Hospital Location Selection: A Systematic Literature Review on Methodologies and Applications

(e increased focus of people on the quality of health care in recent years has led hospital owners to develop strategies and policies to improve medical services through the establishment of new hospitals. For hospitals to be competitive, the hospitalʼs location and proximity to potential patients are considered crucial factors in establishing new hospitals. In this context, evaluating and selecting the most suitable hospital location to establish a new hospital from the multicriteria decision-making (MCDM) perspective is a priority for the entrepreneurs or government to gain a competitive advantage.(erefore, this study aims to present a systematic literature review of the hospital location selection problem considering the applied methods and application areas. (e preferred reporting items for systematic review and meta-analysis statement (PRISMA) are used as a reference framework. Initially, known electronic databases (Web of science, IEEEXplore, Scopus, Science direct, and Google Scholar) were searched up to the early 2021. A number of 47 articles are selected and analyzed under this systematic framework based on inclusion-exclusion points. State-of-the-art developments in adopting MCDM methods and their fuzzy extensions are summarized. All the articles have been examined in a systematic taxonomy to find answers to six research questions (trend, country of origin, outlet journal, MCDM methods used, MCDM environment and criteria type, and decision criteria used). Results show that (1) AHP and GISbased MCDM models are the most contributing approaches to the solution of this problem, (2) location selection criteria are mostly cost, demand, environment, population, government, competition in the market, and distance to some important places, (3) the fuzzy structure is also preferred in addition to the MCDM structure depending on the crisp data type, and (4) the location selection criteria are mostly considered subjective. We pay attention to promising directions that can dominate future research in this field from a methodological or applicability perspective. (is study shows the current views and opportunities for researchers and practitioners and acts as a guide to encourage more creative studies in this field.


Introduction
e decision to choose a hospital location is one of the most crucial policy decisions that government and health policymakers pay attention to. In health care, the priority is to give all patients the right place and exemplary service and be fair. Choosing the optimal hospital location is vital for the effectiveness, quality, and equity of health services [1]. e choice of hospital location is a strategic decision [2]. e location to be selected should be sustainable and capable of eliminating future problems. Choosing the wrong location can lead to significant customer dissatisfaction and increase in cost [3]. A vivid example of this situation is the Covid-19 pandemic.
Nowadays, an overwhelming majority of the world is still fighting against this pandemic. e world has faced an unusual demand for infected people in hospitals. is surge made it mandatory for countries to build hospitals, field hospitals, or specific pandemic hospitals in a considerable and short time. Initially, the Chinese government has recently announced that the two hospitals in Wuhan were built within ten days to meet increased demand. Many other countries, like Turkey, have decided to do so in 45 days. ese unpredictable conditions have obliged states to make these investments and make these decisions that cover various criteria. erefore, many criteria should be considered in a suitable location selection process. is problem depends on many criteria such as environment, demand, population, proximity, competition, government policies, and costs as in other hospital ranking problems [4][5][6][7][8]. For this reason, the selection of a hospital location can be considered as a multicriteria decision-making (MCDM) problem [1]. e problem can be regulated by employing a carrying capacity model, GIS, MCDM, or fuzzy models. MCDM methods have been widely applied to this subject. One of the most applied methods is AHP [1]. Other applied methods are ANP, TOPSIS, VIKOR, ELECTRE, SAW, GRA, EDAS, ARAS-G, CODAS, CRITIC, Entropy, and fuzzy extensions of these methods. ere is an MCDM method specific to each heading (the stage in hospital location selection). For example, pairwise comparison-based MCDM methods are the most widely used method for determining criterion relative importance levels. is is because the relevant stage of the problem shows full compliance with the structure of the method. Figure 1 illustrates the problemʼs implementation stages with the relevant tools for each stage.
Since the topic is handled in many types of research, there is no review paper covering state-of-the-art research. e only attempt by Moradian et al. [9] aims to determine disaster risk criteria in hospital location selection by criticizing 15 studies. ey have merely categorized the criteria without reviewing the methods, criteria type, uncertainty analysis in data type, and some various aspects (publication trend, country of origin, and publication outlet either as a journal article or known conference paper). Moradian et al. [9] stated that there exist two crucial theories in hospital location analysis. While one is the Weberian model, which focuses on a single objective, namely, the minimum cost or maximum profit, the second is concerned with the "behavioral approach" which simultaneously considers numerous criteria to determine the most suitable location. MCDM-based methods are considered under this second theory. Since the literature lacks in terms of presenting studies of hospital location selection from the MCDM perspective, we aim to provide a state-ofthe-art literature review of the hospital location selection problem considering the applied MCDM methods and selection criteria. Known electronic databases were searched up to the early 2021 with the aid of the PRISMA framework. Articles that orient various MCDM methods are selected and analyzed under a systematic framework. State-of-the-art developments in adopting MCDM methods and their fuzzy extensions are summarized in a hospital location selection process. We pay attention to promising directions that can dominate future research in this field from a methodological or applicability perspective. is study is different from the Moradian et al. [9] study by providing a general criteria hierarchy that can be fitted to the possible future hospital location selection research. e existing literature review also reveals the gaps in fuzzy MCDM domain that researchers can benefit from in hospital location selection studies.
In summary, to the best of the authors' knowledge, although there are different state-of-the art reviews on location selection problem in the literature [10][11][12][13][14][15], there is no review paper attempt regarding hospital location selection. As a result, to fill the gap in the literature, we provide a systematic review of publications related to hospital location selection till 2020 with focusing on MCDM concept and the state-of-the-art progress in this subject. Contrasted with the existing reviews on similar subjects, the novelties of this review are as follows: (1) we focus on the problem from the MCDM perspective and as well as GIS tools; (2) we categorize the applied methods, the decision criteria, MCDM environment, and type of criteria by reviewing the publications in detail; and (3) we highlight the advantages and disadvantages of applied methods as well as research limitations, and future research studies are determined. e rest of the paper is organized as follows. e second section gives the literature reviewing methodology. en, in the third section, we provide the results in graphs and tables. e last section offers future research areas that can be studied by scholars.

Review Methodology
A PRISMA framework is frequently used in the literature [16][17][18][19][20][21][22][23] which is followed as the review methodology of this study. e referenced review studies have considered the following issues in designing their review frameworks: (1) search strategy, (2) inclusion-exclusion criteria, (3) study selection, and (4) data extraction and classification. Initially, known electronic databases (Web of science, IEEEXplore, Scopus, Science direct and Google Scholar) were scanned until the early 2021 by using suitable keywords. In this study, a Boolean query search using keywords such as "hospital location selection" OR "hospital site selection" OR "MCDM" OR "GIS" OR "AHP" OR "TOPSIS" is used. Also, the search is executed by different combinations of these keywords and as well as their long versions (e.g., multicriteria decision-making, geographical information system, analytic hierarchy process, and technique for order preference by similarity to ideal solutions). A total of 40 pages were scanned in Google Scholar with these key terms and the famous databases of WOS, IEEEXplore, Scopus, and Science direct. Here, to add Google Scholar as an outlet in our search, many studies regarding our scope are not included in the essential databases. A considerable amount of articles on hospital location/site selection were unfortunately published in journals not included in SCI-Expanded, SCI, SSCI, or ESCI indexes. For this reason, we have scanned Google Scholar to a certain number of pages. e inclusion criteria are that the article is written in English and is a journal or important conference article (e.g., IFAC, IEEE, or ASCE).
Studies with the following criteria are excluded within this review: (1) the location of the hospital is directly determined with a conceptual framework without using any mathematical model or MCDM-based model; (2) it has handled the location problem of any nonhospital health-related facility; (3) the writing language of the article is not English (e.g., Arabic); and (4) the study is a review article, not a research article. After extracting duplicate articles in each database, the articlesʼ titles and abstracts were scanned according to the inclusion/exclusion criteria. At this stage, the articles that appeared out of the subject were sorted out. In the PRISMA eligibility assessment phase, among 49 papers, one paper was excluded [24] since its writing language is Arabic. One another paper [9] seems not a research article. erefore, we exclude it from the current paper pool. Finally, we structured our state-of-the-art review by including 47 papers ( Figure 2). Some essential topics were collected on an Excel sheet to discuss hospital location selection. e column names in this sheet are as follows: Study ID, Study abbreviation (in the format of First author last name et al. [Publication year]), Year of publication, Country of origin, Journal name, Applied method(s), MCDM environment, Type of fuzzy set (e.g., fuzzy, crisp, and rule-based), Type of criteria (subjective, objective, and subjective + objective), and Main criteria. We then carry out some analyses in the light of the reviewed studies. ese headings also show the content of the research questions (RQs) that the article seeks to answer.
ese RQs were identified, and answers were sought. e RQs determined for the analysis are as follows:

Results
is section presents the distribution of papers by the following points of view: (1) publication trend by years, (2) distribution of in terms of country of origin, (3) published journal, (4) applied MCDM methods either in single or hybrid, (5) MCDM environment and type of criteria used in the papers, and (6) the location selection criteria used in the documents.

Publication Trend.
Firstly, the time-dependent variation of the number of papers included in this literature review has been handled. Figure 3 shows the trend of articles by year. It can be easily seen that there are two crucial outliers in the production of papers in 2013 and 2016. e exponential trend has an acceptable R 2 value (40%), indicating that the documents increase through the literature.

Country of Origin for the Hospital Location Selection
Studies. Secondarily, the reviewed papers are analyzed by country of origin. e "country of origin" criterion we examine here refers to the country of the first author of the reviewed study. Figure

Applied MCDM Methods and eir Extensions.
From the point of applied MCDM methods, the following significant findings can be extracted (Table 1). (1) AHP and GIS-based models play a dominant role among all methods (n � 18/47). Among these 18 papers, five studies integrated AHP and GIS in selection. (2) Higher than half of the published papers utilize a hybrid MCDM approach. ANP-TOPSIS and F-TOPSIS are forefronts. (3) Some well-known MCDM methods are not used in the reviewed papers anymore, for example, methods like BWM, PROMETHEE, DEMATEL, and MAIRCA. (4) In hybrid approaches, GIS is the most frequently preferred tool as it can visually analyze location. It is often used in other location selection problems.
AHP is a common MCDM method that has previously been applied to various problems in location selection. It has some characteristic features. It has a decision hierarchy that demonstrates the difficulty in levels.
ese levels include decision goals, criteria, subcriteria, and alternatives. e method is also based on a pairwise comparison manner using the 9-point scale of Saaty. It checks the consistency of decision matrices to build the model on a factual basis. e problem of hospital location selection can be easily solved via AHP and its improvements. is problem has a decision hierarchy, which the AHP method also has. e usage of AHP and GIS in hospital location selection problems stems from their advantages and integration compatibility. A GIS is a visual tool that provides convenience in understanding the criteria of the potential location such as geographic, geological, and distance to central points [25][26][27][28][29][30][31][32]. It also Records excluded a er titles and abstracts were reviewed (n = 445) Studies considered ineligible (n = 2) includes many useful maps with various layers. Apart from GIS modeling, AHP is combined with other MCDM methods in hospital location selection. Among these methods, TOPSIS and its derivations are forefronts [2,[33][34][35][36][37][38][39]. e TOPSIS method was initially proposed by Yoon and Hwang [40] as an MCDM method. It is based on separation from the ideal and negative-ideal solution concept. According to the method, each evaluation criterion has a monotonous increasing or decreasing trend. To create an ideal solution, the largest of the weighted evaluation criteria in the matrix is selected [41]. It uses a criteria weight matrix and a decision matrix consisting of alternatives, criteria, and respected performance measures. In Khaksefidi and Miri [36],   Mathematical Problems in Engineering TOPSIS in evaluating alternative hospital locations [2,34,35,39]. While Miç and Antmen [2] and Baran [34] applied for fuzzy triangular numbers in TOPSIS, Kahraman et al. [35] and Senvar et al. [39] used spherical fuzzy numbers and hesitant fuzzy numbers, respectively, in TOPSIS. Kutlu Gündogdu and Kahraman [35] proposed that spherical fuzzy sets reflect uncertainty and ambiguity in real-world decision problems better than classical fuzzy set theory. ey are mathematically based on a membership function on a spherical surface. ey independently describe the degree of membership, nonmembership, and hesitancy in a larger domain (the sum of these three values must be between 0 and 1). ey are considered as the integration of Pythagorean fuzzy sets and neutrosophic sets. Hesitant fuzzy sets have a flexible style in dealing with uncertainty by allowing many potential degrees of membership of an element to a set. Considering the advantages of these fuzzy sets, an improved TOPSIS model can pay more attention to the hospital location selection problem. ere are still more gaps in the literature regarding this problem using new types of fuzzy set theory. e neutrosophic fuzzy set, Pythagorean fuzzy set, Fermatean fuzzy set, intuitionistic fuzzy set, and interval type-2 fuzzy set are not integrated with TOPSIS. However, they can easily produce reasonable solutions in solving the hospital location selection problem. Lin and Tsai [38] and Lin and Tsai [37] developed an ANP-TOPSIS integrated model where ANP is used to obtain relative location selection criteria weights apart from AHP-TOPSIS models.
ere are also some single and hybrid MCDM methods applied to the hospital location selection. Single methods refer to a single MCDM method without merged with any fuzzy set extension. For example, ANP is used byÖnüt et al. [42]. ey improved the study of Wu et al. [43] by modifying location selection criteria considering the observed countryʼs local conditions. is study is followed by two studies focusing on single GRA methods and VIKOR, respectively [44,45]. In a study by Lin et al. [45], an optimal location for a regional Taiwanese hospital location is determined via the GRA-based model. Wang et al. [46] proposed a model for selection of hospital constructions with probabilistic linguistic MCDM with weight determined by the CRITIC method. Çelikbilek [44] evaluated to select the best hospital location for a private health institution using VIKOR. VIKOR is originally in Serbian (VlseKriterijumska Optimizacija I Kompromisno Resenje) and means that multicriteria optimization and compromise solution. It was initially proposed by Opricovic [47]. It requires a criteria weight matrix and a decision matrix covering alternatives, criteria, and respected performance measures (values of alternatives concerning the criteria). From this aspect, it provides useful solutions to the problems suitable for the MCDM concept. Although it has not contributed a lot to the hospital location selection problems [48], it can solve this problem with its compromised solution concept like TOPSIS.
On the other hand, hybrid methods refer to an MCDM method, often equipped with different fuzzy sets. F-ELECTRE [49], F-SAW [50], and F-ANP [51] are some examples from the reviewed papers. e ARAS-G proposed in Sen [52] puts the ARAS method and grey numbers together to determine a new public hospital location. In a study by Hashemkhani Zolfani et al. [53], a grey-based decision support framework using CRITIC and CoCoSo methods for location selection of a temporary hospital for Covid-19 patients is used. A case study is also performed for Istanbul using the proposed decision-making framework.
In most of the reviewed papers, AHP and its extensions (e.g., F-AHP, D-AHP, and DF-AHP) and some similar methods (e.g., entropy, ANP, and F-ANP) are used to obtain relative importance values (weights) of location selection criteria and subcriteria [3,42,51,[54][55][56]. However, new and novel approaches have been recently released in the literature to overcome the drawbacks of AHP. e best example that comes to mind quickly is BWM. It is a recently proposed MCDM method by Rezaei [57]. It needs a smaller number of pairwise comparisons and more consistent comparisons compared with AHP. Since the crisp values of criteria may be inadequate to model, the problem considering the vagueness and ambiguity and providing pairwise comparisons with less compared data, the BWM is extended with fuzzy triangular sets [58]. e method is applied in various areas from manufacturing to supply chain management and transportation [59][60][61]. Although plenty of research studies are performed in these areas, there is no contribution to hospital location selection analysis.
Various methods are applied to the hospital location selection problem like belief rule-based system and evidential reasoning [62,63]. e evidential reasoning method is developed to handle MCDM problems having both qualitative and quantitative criteria. Unlike classical MCDM methods, it demonstrates the MCDM problem via a decision matrix (in another name, a belief expression matrix). Each decision criterion regarding hospital location selection concerning alternative locations is described by a distribution assessment using a belief structure [62]. erefore, evidential reasoning is well-suited to handle incomplete uncertainty since it uses a belief structure to model an assessment as a distribution [63]. Apart from these, in a study by Kaveh et al. [28], two important metaheuristics algorithms "GA and PSO" are used with AHP and GIS in a hospital location-allocation problem.

MCDM Environment and Type of Criteria.
Another vital highlight included in the literature analysis is the MCDM environment and the type of decision criteria. Crisp, fuzzy, and grey environments are considered under the MCDM environment.
Regarding the fuzzy set version used in the hospital location selection problem, we can say that fuzzy triangular numbers are the most preferred. In 10 of the 12 studies, fuzzy triangular numbers [2, 3, 34, 44, 49-51, 55, 56, 64] were used, while spherical fuzzy numbers [35] were used in one and hesitant fuzzy numbers [39] in one. To illustrate the advantages of joint usage of MCDM and fuzzy sets in hospital location selection, a comparative discussion of the classical MCDM, grey theory-based MCDM, and general fuzzy set theory-based MCDM techniques is performed and presented in Table 2. e classical MCDM does not consider data/information uncertainty, converting expertsʼ judgments into a crisp value. e grey theory-based MCDM and general fuzzy set theory-based MCDM reflect the uncertainty of expertsʼ assessment environment. Still, they do not measure the degree of membership and nonmembership of the evaluated events simultaneously.
Another critical issue in the examined studies is the type of location selection criteria data used ( Figure 6). We made a classification as subjective, objective, studies using both, and nonapplicable ones. While 32 out of 47 studies dealt with the criteria subjectively, 6 of them use both subjective and objective criteria [25,29,33,50,62,67] and one of them use only objective criteria [37].

Selection Criteria Used for Hospital Location.
e hospital location selection is defined by the implementation of a five-stage process which is divided into the following headings [1,9,64]: (i) Determining the selection criteria (ii) Determining the importance of these criteria (iii) Proposing facility location options (iv) Evaluating facility location options (v) Making the final decision As one of the main components of this problem is the decision criteria, it is so crucial to discuss the papers in terms of selection criteria in depth. e reviewed papersʼ decision criteria are analyzed (please refer to Table 3 to see the whole list of main criteria used in selecting hospital location). According to the results, nearly half of the papers (n � 20/42) utilized a "cost" related main criterion. It is mentioned with a name of investment costs, travel costs, construction costs, land cost, building regulation cost, and merely cost in different studies. In 10 papers, demand is mentioned as the main criterion. 18 of 42 papers include a main environmental criterion. It is mentioned in various names such as "Environmental factors," "Environmental situation," "Environmental quality," "Environment," "Environment and safety," "Environmental issues," and "Environmental favor." In 11 papers, the population is also considered as a crucial parameter on hospital location selection. Seven studies focus on competition on the market and distance to some critical locations.
e first two studies on hospital location selection are described by Lin et al. [45] and Wu et al. [43]. In these two studies, six main criteria were determined: (1) factor conditions, (2) demand conditions, (3) firm strategy, structure, and rivalry, (4) related and supporting sectors, (5) government, and (6) chance. In the factor conditions criteria group, authors consider capital cost, labor cost, and land use cost as subcriteria. Capital cost refers to the capital to construct a hospital building [43]. Labor costs cover the costs to be spent on staff, including doctors, nurses, technicians, pharmacists, caregivers, cleaning staff, and all other healthcare professionals regarding quality and quantity [39]. Land use cost is described as an economic value and suitability of land for constructing a hospital building. For example, the cost may vary depending on whether the land is arid or commercial land [29,43,76]. Some scholars have used different names concerning this group (factor conditions).
ese are as follows: In the studies of Miç and Antmen [2], Baran [34], Lin et al. [50], andİnce et al. [74], "an investment cost" term is mentioned to define a total of hiring purchase, facility arrangement costs, and environmental planning costs. Kahraman et al. [35] used the term "installation costs" instead of factor conditions. Almost half of the scholars mention a general term of "cost" in the context of factor conditions group [2, 3, 26, 33-35, 39, 44, 49, 50, 52, 54, 56, 62, 63, 65, 69, 74-76]; İnce et al. 2016. While this term is explained as a fundamental factor of cost of land, land topography, land ownership, and running/ maintenance cost in some reviewed papers [70,75], it is frequently described as a total factor of land and construction costs [26,33,69]. e second vital criteria group concerns with demand conditions in hospital location selection analysis. In this group, three important subcriteria are mentioned [43,45] as population quantity, population density, and population age scattering. e demand for the potential hospital is directly related to the areaʼs population where it will be established. us, the demand in a large city is expected to be higher than in a rural area. Population density calculated as population per square meter is another crucial criterion that corresponds to the hospital scale and type. e type of disease (acute, chronic, and so on) in the potential hospital is related to distributing the populationʼs age. For example, if there is a risk of cancer in most of the population living in that area, it would be appropriate to evaluate building a cancer hospital. Similar to the contributions of the earliest Mathematical Problems in Engineering two works [43,45] and Lin and Tsai [37], many studies consider a demand-based main criterion on hospital location selection problem [1,26,37,38,42,44,50,64,69]. In a study by Şahin et al. [1], two additional criteria are attached to the criteria pool named as income and possibility of population change.
Çelikbilek [44], Khotbehsara and Safari [68], Khaksefidi and Miri [36], Vahidnia et al. [56], Soltani and Marandi [55], Soltani et al. [76], Abdullahi et al. [25], and Yuen [77] took into consideration merely "population density" as a demand criterion. A different subcriterion of "prospective population" is used in a study by Senvar et al. [39] and Adalı and Tuş [33]. As a creative subcriterion under the "population characteristics" main criterion, Kumar et al. [49] suggested four headings: education, earning/economy, the structure of the society, and health awareness. is is differentiated from the others in the literature. From this aspect, for future studies, scholars may pay more attention to these. In some reviewed papers, authors prefer to mention demand conditions under the name of "demographic structure/demographic infrastructure/demographics" [2,[33][34][35]39]. e third criteria group is related to the firm strategy, structure, and rivalry. is group is investigated under three subcriteria as a management objective, competitor hospitals, and policymakersʼ attitude [37,43,45]. e management objective is associated with establishing a mission, vision, and policy statement for the management practices [42]. e rivalry among hospitals or competitor hospitals impacts a new hospital investment project and the location of the new hospital building. e policymakersʼ attitude affects the decision of possible hospital location. For example, the administrators, consultants, and architects may have different opinions towards managementʼs style. And, this conjuncture is likely to affect the decision. Studies by Şahin et al. [1], Yuen [77],İnce et al. [74], and Organ and Tekin [54] considered "competitors" as the leading criterion group. ey determined four subcriteria under this group named medical technology, total beds, units, and whole hospitals. In a study by Assad [26], a group called "administrative concerns" is mentioned regarding firm strategy, structure, and rivalry. It includes policymakersʼ attitudes and hospital personnel. Wibowo [64] suggested the organizational sys-temʼs main criteria covering the subjective assessment of managementʼs attitudes towards business practices and competitors. is is often determined by the management objective, the managementʼs attitude towards competition from other hospitals, and policymakersʼ attitude towards the management style to achieve long-term success. e fourth criteria group suggested by the owner of the first study on hospital location selection [45] is "related and supporting sectors." e existence of related and supporting sectors is an essential factor affecting location choice. It is evaluated on three subcriteria, including medical practices and the pharmaceutical sector, hospital management sector, and health sector [43,50,51]. As the equivalent of this criterion, main criterion group with different names is available in the literature such as "industries and educational institutions" [31], "supply chain sector" [42], "related industry" [1], "sector support" [69], "supporting industries" [64], "support" [72,73], "medical suppliers" [44], and "existing healthcare centers" [76]. All of the terms mentioned above have the same meaning as Lin et al. [45] which is "relevant and supportive sectors." e fifth group covers governmental policy efforts towards establishing hospitals to strengthen their competitiveness [1,33,37,38,42,43,45,50,51,64]. is group is investigated under three subcriteria, including qualifications   of the hospitalʼs establishment and the regulations of the established standard, efforts to promote a medical network, and promulgating tasks that require a hospitalʼs assessment [37,43,45]. In a study byÖnüt et al. [42] and Şahin et al. [1], this group consists of incentives, legislation, policies, and tax. e final group is named as chance. It includes three subcriteria: sharp change in demand in the market, unusual fluctuations in production costs (e.g., bullwhip effect), and sudden change in the financial market and exchange rate. For example, the Covid-19 disease, which has recently emerged in China and has caused a major epidemic worldwide, has caused a sudden fluctuation in demand, directly affecting new hospital investments. is main criteria group has been handled by many scholars with different definitions and scope. For example, while a criterion heading of "Risk" is used in a study by Islam et al. [62] and Mahmud et al. [71], a term of "Disaster risk" is mentioned in a study by Assad [26]. However, the risk concept in the study by Islam et al. [62] and Mahmud et al. [63] is quite different from the meaning explained here. ey mean risk as a function of land risk, construction risk, and time frame and delivery speed. Some papers consider this a change in the market or market dynamics [33,39,64].
Apart from all these, another essential criterion group is the evaluations about the distance of the location to some points. In this context, subcriteria such as distance to existing hospitals, proximity to main roads, distance to dumping areas [28,29], distance to industrial areas [29,76], distance to social centers, distance to medical suppliers, distance to other governmental buildings, easy access to ambulances [28,44], distance to the city center, distance to other neighboring cities [36], distance to main arteries [56], distance to arteries and main roads, distance to other medical centers [54,55], proximity to residential units, and proximity to noise sources [54] are emphasized in the reviewed papers. In the study of Kaveh et al. [28], the selection criteria are directly related with distance. Seven criteria named as distance from the existing hospitals, distance from fire stations, distance from population centers, distance from road and street network, distance from green spaces and parks, distance from strong power lines, and distance from fault are used.
Moradian et al. [9] studied the risk criteria in hospital site selection and performed a systematic literature review. However, the scoping and time period of this review is limited. ey focus only on selection criteria and their classification of the studies published till 2017. at study does not include a comprehensive discussion on applied MCDM methods and their extensions. e criteria classification of Moradian et al. [9]ʼs review is executed under four groups as cost concerns, demand concerns, disaster risk concerns, and environmental concerns. In the light of all discussions as mentioned above and inferences from the reviewed 47 papers, we can suggest a generic hierarchical framework of the hospital location selection problem as presented in Figure 7. We believe this framework can benefit the scholars who will study the MCDM problems of hospital location selection [74,78].

Open Areas and Suggestions for Future
Work. e point to be noted here is that the importance of the "location selection problem" increases due to the need for rapid establishment of hospital investments during critical and disaster events. New criteria should be added to the hierarchical framework for hospital location selection in such disastrous situations, as explained in Figure 7. In this framework, some of them are mentioned within the main criterion of "Chance" and within the main criterion of "Related and supporting sectors." A patient influx in demand, variability in drug, equipment, and device costs, rapid change of economic indicators (financial bottlenecks that may be encountered in global pandemic conditions), and difficulties to be experienced in the pharmaceutical, material, equipment supply chain, and supporting sectors can be counted as new criteria that will directly affect new hospital investments.
From a methodological perspective, this problem is generally addressed through the MCDM concept and the methods created by integrating this concept with fuzzy logic. Apart from that, models supported with GIS have been included in the literature. In this context, it is seen that the innovations in the field of MCDM have not yet taken place at the desired level for this problem. It is expected that methods such as BWM that provide less pairwise comparison and a more consistent subjective evaluation can be used more in obtaining the relative importance weights of the hospital location criteria in future studies. However, at the stage of determining alternative locations according to weighted criteria, the newly proposed fuzzy set versions (such as Pythagorean fuzzy set, Fermatean fuzzy set, and q-rung orthopair fuzzy set) should be preferred more frequently. Well-known and applied MCDM methods such as

Study
Main criteria Shahbandarzadeh and Ghorbanpour [51] Factor conditions, geographical conditions, related and supported industries, government Yuen [77] Population density, land cost, community support, transportation convenience, competing hospitals, environmental favor Kaveh et al. [28] Distance from the existing hospitals, distance from fire stations, distance from population centers, distance from road and street network, distance from green spaces and parks, distance from strong power lines, distance from fault Zolfani et al. [53] Technological, economic, social PROMETHEE, DEMATEL, MAIRCA, MULTIMOORA, and TODIM can be used in this problem with these improved fuzzy set versions. e conventional MCDM methods (like AHP, TOPSIS, and VIKOR) do not consider data/information uncertainty, converting expertsʼ judgments into a crisp value. e general and extended fuzzy MCDM reflects the uncertainty of expertsʼ assessment environment but does not measure the degree of membership and nonmembership of the evaluated events simultaneously. In this case, intuitionistic, Pythagorean, and Fermatean fuzzy MCDM can gain importance and address and solve the problem by providing low, medium, and high uncertainty information levels, respectively.

Conclusion
e hospital location selection problem occupies an essential place among general facility location problems. Hospitals are not ordinary facilities. ey are facilities established for a vital need that manage many complex processes designed for patients to receive comprehensive health care. erefore, while making an investment decision regarding these facilities, determining the appropriate location makes it necessary to consider many different criteria. In this context, such a problem can be easily modeled as an MCDM concept and the most suitable alternative can be determined.
In this paper, a systematic and state-of-the-art review of 47 papers on hospital location selection is presented. All papers were classified by publication trend, published journal, country of origin, methods used to select the location, MCDM environment, type of decision criteria, and the main criteria in the selection process. e ultimate goal is to provide researchers and practitioners with a useful guide on the topic. e following are pointed out in the statistical figures from the retrieved articles. (1) Usage of AHP and GIS-based MCDM models is following an increasing trend compared with others (ones constructed by ANP, TOPSIS, and so on). (2) A scattered distribution is available in the published journals. (3) Turkey ranks first by accounting for almost 28% of all papers related to hospital location selection models. (4) Classical and general fuzzy MCDM concepts are the two most preferred environments based on the data in crisp and fuzzy numbers. Also, fuzzy triangular numbers are the most used fuzzy set version. e newly developed versions (intuitionistic, Pythagorean, and Fermatean) under fuzzy logic theory are suitable for this field. ey are extremely good in terms of the amount of uncertainty information covered. (5) Researchers mostly prefer cost, demand, environment, population, government, competition on the market, and distance to essential locations as main selection criteria. Given the Covid-19 pandemic results, where the health sector is currently in a great struggle, it is clear that the new hospitalʼs location decisions, field hospital, or pandemic hospital investments should be made quickly and reliably. is reinforces the conclusion that a possible rapid demand situation scenario (such as a pandemic, natural, and handmade aphids) that emerge from this study and has not been mentioned by most of the researchers in the review will play a significant role in the location selection decision. We expect and observe that, in the future, the number of applications and approaches related to hospital location selection will rise in the literature. e following suggestions can be enriched by the following future research. is study handles the problem of hospital location selection from the perspective of MCDM in the literature. Researchers could consider different topics regarding healthcare domain such as hospital service quality evaluation, hospital demand forecasting, and hospital facility layout design. ese rarely considered topics with/without MCDM concepts can be further studied. A new topic of future research is the development of MCDM approaches for disaster time location selection problem. is type of problem may vary from the normal time problem in terms of selection criteria. Various risk criteria may exist related to the disaster type.  Nomenclature ABM: Agent-based modeling AHP: Analytic hierarchy process ANN: Artificial neural network ANP: Analytic network process ARAS-G: Additive Ratio Assessment Method with Grey Values CoCoSo: Combined compromise solution CODAS: Combinative distance-based assessment CRITIC: CRiteria importance through intercriteria correlation D-AHP: Delphi analytic hierarchy process DF-AHP: Delphi fuzzy analytic hierarchy process EDAS: Evaluation based on distance from average solution ELECTRE: ELimination Et Choix Traduisant la Realité F-AHP: Fuzzy analytic hierarchy process F-ANP: Fuzzy analytic network process F-ELECTRE: Fuzzy ELimination Et Choix Traduisant la Realité F-MCDM: Fuzzy Multicriteria decision-making F-SAW: Fuzzy simple additive weighting F-TOPSIS: Fuzzy technique for order preference by similarity to ideal solution GA: Genetic algorithm GIS: Geographic information systems GRA: Grey relational analysis HF-TOPSIS: Hesitant fuzzy technique for order preference by similarity to ideal solution P-CNP: Primitive cognitive network process SAW: Simple additive weighting SF-TOPSIS: Spherical fuzzy technique for order preference by similarity to ideal solution TOPSIS: Technique for order preference by similarity to ideal solution VIKOR: VlseKriterijumska Optimizacija I Kompromisno Resenje BWM: Best and worst method PROMETHEE: Preference ranking organization method for enrichment evaluations PSO: Particle swarm optimization DEMATEL: Decision-making trial and evaluation laboratory TODIM: Interactive and multiple attribute decision-making MULTIMOORA: Multiobjective optimization by ratio analysis MAIRCA: Multiattribute ideal real comparative analysis.