to transform the element scales into comparable scales. The normalized rough group direct-relation matrix R is obtained as follows: Additionally, the rough total-relation matrix (T) can be created via: where and are the lower and upper limits of the rough interval tij in the totalrelation matrix, and I is the unit matrix.
Step 4: Calculate the "prominence" and "relation" values Next, the sum of the rows and the sum of the columns are showed by xi and yj, repressively, within the rough total-relation matrix using the following equations: ,i = j

Results
To accomplish the objectives of the study, first the transcripts of the interviews were analyzed according to the steps of the Straussian grounded theory (open coding, axial coding, and selective coding). Following that, "poor service delivery in Shiraz medical tourism" was extracted as the main phenomenon and the basis of the process model. As the interviewees stated, "poor service delivery" was the most significant threat and challenge facing medical tourism in Shiraz. According to Expert 9: "The fact that a medical tourist enters Shiraz and we cannot provide a proper service to his/her is the most serious threat. Even though s/he 15 finally receives the service in question, despite all shortcomings, this problem must be solved for the sake of development." Expert 11 also explained that: "The mismatch between the price and the quality of the services provided is certainly a weakness of the medical tourism industry in Shiraz." The other dimensions of the process model, including causal conditions, contextual conditions, intervening conditions, strategies, and outcomes, were determined based on the main phenomenon. The sub-dimensions and criteria of each dimension are shown in Table 2. High-quality and varied medical services in Shiraz G10 The long history of medical tourism in Shiraz G11 Availability of world-renowned doctors in Shiraz Shiraz's geographical conditions

G12
The strategic geographical location of Shiraz G13 The desirable ecosystem of Shiraz

Intervening Conditions
The important role of mediators I1 Brokers who inevitably mediate the relationship between physicians and medical tourists I2 Medical tourists' poor knowledge and information

I3
The importance of word-of-mouth advertising in target countries The role of government in medical tourism

I4
Problems medical tourists face in the medical visa application process I5 Moving beyond an oil-dependent economy

I6
The expansion of the Cultural Heritage, Handicrafts and Tourism Organization into a ministerial organization Improving medical tourism infrastructure S1 The need to create an appropriate structure for organizing/ monitoring SC members S2 Using cyberspace to promote medical tourism S3 Operating domestic/international flights to Shiraz Gaining competitive advantage by focusing on products and the market S4 Emphasizing domestic and regional tourism As mentioned earlier, in the quantitative stage, the interrelationships among the criteria of the process model were identified using the rough DEMATEL methodology. Based on the steps of rough-DEMATEL, the rough group direct-relation matrix, the normalized rough group direct-relation matrix, and the rough total-relation matrix were calculated and shown in Tables 3-5, respectively. Next, through the equations in Definition 5, the sum of the rows (Xi) and the sum of the columns (Yj) in the rough total-relation matrix were converted into crisp values (xi and yj respectively) to determine the prominence (mi) and relation (ni) values. Needless to say, in the DEMATEL methodology, each criterion is categorized under one of the two main groups: cause and effect. The cause-type criteria refer to the factors that left an impact, whereas the effect-type criteria were the factors that received influence (Fontela & Gabus, 1976). The two groups can be identified according to the values of their prominence and relation; if the relation value is positive, the criterion in question falls under the cause group. However, a negative relation value would mean that the criterion is an effect. On the other hand, the higher the prominence value is, the more important the criterion in question is (Song and Sakao, 2018). Table 6 shows the prominence and relation values of the criteria of the model.

Discussion
According to Table 6, "poor service delivery in Shiraz medical tourism" was the most significant cause, with a prominence value of 3.67 and a relation value of 0.57. As such, this criterion was justifiably selected as the main phenomenon in the process model. Moreover, "lack of transparent pricing" (C1) and "poor anti-corruption measures in medical tourism" (C2) were two other cause-type criteria of the model with prominence values less than that of the main phenomenon. In this regard, Expert 1 explained: "A medical tourist can hardly find the exact prices of the services s/he wants to use through online platforms. We do not have medical tourism packages in which the final price is clearly defined. As such potential tourists remain uncertain about their decisions." Moreover, "a lack of cooperation and coordination between the MTSC members" (C8) was the most crucial effect criterion and was influenced by the cause criteria. Expert 3 stated: "The confrontation between the stakeholders involved in Shiraz medical tourism has slowed down the industry's cycle; they see themselves more as adversaries than allies. This lack of coordination is the most serious challenge to this industry." Similarly, "economic fluctuations and devaluation of the Iranian currency" (G4) was the most critical cause in "contextual conditions", whereas "imposed international sanctions" (G3) was the most "The need to create an appropriate structure for organizing/monitoring SC members" (S1) was also an essential effect-type strategy. As the experts confirmed, the transformation of the Cultural Heritage Organization into a ministry was a turning point in restructuring  the criteria in the model. More precisely, each node represents a criterion, and the arrow between them shows their relation. As an instance, in Figure 4, "economic fluctuations and devaluation of the Iranian currency" (G4) would cause "low prices for medical and tourism services" (G5) with an intensity of 0.1068. Similarly, Figure 5 shows that "the expansion of the Cultural Heritage, Handicrafts and Tourism Organization into a ministerial organization" (I6) would lead to "the need to create an appropriate structure for organizing/monitoring SC members" (S1) (with an intensity of 0.1411) and would help to revisit and refine "inefficient solutions offered by the Iranian Health Tourism Strategic Council" (O1) (with an intensity of 0.0518). Naturally, higher intensity values would indicate the higher importance and significance of the relationship between two criteria in the model.
To improve the situation of medical tourism in Shiraz and thus increase the market share of this city in global medical tourism, the following practical suggestions were raised during the interviews with the experts. The first solution is to strengthen the Iranian Health Expert 3 also explained: "Human resources serving medical tourism need to understand cross-cultural differences and behavioral habits that travelers may exhibit." Moreover, some of the most critical solutions proposed by the experts were launching a comprehensive medical tourism portal, facilitating the medical visa issuance process, and paying attention to academic research in medical tourism.
Conducting face-to-face interviews under COVID-19 circumstances was a major challenge to this study and disturbed the research schedules. Furthermore, it was difficult to contact and interview more experts due to their busy schedules and management responsibilities. In two cases, the experts, for personal reasons, did not agree to have their voice recorded. Future investigations are strongly advised to probe into the relationships among the dimensions, sub-dimensions, and criteria, using various statistical methods/techniques.

Conclusion
This study conducted a comprehensive analysis of the MTSC in Shiraz, Iran, by employing a GT-rough DEMATEL approach. The analysis revealed that Shiraz's medical tourism suffered from shortcomings in its service delivery. Meanwhile, two groups of strategies, namely improving the medical tourism infrastructure and creating competitive advantage by focusing on the product and the market, was unfruitful and did not help to increase the market share of Shiraz in global medical tourism. Hence, to improve the situation of medical tourism in Shiraz and thus increase its market share, several solutions are possible, such as strengthening the Iranian Health Tourism Strategic Council, creating validation platforms, empowering the private sector and human resources in this industry, launching a comprehensive medical tourism portal, facilitating the medical visa issuance process, and 24 paying attention to academic research in medical tourism. Furthermore, the findings and results of this study could suggest important practical implications as they can help different stakeholders in Shiraz's medical tourism industry to improve their current status. The GTrough DEMATEL methodology applied to a tourism context represented another innovation of the study.

Declarations Ethical Approval and Consent to participate
Not applicable.

Consent for publication
Not applicable.

Availability of supporting data
The datasets supporting the conclusions of this article are included within the article and in supplementary file.

Competing interests
Not applicable.

Funding
The present study did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors.

Authors' contributions
A