Designing a Framework to Support Multi-Criteria Decision Analysis for Sustainable Public Facility Location: Insights from Centro Hospitalar Oeste, Portugal

Abstract

Finding the optimal location for regional public facilities has historically been a challenge. Numerous studies have sought to find the best solutions, yet few of them provided a framework to incorporate criteria and indicators that look beyond spatial dimensions, aiming at a more sustainable solution fostering economic development, territorial efficiency, environmental sensitivity and, of course, social justice. This research is based on the issues raised by the implementation of the new Centro Hospitalar Oeste (CHO) in Portugal, which has been extensively evaluated in the decision-making process for many years. Recognising that location decisions involve more than spatial concerns, this article proposes the identification of criteria and indicators that integrate social, economic and territorial perspectives to support the construction of a decision-making model based on sustainability principles. A four-phase, sixteen-stage multiple-criteria methodology is proposed to build a technical instrument for overcoming political conflicts and ensuring efficient, rational and participatory processes for public investment allocation. This article offers evidence from a real spatial planning issue to provide a useful framework to promote a more significant and comprehensive social and economic impact of the public investment made in a regional-level facility. This framework helps to determine the fundamental criteria of evaluation and operationalise sustainable and balanced models for the location of public facilities that take into consideration an integrated territorial approach. The methodology presented is open and flexible so that it can be applied and adapted to different contexts and facilities.

1. Introduction

Finding optimal locations for settling or relocating regional public facilities is always challenging. It is a complex decision-making process involving multiple stakeholders—including national, regional and local authorities, developers, public institutions, citizens and workers—that demands political trade-offs, extensive planning and negotiations [1]. This process can be pretty controversial, so sustainability is crucial to justify the expenditure of large amounts of public resources. Therefore, numerous instruments have been developed to support the most accurate technical solution for locating new public facilities and achieving more sustainable results.

However, little research takes a comprehensive perspective of the factors influencing decision-making and the regional consequences of a new facility implementation. Building on the available methodologies, this study proposes a framework to identify criteria and indicators that are an essential step for the preparation of a multi-criteria decision model that aims at a sustainable territorial solution. This framework takes into consideration a combination of social, economic and territorial dimensions, and therefore an integrated and well-balanced perspective of many factors influencing decision-making, while understanding the impacts of locating and/or relocating regional public facilities.

This research is based on the case of the new Centro Hospitalar Oeste (CHO) in the Oeste region of Portugal, which for many years had its implementation process discussed. Despite the political issues involved in this process, this article focuses on the development of an efficient framework to support the construction of a model for defining a sustainable location for a regional public facility. The main assumption driving the work was that the models traditionally adopted to support locational decisions, being monothematic, end up reducing the efficiency of public investment and limiting the impact on the regional and local development of the facility in question.

This contribution takes into consideration an integrated approach that looks into the configuration of the regional urban system and the existing healthcare, social, environmental and economic conditions.

The methodology is based on a phased approach, considering comprehensive aspects for locating or relocating public facilities This broader, integrated and sustainable evaluation framework is proposed in four phases and sixteen steps. The first phase investigates the starting assumptions to define the best location, identifying the influence area of the facility and the urban centres’ possibilities. Then a multi-criteria evaluation is proposed in phase 2 at a macro-location scale, to indicate the best location among analysed urban centres. Phase 3 addresses the micro-location, evaluating the best implementation site for the given facility. The final phase considers the mitigation measures.

After the introduction, the article is divided into five sections. The second part discusses the challenges of locating central public facilities, presenting a literature review on methodologies and suitable criteria. The third section presents the materials and methods. The methodology is presented in the fourth section, discussing the criteria, sub-criteria, steps and procedures adopted. The fifth section presents the discussion of the results presenting the key research findings and the advantages and limitations of the method. The sixth and final section offers the conclusions.

2. The Challenge of Locating Regional Facilities

2.1. Principles of Location Theory

Since the 17th century, the optimal location problem has been an issue under investigation. Initially, through mathematical, quantitative and theoretical approaches (Pierre Fermat, 1623–1626), foundations for locational studies have developed towards more sophisticated methods that consider numerous variables, such as economic, social, environmental and political criteria. The literature review gathers examples of methodologies and demonstrates that most of them are still based on a purely spatial approach, leaving aside other relevant aspects to determine efficient location.

The history of locational analysis takes us to the initial models developed by von Thünen [2] and Weber [3,4]. They were followed by Christaller [5] and, subsequently, Lösch [6], who made seminal contributions to location theory known as the Central Place Theory (CPT), seeking to explain the spatial logic of the system of cities, addressing the number, patterns, size and functions of commercial centres. Another relevant contribution is that of Reilly [7], who described the law of retail gravitation and argued for the attractiveness of large retail centres for customers.

Peter Haggett’s Location Analysis in Human Geography [8] provided an organized textbook that became a classic in the field [9]. Considering the relationship between spatial efficiency and spatial equity in the definition of optimal locations, Morrill and Symons [10] propose to understand the conditions under which an efficient location could be also equitable or the circumstances that could cause an efficient location to be inequitable, analysing the role of density, income and accessibility.

In 1993, Harris and Batty [11] started to question the possibilities and difficulties that Geographical Information Systems (GISs) could provide for the simulation, organisation and design of activities concerned with spatial planning. They provided initial frameworks for locational models using GIS, indicating requirements for planning support systems and inaugurating a new path in computational analysis.

Addressing the issue of equity in facility location models, Marsh and Schilling [12] contributed with a framework to organise ways of measuring it. Drawing on previous studies and their limitations, the authors propose a base from which equity concerns can be integrated into facility-siting models.

Concerned with a regional planning context, Lopes [13] discusses the problems related to operationalising regional development policies addressing spatial issues, regional interdependence and multidisciplinary approaches. On an urban scale, Camagni [14] deals with organising productive activity and incorporating sustainability and economic development themes.

Deverteuil [15] analysed the legacy of locational theory, describing the evolution of research from refuting and sustaining those initial studies. According to the author, from more quantitative approaches aiming at efficiency and equity, studies have progressively moved towards more sophisticated qualitative approaches overseeing the distribution of costs and benefits and valorising the contextual dimensions.

Moving beyond the significance of spatial location, J.C. Ribeiro and Santos [16] argue for the importance of relational proximity to explain the advantages of dynamic location (not only physical location). The recent advances in analytics and big data have also had an impact on location analysis. Pick et al. [17] offer an updated literature review on location analytics (LA), non-location analytics (NLA) and decision support (DS), covering the new problems, methodologies and applications in the light of the complex datasets available.

The models proposed to solve the location problem show that quantitative spatial approaches are still dominant to support location theories. Complex mathematical models and sophisticated GIS algorithms support public facilities’ location or relocation strategies. Yet, other knowledge bases might also inform the decision-making process of facilities’ location. Barnes [4], for instance, has already criticised a pure rationalist interpretation of location analysis, arguing for the relevance of incorporating local knowledge. The most common technique, the multiple-criteria decision analysis, is described in the following subsection.

2.2. Multiple-Criteria Decision Analysis (MCDA) for Locating Regional Facilities

Multi-criteria analysis is a widely used technique in decision-making processes, involving the analysis of multiple criteria, which can be qualitative or quantitative, and assigning weights to each criterion based on its relative importance concerning the objectives or externalities (positive or negative) resulting from the decision-making process. MCDA, also called MCDM (multi-criteria decision-making), has different methods and tools to define the best solution among options from a decision-making point of view [18].

In the public sector, multi-criteria analysis is relevant to support investment decision-making, assisting policy decisions with a technically grounded assessment that ensures the most effective, efficient and sustainable use of public resources. These approaches provide a structured framework for evaluating different alternatives (options), considering multiple criteria and preferences, allowing for dealing with uncertainties and risks, and identifying the most appropriate investment decision.

Due to the complexity and diversity of aspects to be considered in choosing the optimal location of hospital units, this public policy issue has been the subject of intense reflection and research. MCDA are the most used among the multi-criteria approaches and has different types of analysis such as the Analytic Hierarchy Process (AHP) and the Analytical Network Process (ANP) [18]. The production of scientific knowledge by applying MCDA focuses on two different dimensions. On the one hand, studies have been produced to design integrated analysis models that contemplate various variables. On the other hand, research encompasses analysis based on specific criteria such as accessibility.

The following subsection will focus on the problem of the location of healthcare facilities, as it has been the focus of the methodology of this article. From our perspective, the decision regarding the location of a regional hospital involves much more than an equitable spatial measure, and this issue will be discussed further.

2.3. The Problem of the Location of Healthcare Facilities

Models for location optimisation and service planning of healthcare facilities have been put forward since the 1960s [19]. Some examples can be seen in the works by Godlund [20], Teitz [21] who inserted a systemic viewpoint as well as Morrill and Symons [10] and McAllister [22].

Lately, GIS models have been adopted using a single or a multi-criteria approach. Using a specific criterion to determine an optimal facility location, different studies have focused on accessibility. V. Ribeiro et al. [23] evaluate different location-allocation models using GISs to support the planning of healthcare facilities regarding accessibility level. The study by Alhothali et al. [24] also used accessibility to define the optimal location of COVID-19 vaccination centres. Other studies will focus on the incorporation of transportation modes such as the method proposed by Mao and Nekorchuk [25], in which a model is proposed based on the 2-Step Floating Catchment Area Method (2SFCAM) framework.

A literature review by Dell’Ovo et al. [1] relating MCDA (Multi-Criteria Decision Analysis) and GIS methods demonstrated that few published papers had focused on the location of healthcare facilities. Further references on the location of healthcare facilities and evaluation tools can be found in a book by the same authors [26]. One important study on this topic proposed multiple linear regression and geographically weighted models to analyse the optimal distribution of hospitals and other healthcare centres in a city [27].

Addressing the problem of hospital location selection adopting multiple criteria, the research conducted by Teixeira and Antunes [28] focuses on accessibility, demand, hierarchy of facilities and maximum and minimum capacity to provide a hierarchical location model for public facility planning. It considers different demand levels and varied types of facilities related to the p-median objective. Batta et al. [29] adopt a p-median objective to demonstrate that the appropriate use of dispersity, population and equity parameters can lead to good results when locating public facilities.

Mestre et al. [30] discuss two different location-allocation models for the strategic planning of hospital networks in the face of uncertainty. Considering first-stage decisions for location, the first model is proposed as the second takes into consideration location and allocation. The study points out that the latter can be more flexible as it is not scenario-dependent.

Gray Relational Analysis (GRA) and Analytic Hierarchy Process (AHP) were adopted by Şen and Demiral [31] to provide solutions towards hospital locations. The criteria used are site conditions and surroundings, accessibility and traffic, patient/emergency access consideration, cost, future considerations (such as expansion capacity) and nuisance.

The proposal by Dell’Ovo et al. [1] is to combine spatial analysis with MCDA to provide a decision support system for locating healthcare facilities. The work considers functional, locational, environmental and economic aspects, providing useful indicators and measurement criteria. Caprioli and Bottero [32] propose a fuzzy spatial multicriteria analysis for identifying suitable locations for urban infrastructure, including new healthcare facilities. The study is conducted using the Analytic Hierarchic Process (AHP) and Fuzzy AHP (FAHP) and focuses on an evaluation model based on criteria for locational analysis clustered in four groups: location, mobility, environmental and social.

A systematic review elaborated by Gul and Guneri [33] identified that the AHP and GIS-based MCDM (multi-criteria decision-making) methods are the most common approaches applied. The study also identified the most used criteria as cost, demand, environment, population, government, competition in the market and distance to important places.

Loumeau [34] adopts a multi-criteria approach to evaluate public facilities’ location decisions. The approach suggests combining the standard distance-minimising problem with particular location decisions of individuals, exemplified as housing costs, commutes to work and the presence of existing public/private facilities. Most recently, focusing on equity and fairness in spatial location, Giovanniello and Tonin [35] propose an envy-free model with heterogenous individuals that accounts for differentiation in quality (vertical) and location (horizontal) of public good facilities.

Tackling the Portuguese context, Gonçalves et al. [36] have already discussed the challenges of locating the Hospital do Oeste Norte. On that occasion, the approach suggested a complex hierarchy-based process, proposing the segmentation into macro-location and micro-location and specific indicators for each. This idea will be recovered in the methodology suggested in this paper.

However, studies show that spatial distance might not be the only important factor in evaluating hospital use. Moscelli et al. [37] and Varkevisser and Van Der Geest [38] focus on patients’ decisions regarding which hospital to visit according to the presented health issue, understanding which factors influence this choice according to evidence from England and Sweden, respectively. Both studies show that numerous patients did not use the nearest hospital. The hospital’s good performance, lower waiting time and patient age and social status affect the choice. A study developed in United States metropolitan areas has shown that socioeconomically disadvantaged neighbourhoods might also affect hospital quality [39].

2.4. Research Gaps Identified and Shortcomings

The literature review shows a variety of methods and tools to perform types of multi-criteria decision analysis, whether involving GIS procedures, AHP or ANP techniques or taking into consideration one or more variables. Most of the studies gathered here take spatial features such as accessibility, mobility, land value, presence of infrastructure and green areas, among others to evaluate possible locations for a new facility. Multi-criteria evaluation is, therefore, a handy tool to assist in this decision-making process since it allows for the simultaneous consideration of various relevant factors and viewpoints.

The referred studies show that in hospital location decisions, the most common criteria include accessibility, demography, infrastructure availability, plot costs, qualified workforce availability and security [1,25,26,27,30,31,33,40,41]. Yet, those criteria can be limited, missing the whole picture when taking into consideration the spatial, economic, environmental and social effects of a new regional facility.

This article proposes to fill this gap, expanding the set of criteria to be adopted as well as the indicators. If new criteria are added, the regional impact of a new facility can see more ambitious consequences. On the regional scale, the inclusion of a new public facility can promote new dynamics of development. In this way, our concern in this article is to provide an important step before the definition of a decision-making model. We believe that defining correctly a set of criteria and indicators to address these challenges is essential for the efficient operationalisation of a given model [42].

Moreover, the location problem concerning public facilities has two crucial shortcomings that must be emphasised. On the one hand, there has been a reduction in new significant public investments because territories in Europe have been progressively equipped over decades [43]. There has even been a reduction in the number of healthcare facilities, as advances in medicine and associated technology have led to hospitalisation being dispensed and treatments being favoured on an outpatient basis or even at home [44].

Exceptions to this framework arise from the need to overcome the obsolescence of existing facilities and rationalise the supply directed at a given area [45]. Thus, although less frequent, the construction of new hospitals is still a significant investment, especially in a context of scarcity and competition for public resources, aggravated by restrictions on public indebtedness imposed by national and international rules [46].

On the other hand, public investments, precisely because they become a collective endeavour, incessantly seek the best levels of efficiency [47]. In the case of healthcare, a hospital, alongside its role as a service provider, must be able to expand and diversify this offer, with effects on science (in combination with research activities), education (in conjunction with health teaching activities), the economy (in conjunction with economic investments in the pharmaceutical and biomedical engineering areas) and the territory (in conjunction with urban and regional planning and development instruments and complementary infrastructures and services). In other words, a hospital can be seen as more than just a healthcare provider, fostering competitiveness and territorial cohesion [48].

This article offers a new framework to analyse regional facility location, questioning the methods of locational determination based solely on a distance–time problem that has been applied to the potential demand for health services. It proposes designing a new model capable of articulating the renewed complexity with which the location of a regional hospital should be considered and offers a comprehensive interpretation of the decision-making process that considers both spatial features and social, economic and environmental aspects.

3. Materials and Methods

This section is divided into two parts. The first addresses the case study, pointing out the issues identified in that decision-making process that were used for the elaboration of the methodology. The second part focuses on the description of the methodology framework providing a scheme of the necessary procedures in decision support systems. Decision-making regarding the location of a hospital is a complex problem that involves considering multiple criteria and preferences of various stakeholders (patients, workers, suppliers, valued sectors, local authorities, etc.).

3.1. Case Study Presentation: Centro Hospitalar Oeste

The contribution offered in this article was based on the proposal of an efficient decision-making process to solve the relocation problem of a regional hospital located in the Oeste region, Portugal. The Oeste region is composed of thirteen municipalities: Alcobaça, Alenquer, Arruda dos Vinhos, Bombarral, Cadaval, Caldas da Rainha, Lourinhã, Nazaré, Óbidos, Peniche, Rio Maior, Sobral de Monte Agraço and Torres Vedras (Figure 1 shows the composition of the region). The Oeste region is located in a corridor between the two metropolitan regions of Porto and Lisboa and has a population of almost 380,000 inhabitants, representing 3.6% of the Portuguese population. Lately, the region has seen relevant economic and demographic growth if compared to the national context. This expansion of economic activities in the region is based on the growth of agro-industrial segments and tourism.

The deterioration in the quality of hospital services in the Oeste and their inadequacy to meet the needs of the population is a consensual problem, present in the most diverse studies and opinions carried out over the last two decades, which makes it urgent to decide on the location of a new hospital. The shortcomings of the public offer, together with the growing functional integration of the municipalities in the southern sector into the Lisbon Metropolitan Area, has made the population of the southern municipalities seek out the hospitals in the neighbouring region; in the northern sector, the municipality of Nazaré and part of the municipality of Alcobaça have been integrated into the area of influence of the city of Leiria; and that in Torres Vedras the private hospital offer has expanded.

The delimitation of the Centro Hospitalar Oeste’s (CHO) area of influence, in coherence with the reality of current hospital demand, is a decisive aspect of the credibility of decision-making on the location of the hospital. The development of this methodology included the analysis of the health facilities offered in the Oeste region and the systematisation of the previous studies elaborated to define the relocation of the CHO. Three studies have been analysed: one prepared in 2007 [49], another in 2009 [50] and the last one prepared in 2022 [51]. Location proposals can be seen in Figure 2.

Controversies identified in this decision-making process to locate the new facility led local administrations to reach for a scientific-based solution that could offer a more adequate, just, robust and rigorous method to support the location of this new regional facility. This article emerges from a reflection about the evaluation process leading to location decisions. In this way, the methodology proposed to evaluate the benefited population, hospital users and workers, and city residents.

In this article, we chose not to discuss all the political, socioeconomic and spatial implications of previous studies. Instead, we focus on the methodology developed as an answer to the challenge posed by the municipalities involved in this decision-making process. It was important to create a decision-making process that contemplated multiple users and stakeholders, moving beyond spatial dimensions. Drawing from the positive aspects and the limitations of previous studies, this methodology has been proposed as a solution towards a more comprehensive and transparent decision-making process.

3.2. Methodology Design Framework

The following paragraphs describe the elements taken into consideration for the methodology creation and the way it was structured. Bearing in mind that a variety of stakeholders should be attended to by this method, the methodology considers the accessibility of potential users, the community of workers and the conditions for those who are dislocated, thinking about the needs of their families to reside in proximity to their job with a good quality of life and the conditions to attract and settle qualified new professionals. The proposed framework also includes environmental externalities, accessibility costs, mobility conditions, potential economic externalities (negative and positive) and research and development opportunities to boost regional development.

This methodology sheds light on the decisive assumptions and the criteria (with associated indicators) to make the location decision-making process possible, comprehensive, rigorous and robust and to facilitate an objective, transparent, systematic and publicly accountable assessment of the available options.

Firstly, it is essential to acknowledge the structuring role of the regional development and spatial planning framework and guidelines that determine the territorial models and the possible areas with the capacity to provide these regional central functions. Secondly, aligning with other stakeholders, the area of influence and the type of facility is a crucial point of departure to initiate the location analysis. Thirdly, adopting a multi-criteria framework is the most suitable approach for attending to as many dimensions as possible, such as social, economic and environmental towards a sustainable solution.

The methodology is based on a four-step approach. The first phase focuses on the starting assumptions, such as the definition of the influence area of the facility and the identification of possible urban centres that fulfil all essential needs. The second phase is a multi-criteria evaluation of the proper urban centres based on a macro-location analysis, proposing a performance ranking according to several criteria. Phase 3 investigates the alternative location within the urban centres, focusing on the micro-location to indicate the best location for the facility implementation. The fourth and final phase considers the mitigation measures.

This study was conducted to support the decision to the location of a regional hospital facility using a Portuguese real case study carried out by the authors however, it might be useful not only for different contexts but also for supporting the location of other types of highly differentiated urban functions. The methodology shows how academia can produce applied research to inform and support planning practices and political decisions. It offers a broad vision of a location problem that moves beyond obvious spatial effects, also pondering the possibilities of attraction of investments for the city-region development and the sustainability of the spatial planning framework. Considering economic and social externalities, demographic circumstances and attending planning regulations, the methodology offers the possibility to conduct a rational, robust and participating decision-making process.

Currently, decision-makers are called to examine solutions for problems that have become even more complex with time. Models created to provide resources to structure these problems should include the total of parameters that organise them, given their systemic character. Decision-making models should be able to classify problems according to their typology, map territorial and inter-section interactions and understand the challenges of local, urban and regional development. We designed a scheme that addresses the relevant characteristics that are part of the decision-making environment (Figure 3). The capacity to comprehend and integrate these broad dimensions should indicate a better result for robust decision-making and ensure clear recommendations.

The scheme is based on three steps (i) inputs, (ii) processing and (iii) outputs, to which different activities and functions are attributed. The inputs include data storage and models library, both oriented for decision-making. The processing step addresses the organisation of the parameters of the problems, the typology of the problem, events and policies simulation and the proposal of solutions for a given decision-making problem. Processing feedback can be used as input. The last step addresses the outputs of decision-making processes, such as the status and progression reports, parametrisation and results and finally, the recommendations. The decision maker should make informed decisions based on the support of technological platforms, programming and interfaces.

A variety of MCDA methods can be adapted to move further with the decision process, given that a robust set of criteria and indicators are structured. The first benefit of MCDA methods is their ability to incorporate the perspective and preferences of multiple stakeholders regarding an investment, resulting in more transparent, inclusive and acceptable decisions for all parties involved. Generally, the benefits and externalities of public investment are multidimensional, generating different viewpoints from stakeholders and interest groups on the aspects that should be considered for decision-making.

Another benefit of multi-criteria analysis is its capacity to deal with uncertainties and risks affecting the expected outcomes. Due to their complexity and scale (space and time), this is even more crucial in investment processes that only materialise the impacts much later than the decision-making moment and spread out in distinct spatial, political, social and economic contexts.

Multi-criteria analysis can also assist in identifying the most suitable investment option from a set of alternatives, providing a structured framework for evaluating and comparing these options based on multiple objective descriptors. This enables decision-makers to identify the most suitable choice that maximises benefits and minimises risks and costs. The next section describes the methodology designed to support the creation of more comprehensive decision-making models.

4. Designing a Multi-Criteria Methodology to Operationalise Public Facility Location Models

4.1. A Step-by-Step Methodology Proposal

The construction of a new regional hospital unit has relevant implications for the urban centres where hospitals are currently located in the Portuguese Oeste Region. The fact that it involves the relocation of healthcare services raises security issues for local communities and it enhances the natural competition among municipalities to capture this investment. Adopting a comprehensive, transparent and stakeholder-inclusive approach in the decision-making process is essential for legitimising the decision and promoting a sustainable regional framework.

For this objective to be achieved, this methodology is proposed as a robust tool to facilitate the construction of a sustainable decision-making model. We argue that the process of conducting the multi-criteria assessment should be preceded by a clear and objective definition of the catchment area of the new hospital facility. It should be based on the territorial model planned by regional and local regulations with the identification of urban centres that, due to their size and role in the regional urban system, are suitable for hosting a regional-scale facility.

The methodology suggests that the evaluation should be conducted in two steps and at two scales. Firstly, an analysis is performed at a regional scale (macro-location), ranking suitable urban centres based on criteria reflecting various objectives and conditions for the hospital unit. This aims to provide the best conditions of safety, cost and comfort for its users and workers, to minimise environmental externalities and generate economic benefits for the region. Secondly, at an urban scale (micro-location), the different alternative sites are evaluated within the highest-ranked urban centres.

Figure 4 represents the flowchart of the decision-making process supported by a multi-criteria analysis divided into sixteen stages. The first phase addresses the initial assumptions, including Stage 1, the definition of the catchment area of the hospital and Stage 2, the identification of urban centres with the capacity to accommodate such a facility, according to national and regional territorial plans.

The second phase of the multi-criteria evaluation process aims to rank the regional urban centres identified in Phase 1 based on their performance, which is measured using a set of criteria and subcriteria to assess the existence of favourable conditions, addressing the macro-location.

Stage 3 proposes the definition of criteria and subcriteria for the evaluation. The literature review highlights accessibility, economy, demography, mobility and urban infrastructure as some of the most important criteria. Contextual aspects such as tourism could also be relevant, an essential criterion for the Portuguese case.

Stage 4 addresses the establishment of weights for the chosen criteria. Several techniques can be used to determine the weights, including (i) the direct judgment method, where the experts responsible for the evaluation assign weights directly to each criterion based on their relative importance; (ii) the pairwise comparison method, where experts compare each criterion with all others and assign weights based on these comparisons; (iii) hierarchical analysis method where criteria are grouped into hierarchical levels and weights are determined through an analysis of relative priorities.

Stage 5 is the identification of indicators to measure the subcriteria. The measurement of each subcriterion should be carried out using statistical or other quantitative indicators obtained from official sources, which should be as updated as possible and at the most relevant and precise spatial scale possible.

Stage 6 proposes quantifying indicators of analysis, using the most recent available information. Stage 7 addresses the normalisation of scores, an essential process to ensure that different criteria and subcriteria can be used in the final assessment. Other normalisation methods can be used as long as they are consistent throughout all the requirements.

Stage 8 is the classification of the compatible urban centres. The hierarchical ranking of urban centres is obtained from calculating the final score for each urban centre under analysis as a result of quantifying the indicators and their normalisation based on assigning weights to each criterion and subcriterion.

The following phase, number 3, addresses the micro-location. After completing the multi-criteria evaluation conducted in Phase 2 and identifying suitable urban centres offering the best conditions, a comparative analysis of the available sites within the chosen urban centre should be performed. Stages 9 to 15 are part of this phase.

Stage 9 is the identification of possible locations within the qualified urban centres. With the support of local authorities, the different locations within the chosen urban centre that are suitable, available or meet preliminary conditions for installing the hospital facility should be identified.

Stage 10 proposes the definition of criteria and subcriteria to evaluate the possible locations. The alternatives identified previously will undergo a multi-criteria evaluation process identical to that conducted in the previous phase but related to attributes that are predominantly local. The criteria and subcriteria used should address the physical aspects of the location, its suitability for the construction of a hospital facility and its integration into various urban infrastructures and services. Criteria related to the environment, accessibility and transportation, formal land registration and ownership conditions, physical characteristics and the need for preliminary work could be used.

Stage 11 establishes the thresholds and weights for admissibility. Establishing weights should follow the previous process and methodology for evaluating suitable urban centres. Considering that some criteria are descriptors of values (e.g., classified areas or areas subject to constraints) or risks (e.g., flood risk) where the existence of thresholds for admissibility to the construction of a new facility may be placed, a minimum acceptable value of exposure or affectation can be established from which the option is no longer adequate.

In stages 9, 10 and 11 it is proposed to listen to stakeholders, with objective implications for the respective contents. Given the delicacy and subjectivity of these contributions, especially in more complex social and economic environments, we suggest adopting public participation methodologies capable of mitigating the effect of individual opinions by valuing collective readings, as suggested by Amini et al. [52].

Stage 12 seeks to identify the indicators to measure subcriteria. Given their specific focus, the construction of these indicators cannot be carried out solely from statistical sources; the work must be conducted with the support of the municipality of the selected urban centre from the previous phase. A characterisation sheet for each alternative should identify the information used and its sources.

Stage 13 is the normalisation of scores, and stage 14 is the ranking of possible options for choosing micro-location. The methodologies should be the same as used in stages 7 and 8. The 15th and last stage addresses the recommendation for a decision with the best location determined.

The final phase, Phase 4, defines impact mitigation measures. Defining mitigating measures for the impacts in different cities and regions, depending on where the new facility will be located or relocated, is indispensable, as public facilities such as hospital units establish numerous relationships with the social and economic context in which they operate. Whether on access to hospital services or the local economic base, preventive measures aimed at avoiding or reducing negative impacts before they occur should be planned to minimise impacts. In the same way, mitigating measures aimed at reducing or minimising adverse effects that may arise should be foreseen. The definition of mitigating measures should be supported by a detailed assessment of the potential impacts resulting from the closure of hospital units as well.

4.2. Macro-Location: Criteria and Indicators

This methodology includes a proposition of the criteria, subcriteria and indicators that can be used for assessing suitable urban centres. Firstly, the macro-location and the requirements for accessibility, mobility, economy, demography and infrastructure are justified and detailed. A synthesis of all criteria and indicators proposed is provided in

Table 1. Matrix of criteria and indicators proposed by the methodology for the macro-location assessment.

Subcriterion		Indicators
Accessibility	Time distance from the population that used the urgency to the facility to the suitable urban centre	The sum of time distance from the origin of urgency episodes (last five years)
	Time distance from the population that went to the facility for scheduled appointments to the suitable urban centre	The sum of time distance from the origin of the population with scheduled appointments (last five years)
	Time–distance in commuting for workers in the facility	The sum of the time–distance of residence of workers
	Costs of transportation for the population residing in the influence area of the hospital to the suitable urban centre	Average costs of transportation (potential users)
	Costs of transportation for the workers of the hospital to the suitable urban centre	Average costs of transportation (workers)
	Estimated volume of emissions of the potential users to the suitable urban centre	Total emission (average value/km2)
	Estimated volume of emissions of the workers to the suitable urban centre	Total emission (average value/km2)
Mobility	Population served by public transportation in the suitable urban centre	Population in a buffer of 1000 m from the public transportation lines
	Frequency of service of public transportation in the area	Hierarchy of public transport lines for the frequency they run
Economy	Total jobs in the suitable urban centre	Total number of private jobs
	Companies in the suitable urban centre	Total number of companies
	Jobs in healthcare services in the urban centre	Number of jobs in health services
	Presence of higher education facilities	Number of higher education students
	Capacity of touristic accommodation	Number of tourist beds
	Economic effects of closure/relocation of the facility	Number of job losses
Demography	Proximity to the residing population included in the influence area of the facility (present)	Demographic centre of gravity of the total population
	Proximity to the elderly population included in the influence area of the facility (present)	Demographic centre of gravity of the elderly population
	Proximity to the residing population included in the influence area of the facility (in 2040)	Demographic centre of gravity of the total population
	Proximity to the elderly population included in the influence area of the facility (in 2040)	Demographic centre of gravity of the elderly population
	Proximity to the women of reproductive age population included in the influence area of the facility (present)	Demographic centre of gravity of the women of reproductive age population
	Proximity to the young population included in the influence area of the facility (present)	Demographic centre of gravity of the young population
Infrastructure	Housing offer	Housing offer
	Primary, middle and high school availability	Schools’ offer
	Pre-schools availability	Pre-schools’ offer
	Nurseries availability	Nurseries’ offer
	Primary healthcare units’ availability	Primary healthcare units offer
	Cultural offer	Total capacity of cultural centres
	Green areas availability	Green areas/inhabitant

.

Concerning users’ and workers’ accessibility, the subcriterion time distance from the facility to the suitable urban centre is proposed. The time distance measures used are that from the population that used emergency services at the hospital facility in the last five years and that had doctor appointments in the previous five years and from the dislocation of personnel from home to work. The sum of time distance is the indicator suggested.

Costs of access from residents and personnel to the suitable urban centre is another criterion suggested, and the average cost indicator is suggested. Another relevant criterion related to accessibility is the estimated emissions associated with users’ and workers’ dislocation to suitable urban centres. Total emission in an average value/km should be adopted as the indicator.

Regarding the mobility criterion, two subcriteria are suggested. The first is assessing the population served by public transport in the suitable urban area, and the second is the frequency of that public transport service. Examples of indicators that could be adopted are the population in a buffer of 1000 m around public transportation lines and the hierarchy of public transportation lines.

The economy criteria are related to understanding the economic impact of the implementation and or/relocation of a healthcare facility in an urban centre. Subcriteria suggested are the number of jobs, number of companies, number of jobs in healthcare, presence of schools and universities, capacity of touristic accommodation and economic effects of closure/relocation of the facility (measured by job losses).

The demography criterion seeks to understand the potential population to be served. The subcriteria listed relate to the proximity of residents, elderly, women of reproductive age and youth residents (below 14) at the moment of analysis and in 20 years. The indicator demographic centre of gravity is used.

The last criterion, infrastructure, is used to analyse the conditions of the suitable urban centre. The subcriteria suggested are housing availability, educational availability (pre-school, middle and high school), nursery availability, primary healthcare units’ availability, cultural offer and green areas availability.

4.3. Micro-Location: Criteria and Indicators

The methodology also proposes criteria to assess possible locations within the selected urban centre, addressing the micro-location. The evaluation criteria to be used have a very different nature from the previous ones. They focus on specific areas and their purpose is to assess the suitability of a given land for the construction of a hospital unit.

Similarly, the indicators to be developed result from cartographic analyses that can be carried out with the support of municipalities, distinguishing them from the previous ones where there is a major use of statistical indicators. The following criteria, subcriteria and indicators are presented without prejudice to privileging a participatory approach of all stakeholders in their identification. A matrix of the criteria and indicators used for assessing the micro-location is shown in

Table 2. Matrix of criteria and indicators proposed by the methodology for the micro-location assessment.

Subcriterion		Indicators
Environment	Exposure to risks	Extension of the area exposed to flooding, seismic and landslide risks
	Exposure to noise	Extension of the area exposed to daytime, afternoon or night risks
	Environmental conditionals	Extension of the area located in conservation areas (Ecologic or Agriculture)
	Other conditionals	Location of water lines, power lines or environmental infrastructure
	Land use classification	Percentage of urban land available
Accessibility and Transportation	Accessibility	Distance to roads with higher hierarchy and the number of urban accesses to the implementation area of the new facility
	Transportation network	Distance to the interface of road transportation, the distance to the railway station and the number of connections to the public transportation system
Formality	The legal nature of the land is measured	percentage of the plot composed of public property
	Cadastral situation	number of parcels to be acquired and the number of landowners
	Cost	Cost of land
Physical	Available area	Extension of available area
	Configuration	level of adaptation of the plot to the use
	Physical conditions	suitability in terms of solar exposition, slope and lithology
Necessity of previous work	The extension of demolitions and adaptations	Extension of demolitions and adaptations to cover
	Procedures for exclusion	Extension of the area to exclude
	Security and civil protection	Distance to the firefighters’ station

.

The first criterion, Environment, is composed of five subcriteria. The initial subcriterion is the exposure to risks, measured by the indicators of the extension of the area exposed to flooding, seismic hazards and landslides. The following is noise exposure, measured by identifying the extension of the area exposed to elevated noise levels during daytime, afternoon and night. The next subcriterion is environmental conditions, which can be measured by understanding the extent of the area situated in conservation areas (that might be either ecological or agricultural). Other conditioning circumstances can be the location of waterlines or power lines. Land use classification is another subcriterion that can be used to understand the percentage of urban land available in the possible locations.

The second criterion is Accessibility and Transportation. The subcriteria Accessibility uses the indicators distance to roads with higher hierarchy and the number of urban accesses to the implementation area of the new facility. Regarding Transportation, the distance to the interface of road transportation, the distance to the railway station and the number of connections to the public transportation system are the indicators proposed.

The third criterion for evaluating micro-location is Formality. This criterion intends to measure the legality and costs of the land available for the new facility location. The subcriterion related to the Legal nature of the land is measured by the plot percentage of public property. The next subcriterion is the Cadastral situation, in which the number of parcels to be acquired and the number of landowners should be measured. Cost is the last subcriterion.

The physical characteristics of the area are the fourth criterion. The extension of available area, the level of adaptation of the plot to the use and the suitability in terms of solar exposition, slope and lithology should be assessed.

Finally, the last criterion relates to the necessity of previous works to implement the new facility. The extension of demolitions and adaptations should be carefully assessed. Distance to the firefighters’ station should also be taken into consideration.

5. Discussion

The methodology provided in this article is used to fill a gap regarding the location of public facilities. Most available methods provide a spatial location model, not considering other dimensions that can be fundamental for a spatial support system based on a multi-dimensional sustainability assessment. Even though a final model is not the purpose of this research, we provided an important step for a solid location model design, which is a decision-making framework with a designated set of comprehensive criteria and indicators aimed at a sustainable location solution.

The location of the new regional hospital should be supported by a multi-criteria evaluation of various alternative options since the choice of a hospital location is a complex decision that must consider a wide range of aspects. An integrated model should be capable of assessing different aspects such as spatial, social, economic and environmental and propose sustainable solutions to balance territorial issues.

These aspects were discussed throughout the article and some of them were already addressed by previous studies discussed in the literature review section, such as accessibility for potential users and relocated human resources, accessibility costs, environmental externalities and mobility conditions. Yet, other aspects should also be part of the assessment, for instance: conditions to leverage economic and Research and Development effects that accelerate regional development; conditions for relocated workers and their families to reside nearby with quality of life; conditions for attracting and retaining new skilled professionals in the future, meeting their life expectations. Considering those extra conditions is essential for sustainable regional planning.

Moreover, this methodology has considered the assessment of a new location and the externalities related to the relocation of an existing facility. The need for a comprehensive, rigorous, transparent and objective process is reinforced by the fact that the decision-making could also face the closure of the current hospital units. This would probably lead to negative impacts on the local economy and a potential perception of losing access to healthcare services.

The methodology suggested comes into place after reviewing the academic literature related to locational decision-making and identifying gaps and shortcomings. The comprehensive vision proposed allows for an objective, auditable, transparent and systematic framework for evaluating options and the participation of various stakeholders and interest groups. The vast publications on MCDA, location-allocation models and locational decisions can provide different mathematical models to support evidence-based locational decisions, however, most of the literature is based mainly on spatial features.

Moreover, many studies will focus on one case study, providing data and criteria that are suitable to one context and that should be modified if adopted in other cities. Despite the complexity of the models offered, most of the referred examples do not reflect on indicators and criteria that address the multiple stakeholders and communities affected by the location decision. Considering that other aspects that are not directly connected to the decision can still influence urban development and communities was the fundamental departure point of this research.

The article sheds light on challenges identified in previous studies: (a) the scale of evaluation, to which the methodology suggests a two-scale approach addressing the macro- and micro-location [36]; (b) the decision-making environment, taking into consideration the multiple stakeholders involved in the process, considering socioeconomic (population, hospital users and workers) and spatial aspects reflected on defined criteria and indicators; (c) the introduction of a phased method to approach the problem; (d) an adaptable and scalable methodology that can attend different spatial contexts and types of public or collective facilities—indicators can be hierarchised according to the context as well pondering and weighting that can be customised for each specific case study

The limitations of this methodology relate to the availability of information that might be accessible to the realisation of an efficient assessment since data might not be updated or not available at all. Another relevant limitation to be acknowledged is that this methodology demands a high level of governance capacity in the analysed case study. Aggregating dimensions and stakeholders make sense in a location analysis when a mature civil society with active participation and with capacity to integrate synergies and bottom-up or side-by-side strategies is at stake.

Even though this methodology emerged from a specific context of relocating a hospital in Portugal, it is adaptable. The difficulties faced in this process are probably similar to the situations encountered in cities, and therefore, the methodology could be used in different contexts and circumstances. Criteria, subcriteria, indicators and weights could be adapted to meet the needs of other case studies.

6. Conclusions

This methodology is suggested as an effective tool to strengthen the decision-making process addressing a sustainable location that considers a regional system. In the face of location problems referring to a regional scale, it is fundamental to attend to the local necessities and understand the resonance and implications at a larger scale. In the case of locating a healthcare facility such as a regional hospital, a spatially efficient location is crucial because it is reflected by a set of multifactorial aspects such as local development, local dynamics, local characteristics and infrastructure that might influence regional dynamics as well. This study proposes a multi-criteria framework in two scales, observing the macro-location (hierarchy of possibilities) and the micro-location (hospital implementation) concerning the externalities.

Competition between municipalities triggered by the capture of the local investment has led to years of controversy in deciding the Portuguese case. The methodology offers a technical instrument to overcome political rivalry and efficiently allocate public money.

The location and/or relocation of specialised functions cannot ignore the factors that enable or enhance their operation, considering the interaction between service efficiency, other multiplier effects and positive externalities. The recognition of this interaction between cluster size and the level of equipment or service has long been acknowledged in regional science. It has passed into public policies for spatial planning and territorial organisation. This is evident in documents defining healthcare equipment networks, as well as in territorial models and guidelines for regional territorial management instruments, which not only determine the hierarchy of places but also establish the type of functions that should be located there, binding the entire administration.

This sustainable decision-making framework proposes that a multi-criteria analysis should be supported by a precise definition of the facility’s catchment area provided by a higher instance of power (such as the Ministry of Health, in the case of a healthcare facility). In the same way, the location of the new facility should be aligned with policies and territorial plans, following norms and guidelines related to structuring urban systems to ensure the project’s success.

This article argues that the decision regarding the location of new healthcare facilities should be supported by an integrated approach and an innovative structured multi-criteria evaluation process in four phases: (i) definition of initial assumptions, (ii) macro-location evaluation, (iii) micro-location evaluation and (iv) mitigation measures. The availability of methodologies capable of including comprehensive and integrated dimensions of analysis within their scope may have the potential to support complex planning decisions and amplify their multiplier effect with repercussions on the development of the territories where they are applied. As the study has demonstrated, the location of a regional facility might influence not only spatial conditions, but it stretches over socioeconomic, political and environmental consequences. The implementation of large public facilities is instrumental in the delivery of efficient policies. More objectively, this study shows that the location of a facility of regional or sub-regional relevance should not only respond strictly to its target audience but should also consider its effects in related areas, with an impact on local and regional development. The multiplier effects of public investment will thus be broader.

The presented case study concerns mostly spatial and health policies, but it was demonstrated that socioeconomic effects might follow, as well as environmental impacts. Considering that the methodology might be used for other types of facilities, it can impact other public policies like education, housing, etc. We argue that the design of policies should have a transversal and broad vision, considering that decision-making in one sector might affect others. The capacity of policy articulation should be part of the decision-making process.

These methodological steps and the aspects that integrate them were deemed the most appropriate for the case study.

It must be stressed that although this study was conducted with great methodological rigour, it does have some limitations, including the fact that it has not yet been empirically validated and the possibility that some public investments in equipment or infrastructure, due to their specific nature, are not suitable for the application of this methodological proposal.

Likewise, other territorial contexts may warrant a critical review of the methodology, particularly concerning the four phases described above or the indicators considered therein. This transparency and rationality could help make locating facilities more consensual or, at least, more comprehensive. Nevertheless, this study could benefit from an application of the methodology to a case study of a different type of facility or located in different regions, to understand the validity of criteria and indicators proposed and develop them further into a proper locational decision model.