A New Dimension of Health Sustainability Model after Pandemic Crisis Using Structural Equation Model

Abstract

Since the coronavirus (COVID-19) pandemic, it has been clear that the health dimension (HEDm) has a severe impact on sustainability, which was originally considered from the pillars of society, environment and economy. Hence, the integration of the health dimension into the other three pillars is plausible to define guidelines and criteria for progress monitoring and policy assessment towards a health-sustainable city. The objective of this study aims to present The Health Sustainability Model (HSM), a four-dimensional model for health sustainability (health, economy, environment, and society), using the Del-phi method to determine potential indicators agreed by eighteen experts, including physicians who deeply understand issues on health sustainability, and assess complex dimensions of health in the context of sustainability. The researchers have found that 45 indicators, later grouped into 15 elements and 4 dimensions, have a high level of agreement with Kendall’s W (KW) at 0.36. The HSM was then examined by the structural equation model (SEM) with reliability and validity shown as follows: the absolute fit with CMIN/DF = 1.44, RMSEA = 0.033, GFI = 0.96, AGFI = 0.94, RMR = 0.025, and the incremental fit with NFI = 0.94, CFI = 0.98, TLI = 0.97, and IFI = 0.98. Based on the results, the model is valid, in line with the empirical data. For further application, the HSM is expected to support city planners and decision makers by identifying room for improvement in each dimension through the indicators employed in the model. In contrast to existing studies that mainly use qualitative data, by conducting quantitative assessment, the model enables policy makers to objectively evaluate conditions and appropriately design policies to improve residents’ well-being.

1. Introduction

Sustainable development (STD) was first introduced in the Brundtland Report in 1987. The term later came to denote a concept of sustainable development, receiving much attention globally [1]. From literature reviews, “sustainable development” has been defined by many scholars; however, the definition of sustainable development remains ambiguous because of its distinctive characteristics, [1] varying depending on the perspectives, goals, and contexts of cities [2,3]. Nonetheless, a sustainable city encompasses the two following aspects: enhancing human well-being and using ICT infrastructure in city management. The concept of STD with reference to “a development that meets the needs of the current generation without compromising the ability of future generations to meet their own needs” [4] is commonly used in three dimensions of development, society, environment, and economy [5], by primarily focusing on environmental development to reduce the impacts on society and economic growth [1,6]. Existing studies show that the three dimensions of STD are interrelated [7] and that promotion of STD activities is needed to conduct STD comprehensively, particularly in the social dimension (SODm) that interrelates with health dimension (HEDm) [8,9]. For the purpose of this study, STD means “development toward goals in consideration of the context of a city while strictly emphasizing the participation of all stakeholders and efficiently managing the effects on the health, environment, economy and society dimensions to achieve a new normal”. The definition used in the manuscript is consistent with the Brundtland Commission, stating that health is an integral part of social sustainability [10], making ‘Health for All’ an ultimate global goal after the World Health Assembly announced in 1977 that STD is the main goal of global development. Published in The Publication of the Brundtland Report in 1987, STD was turned to consider health development on the basis of equality, differently from the former concept that primarily focused on future generations and the environmental dimension (ENDm). The concept of health equality and STD was consequently introduced since 1989, marking a significant step for policy planning at national and local levels [11]. The concept of health promotion integrated with sustainable development was introduced through three models: (1) “The Mandala of Health: a model of the human ecosystem” studying the human ecosystem and the intersection between human culture and ecological environment. The model presents a belief that health depends on culture and the biosphere, (2) “Human development: focusing on the sustainable health”, presenting sustainable development with HEDm, ENDm, and an economic dimension (ECDm), (3) “Health and community ecosystem”, presenting a concept of the integration of health with the SODm, ENDm and ECDm in the context of healthy community [12]. Taking into account that health services are not only a single factor contributing to good health, the model also relates to other components such as the economy, lifestyle, behavior, and environment that may affect human health [13,14,15]. Later in 2000, the United Nations (UN) proposed the eight Millennium Development Goals (MDGs) to develop quality of life (QoL), namely; (1) to eliminate extreme poverty and hunger; (2) to achieve universal primary education; (3) to promote gender equality and empower women; (4) to reduce child mortality; (5) to promote maternal health; (6) to fight malaria, HIV/AIDS and other diseases; (7) to promote environmental sustainability and (8) to develop a universal partnership for development. The proposal was the official beginning of the integration of health and sustainability globally [15]. However, the MDGs ended in 2015 when the UN established “the Sustainable Development Goals (SDGs)” as a new target to drive global STD until August 2030 (a period of 15 years). The current SDGs are developed on three dimensions of sustainability: ECDm, SODm, and ENDm, encompassing 17 goals [16]. Nevertheless, the occurrence of COVID-19 has drawn a reconsideration as to whether the current concept of SDGs is sufficient [17]. The pandemic shows that HEDm substantially affects all three dimensions. Particularly, SODm is directly affected by social distancing to reduce the spread of infections, which consequently affects ECDm. In fact, the lack of socialization, travelling discontinuity and supply chain disruption in all affect the economy globally. Furthermore, ENDm has been affected by the disposal of a large amount of infectious and hazardous waste during the epidemic [7]. Considering the fact that, during the pandemic, each country has had to inevitably manage its healthcare system and resources sufficiency to control the spread of the coronavirus [18], healthcare was in a critical state around the world. Therefore, the integration of HEDm into the three pillars of STD is a plausible step to determine guidelines and criteria for progress monitoring and policies assessment to prepare a health-sustainable city for possible future outbreaks [13].

Table 1. Literature reviews matrix.

Objective	Initial Factors	Sustainability Dimension Focus					Result	Ref.
		Soc.	Eco.	Env.	Hea.	Etc.
Investigate the possibility of comparing, in a transparent way, urban quality of life using sustainable development indicators	ISO37120	✓	✓	✓			IDS	[19]
Identify indicators for the sustainable development of cities that have the greatest potential for their underlying data to be measured by means of remote sensing.	ISO37120	✓				✓	IDS	[20]
Analyze the indicators of the São Paulo City Observatory (Observa Sampa), confronting them with the ISO 3712x series (sustainable, smart, and resilient cities) standards	ISO37120 ISO37122 ISO37123	✓	✓			✓	IDS	[22]
Applied indicators at municipal-level regulatory plans submitted to SEA in Chile.	ISO37120			✓			IDS	[21]
(1) The three ecological models are the links between health, environment and economy (2) All models link the social, environmental and economic dimensions of a healthy and sustainable community.	-	✓	✓	✓	✓		CM	[13]
Understanding the linkages between health promotion and sustainable development for health and food	-	✓	✓	✓			CM	[15]
Evaluates the sustainable development level with the phenomenon of a spatial clustering case study in China	LR	✓	✓	✓			EM	[23]
Evaluation of urban sustainability with grey relation analysis is used to reduce the uncertainty existing in the process of an evaluation case study in China	LR	✓	✓	✓			EM	[1]
Assess Sustainable Urban Development in an Emerging Economy with Fuzzy-TOPSIS in a case case study in Vietnam lacking clear data	LR	✓	✓	✓			EM	[24]
Assess a healthy and safe built environment with integrated MCDM methods in a case study in Lithuania	LR/Unitednation	✓	✓	✓			EM	[25]
Sustainability assessment method that integrates the MCDM approach with the variability of the alternatives’ performance measurement in a case study in Europe	LR	✓	✓	✓			EM	[26]
Verify the performance of Brazilian municipalities in the three dimensions of sustainability in a case study in Brazil	LR	✓	✓	✓			EM	[27]
Developed an evaluation index system that satisfies the requirements of green development in coal-resource-based cities by considering four dimensions in a case study in China	LR	✓	✓	✓		✓	EM	[28]
Evaluate the progress of a single city towards the concept of sustainable development in a case study in Poland	Warsaw policy	✓	✓	✓			EM	[29]
Present a new model, called Geo Umbria SUIT, integrating Multicriteria Analysis and Geographic Information Systems, specifically developed for helping Decision Makers to take policy decisions about sustainability in planning in a case study in Malta	the Geo Umbria Suite	✓	✓	✓			EM	[30]
The tested model, Geo Umbria SUIT, was found very suitable for territorial sustainability assessment, for evaluating sustainability at the territorial level in two different European countries, i.e., Italy and Spain.	National Statistics Office	✓	✓	✓			EM	[31]
Assessment of localizations in the Besancon area in terms of sustainable urban development; case study in France	GIS grid	✓	✓	✓			EM	[32]
This paper: presents a new general model of health sustainability that integrates the health dimension with the other three sustainability dimensions for a Thai context.	ISO37120 ISO37122 ISO37123 U4SSC	✓	✓	✓	✓		SEM SMM

Note: Literature review (LR); Social (Soc.); Economic (Eco.); Environment (Env.); Health (Hea.); indicator set (IDS); conceptual model (CM); Structural equation model (SEM); Sustainable measurement model (SMM); evaluation model with MCDM (EM).

shows the literature reviews of a sustainability model with health dimension. The study was firstly presented as “the three ecological models”, a descriptive model demonstrating interrelations among health, environment, economy and sustainable development [13,15]. As it primarily aimed to present a conceptual model, the indicators were not sufficiently considered, which makes the model difficult to use for further application and quantitative monitoring. Later, academics proposed a framework to rely on standard indicators as a Sustainable Model Measurement (SMM). The indicators were, namely, “ISO37120: Sustainable development Indicators” for city services and quality of life in communities. The ISO37210 relates to quality of urban life by using indicators for sustainable development [19], identifying indicators for city development [20] and applying the indicators into the plan of municipal level regulatory submitted to SEA in Chile. [21]. With the rise of the smart city concept, “ISO37122: Indicators for smart cities” and “ISO37123: Indicators for resilient cities” were introduced in model building as the standards in relation to the concept of smart city and urban renewal, mentioned in “Application of Open Government Data to Sustainable City Indicators: A Megacity Case Study [22]”. At the time, SMM was frequently used as a sustainability assessment model with multi-criteria discission making (MCDM) when evaluating the level of sustainable development [1,23,24,25,26,27,28,29,30,31,32]. Our literature reviews have also found that while the existing studies aim to present SMM with three sustainable dimensions, they have not described SMM with health sustainability. Furthermore, the indicators, though selected by a group of experts, lacked consensus on verification of validity and reliability. Hence, there is a possibility that the selection was conducted with biases such that a verification on compliance with empirical data to confirm the accuracy of the proposed SMM becomes necessary.

To fulfil this academic gap, this research presents a general model of health sustainability as an assessment tool that integrates the health dimension with the other three sustainability dimensions. The model is composed of standard indicators such as ISO37120, ISO37122, ISO37123 and U4SSC that were agreed through the Del-phi technique with experts from various fields. Then, the model is examined for its accuracy and generalized by the structural equation model. As the researchers also consider the context of Thailand, the model is designed to be functionable with quantitative indicators and to be able to use in MCDM, the model is expected to support strategy formulation for sustainable city development by identifying room for improvement on residents’ well-being in all related dimensions. For more details, the paper is structured into four sections with the aim to present a new model of sustainability. First, Section 1 introduces a literature review. Section 2 elaborates on methodologies of the study, starting from establishing indicators and expert panels and ending with the HSM structural equation analysis. Then, Section 3 shows the results of the study. Lastly, Section 4 conclude the study with a discussion on limitations and future research.

2. Materials and Methods

2.1. Delphi Method (DM)

DM is adopted to thoroughly investigate in-depth issues and generate a consensus from an expert panel. So, it does not need a large sample size [16], and mainly conducts investigation with the help of experts. Purposive sampling is used. The experts verify the structural validity, interpretation and variable classification [33], and generate a consensus towards the complex issues by cooperating using the questionnaires [34], and anonymity of experts [35] is enforced to avoid acquiescence of the experts from the beginning to the end of the study. The effectiveness of DM is contingent on selecting quality experts because this may alter the outcomes and accuracy [33,36]; it therefore takes into account the related information on education, experience, and competence of the experts importantly [37] with respect to the consistency with the respective topics studied [38], and diversified expertise and the in-depth data of the area studied should be inclusive to ensure data clarity in all dimensions [39]. In this study, a structured questionnaire was used to analyze the indicators derived from the literature review and standard indicators expected to have a relationship that may affect the 45 indicators. The procedures are as follows:

2.1.1. Establishing Indicators & Expert Panels

• The researchers gathered indicators from literature reviews and standards in relation to a sustainable city, such as ISO37120, ISO37122, ISO37123, and U4SSC [40,41,42,43], to select indicators (IDC) with consideration of the context of Thailand.
• While the number of experts for the DM should be contingent on 7-8 persons in theory [44], the researchers requested 18 experts to provide their opinions. The group of experts included physicians with in-depth understanding of overall health and sustainability issues at both the national and local level. For reference, The CVs of each expert are shown in

  Table A1. Experts profile & qualification.

  Number	Title	Experience	Expertise
  1	M.D.	Medical physician, professional level of government hospitals Medical lecturer, department of emergency medicine and disaster operations, faculty of medicine	Pulmonary medicine and pulmonary critical care, emergency medicine, community medicine, environmental medicine, internal medicine Specialty: Respiratory Disease
  2	M.D.	Medical physician, professional level and deputy director of government hospitals	Surgery, emergency medicine, pathology
  3	M.D.	Medical lecturer of department of public health nursing and urban medicine Deputy dean for administration and quality assurance faculty of nursing	Maternal Nursing Infants and Midwifery, community medicine, environmental medicine
  4	M.D.	Medical doctor, professional level of government hospitals	Internal medicine, rheumatology and rheumatism, community medicine
  5	M.D.	Professional level officer at strategic and planning division department of disease control	Emergency management, disease control planning, animal science
  6	Dr.	Public health academician, professional level at department of disease control, ministry of public health	Health promotion, community health
  7	Dr.	Public health academician, professional level at department of disease control, ministry of public health	Nutrition ology and dietetics, food science for health, sanitary
  8	Dr.	Professional level officer at fiscal policy office ministry of finance	Macro & micro economic policies analysis, international economy
  9	Dr.	Academician professional level at office of the national economic and social development council	Analyze and plan work to drive economic and social development.
  10	-	Professional level officer at Ministry of social development and human security	Social welfare system, human and social development policy.
  11	-	Professional level officer at department of public works & town country planning	City planning analysis and development.
  12	Dr.	Academician in city architecture Chairman Import and export company	Smart sustainable city, smart city, sustainability, economic, sustainable development.
  13	Dr.	Professional level officer at digital economy promotion agency	Smart city, smart city characteristics, smart city project analysis, economic
  14	Dr.	Environmental academician, master level at ministry of natural resources and environment	Carbon footprint, water quality, air quality, noise, toxic substances, environmental quality standards.
  15	Dr.	Environmental engineer at ministry of ministry of natural resources and environment	Plan and control environment, environment risk factor examine.
  16	Dr.	Lecturer professional level at Kasetsart university Bangkok Thailand	Sustainable development, environmental Science
  17	Dr.	Lecturer, professional level at Srinakharinwirot University Bangkok Thailand	Sustainable measurement, Sociology and Anthropology
  18	Dr.	Lecturer, professional level at Chiang Mai University	Social Science, Sustainable Development

  .

2.1.2. Delphi Procedures & Consensus Indicators

First, we contacted qualified experts to request support as a sample participating in the research study. Next, the questionnaires were distributed for the first time. This was the Item-Objective Congruence (IOC) closed-ended developed questionnaire [45], including 45 indicators validated by the experts for accuracy and appropriateness that will be used to set the HSM (−1 = inappropriate) (0 = uncertainly) (+1 = appropriate). Each indicator must be close to 1, and the minimum value must be ≥0.05 or higher, but if the minimum value is ≤0.05, item adjustment or elimination may be needed [45]. The closed-ended questionnaire with qualified indicators is distributed for the second time. The experts expressed their opinion on the indicators by 5-point Linkert rating scale [46] (from 5 = very important to 1 = least important), halting the study when the expert consensus was consistent. The following statistical tests were conducted, finding as follows: Median (MD) ≥ 4 “highly important” [47], Interquartile Range (IQR) ≤ 1 and Standard Deviation (SD) < 1 [48]. KW ≤ 0.5 was used to test the degree of consensus in answering questions [49].

2.2. Structural Equation Model (SEM)

Over the past decades, SEM has been very popular in all academic arenas, including science, engineering, medicine, marketing, and education. SEM is a statistical analysis tool simultaneously encompassing the multivariate analysis techniques, focusing on examining the modification of the theoretical model [50], and the main goal of using SEM is to test the hypothesis through a causal relationship line [51], combining factor analysis and path analysis. It includes two main components (1) measurement model (MM) and (2) structural model (SM) [52]. The process of analyzing structural equations typically starts with exploratory factor analysis (EFA) to examine the structural relationship between variables to reduce the number of elements through grouping of elements into dimensions. If the structural relationship between variables is known from the review of literature, theories, or hypotheses based on studies, confirmatory factor analysis (CFA) may not be necessary. CFA is often used to confirm the relation between a theoretical variable or a given assumption [53]. The result of CFA is called the measurement model; afterwards, it is used to analyze SM by Regression Analysis, called ‘SEM’. The analysis is conducted using statistical programs such as SPSS for EFA analysis, and AMOS for CFA and SEM

SEM Procedure & Consensus Criteria

• The research instrument was designed after the analysis with DM where the final variables were concluded by the consensus from experts in 4 dimensions and 15 elements. To obtain a rating from the experts, a closed-ended 5-point Likert scale questionnaire [46] was presented to check for consistency and validity. An indicator was deemed acceptable when the IOC value was 0.5 or higher [45]. Next, the researchers tested the reliability of the questionnaires (40 questionnaires) where Cronbach’s Alpha must be 0.7 or higher [54]. Finally, the questionnaires were distributed to the samples, for which the number was determined with reference to Taro Yamane’s Sample Size Table [55]. The distribution of the questionnaire was conducted using the Snowball Sampling method, where the first group of samples or the experts refer to and nominate similar individuals who were qualified and suitable to complete the questionnaire.
• Factor Analysis (FA) is a process where the data obtained from the questionnaires are put into testing for the suitability with a statistical program. When the Kaiser–Meyer–Olkin (KMO) value ≥ 0.7 [56], the data would be suitable for EFA to determine the structural relationship between elements. To reduce the number of elements, the researchers have grouped them into dimensions, and CFA analysis was conducted to confirm theoretical variable relations and check the hypothetical model’s fit with the empirical data using two groups of statistics, namely 1) Absolute fit index, including CMIN/DF, RMSEA, GFI, AGFI, and RMR, and 2) Incremental fit index, including NFI, TLI, CFI, IFI. Finally, SM analysis was performed to test the research hypothesis.
• Model validity and reliability: After completing the factor analysis, the researchers had verified the reliability convergent validity of the model with Composite Reliability (C.R.) to assess the accuracy of elements and latent variables. The preferable C.R. is above 0.7 and more than the value of AVE, where $L_i$ is standardized factor loading and $n$ is the number of items. The C.R., calculated from the squared sum of factor loading $L_i$, is often applied in SEM models. The C.R. was then used to construct the sum of the error variance, shown as $e_i$, which can be calculated from (1). The Average Variance Extracted (AVE), the average variance of the extracted variables, was calculated per formula (2). While the AVE should be more than 0.5, the Cronbach’s Alpha, which measures the internal consistency of the variables calculated by the SPSS program, should be more than 0.7. Furthermore, maximum shared variance (MSV) was tested on the relations between latent variables. MSV is the square of the highest correlation coefficient between latent constructs, where the acceptable MSV should be less than the value of AVE [56]. The above procedure is shown as a conceptual framework in Figure 1.

  $$\mathrm{C}.\mathrm{R}.=\frac{{\left({\displaystyle \sum _{i=1}^n{L}_i}\right)}^2}{{\left({\displaystyle \sum _{i=1}^n{L}_i}\right)}^2+\left({\displaystyle \sum _{i=1}^n{e}_i}\right)}$$

  (1)

  $$\mathrm{AVE}=\frac{1}{n}{\displaystyle \sum _{i=1}^n{L}_i^2}$$

  (2)

3. Results

3.1. Del-Phi Reliability

As shown in

Table 2. Del-phi Results.

Dimension	Element	Indicators	Unit	IQR	M.D.	AVR.	S.D.	CODE
Health	Health Status (HS)	Life expectancy	Year	0	5	5.00	0.00	VR01
		Low-birth-weight newborns	%	1	4	3.56	0.50	VR02
		Number of deaths (rate)	1 K	1	4	4.00	0.67	VR03
		Infant mortality livebirth (rate)	1 K	1	4	3.56	0.50	VR04
		Suicide mortality (rate)	100 K	1	4	4.22	0.63	VR05
	Communicable Disease Control (CDC)	HIV/AIDS mortality (rates)	100 K	1	5	4.56	0.68	VR06
		Tuberculosis mortality (rates)	100 K	1	4	3.78	0.63	VR07
		Pneumonia mortality (rates)	100 K	1	4	3.44	0.68	VR08
		Diarrhea mortality (rates)	100 K	1	4	3.56	0.50	VR09
	Non communicable Disease Control (NCDs)	Cancer mortality (rates)	100 K	0	5	4.89	0.31	VR10
		Stroke mortality (rate)	100 K	1	4	3.67	0.67	VR11
		Ischemic heart disease Mortality (rate)	100 K	1	4	4.33	0.67	VR12
		Diabetes mellitus mortality (rate)	100 K	1	5	4.56	0.50	VR13
		Chronic obstructive pulmonary disease mortality (rate)	100 K	1	4	4.22	0.63	VR14
		Chronic kidney disease mortality (rate)	100 K	1	4	3.44	0.68	VR15
	Health Resource (HS)	Physicians (rate)	100 K	1	5	4.56	0.68	VR16
		Hospital beds (rate)	100 K	1	4	4.33	0.67	VR17
		Nursing & midwifery personnel (rate)	100 K	1	4	4.33	0.67	VR18
		Psychiatric physicians (rate)	100 K	1	5	4.44	0.68	VR19
		Total Health worker (rate)	100 K	1	4	3.56	0.50	VR20
		Ambulance (rate)	100 K	1	4	4.00	0.67	VR21
		Electronic Medical Records	%	1	4	4.44	0.50	VR22
Environment	Environment Risk Management (ERM)	Environmental risk Management	%	1	4	4.33	0.47	VR23
	Air pollution management (APM)	Average of AQI Index	Index	1	4	3.44	0.68	VR24
	Protected Natural Areas (PNA)	Forest area rate	%	1	5	4.56	0.50	VR25
	Water Management (WM)	Water management Index	Index	1	5	4.56	0.50	VR26
Social	Health Service Standard (HSS)	Health resource management	%	1	5	4.56	0.50	VR27
		Transparency in public health	%	1	4	4.33	0.67	VR28
		Green & Clean hospital administration	%	1	4	3.56	0.50	VR29
		Management of public health crises	%	1	4	3.44	0.68	VR30
		Community hospital quality	%	1	4	3.56	0.50	VR31
		Control of acute infectious diseases	%	1	4	4.00	0.67	VR32
	Social security (SS)	Smoking mortality (rate)	100 K	1	4	3.56	0.50	VR33
		Alcohol drinking mortality (rate)	100 K	1	4	4.00	0.67	VR34
		Traffic accident mortality (rate)	100 K	1	5	4.56	0.50	VR35
		Crime mortality (rate)	%	1	5	4.44	0.68	VR36
	Health Promotion(HP)	Universal Health Coverage Service	%	1	4	4.67	0.68	VR37
		Desirable health behaviors	%	1	4	4.33	0.67	VR38
		Obesity (BMI > 30.0 kg/m2)	%	1	4	3.67	0.67	VR39
		Management of Glycemic control	%	1	4	4.00	0.67	VR40
		Management of Blood pressure control	%	1	4	3.67	0.67	VR41
Economic	City’s Employment (CEM)	Unemployment rate	%	1	5	4.44	0.68	VR42
	Poverty Reduction (PR)	Population living in poverty	%	0	4	4.11	0.31	VR43
	Household debt (HHD)	Household debt per income ratio	%	1	4	4.33	0.67	VR44
	Economic growth(ECG)	Gross Provincial Product growth rate	%	1	5	4.44	0.68	VR45
		Summary		0.93	4.30	-	0.58	45
		KW = 0.36/Sig 0.00/N 18

Note: 1 K = 1000 Population; the Interquartile Range (IQR); median (M.D.); average (AVR.); standard division (S.D.); Kendall’s Coefficient of Concordance (KW); Level of significance; number of expert (N).

, it was found that the 18 experts had a consensus on all 45 indicators overall, IQR = 0.93, S.D. = 0.58, and M.D. = 4.3, and on each variable. Based on the results in Table 2, all indicators have IQR ≤ 1, M.D. ≥ 4, and S.D. < 1 where, in the DM study, all variables should have IQR ≤ 1 and S.D. < 1. Hence, it can be interpreted that the experts had a consensus on consistency, with all indicators of high significance [45]. When KW = 0.30 < 0.50, the answers were unanimous [46]. The results of the DM analysis can be summarized in Table 2.

3.2. The Health Sustainability Model (HSM)

The HSM encompasses four dimensions and 15 elements as follows:

3.2.1. Health Dimension (HEDm)

Table 2 reflects the basis of good health. Being perfect physically & mentally healthy since birth, including main disease control and the urban health service resources that result in urban public service from birth to the end of life, health status, and well-being are the heart of to human happiness and contribute greatly to economic and social development, consisting of four elements as follows:

• Health Status (HS) is one’s health status from birth to the end of life that affects longevity and health; it is influenced by several factors, including genetics, behavior, environment, health service system. and social factors, consisting of five indicators; VR01–VR05. Within HS, life expectancy has the highest average of five, implying that life expectancy is absolutely agreed to be an indicator for HS.
• Communicable Disease Control (CDC) reflects health promotion by protecting the general public from the effects of major communicable diseases. The goal is to monitor and control the cause of germs’ spread to minimize new infected cases and risk of recurrent and severe disease, consisting of four indicators; VR06–VR09. Table 2 shows that, within CDC, HIV/AIDS mortality has the highest average of 4.56, implying that is largely considered as an indicator for CDC.
• Non-Communicable Disease Control (NCDs) reflects the broad exposure to risk factors causing chronic diseases which are a major public health problem around the world because they cause sickness, disability, and premature death. The monitoring of chronic disease situations can reflect the health risk factors that have been emerging in the cities in a certain way, for instance, the tendency in urban populations to die from lung cancer mostly, etc.; it consists of six indicators, VR10–VR15. On the indicator of NCDs, Cancer mortality is widely agreed to be an indicator, with an average of 4.89.
• Health Resource: (HR) reflects efficiency and fundamental resources in the delivery of health services to the public and society; for instance, medical personnel on duty either directly or indirectly contact with patients, ward beds, ambulances, etc. The standard health service should possess resources in an adequate quantity to meet the demand, which affects the quality of public health services in the wider society. It consists of seven indicators, VR16–VR22. Among the seven indicators, the rate of physicians is largely agreed to be an indicator with an average of 4.56.

3.2.2. Environment Dimension (ENDm)

ENDm reflects the natural resources situation and urban environment, focusing on the environmental management of specific problems such as water pollution, air pollution, and environmental risk factors that are significant to human life. Urban communities affect well-being as well as the health and integrity of the ecosystem. The dimension consists of four elements as follows:

• Environment Risk Management (ERM) reflects the physical management factors, including the use of chemicals, biological and work that impact health, pollution, radiation, noise, land use patterns, and a working environment that impacts climate change, except for natural disasters beyond human control. These risks are attributed to non-health sectors such as energy, industry, agriculture, and transportation, which affect the incidence of chronic diseases (NCDs), consisting of one indicator, VR23.
• Air Pollution (APM) reflects air contamination by chemicals, dust, fumes, and pollution. It is a major public health concern, including particulate matter (PM), carbon monoxide (CO), ozone (O₃), and nitrogen dioxide (NO₂). Sulfur dioxide (SO₂) affects morbidity and mortality from respiratory diseases and subsequent chronic diseases. It consists of one indicator, VR24.
• Protected Natural Areas (PNA) reflect protected natural areas, including plants and wildlife that are diverse in ecosystems on land, underwater and in wetlands; such areas can slow the loss of biodiversity because the areas are not disturbed by humans. It consists of one indicator, namely VR25.
• Water Management (WM) reflects eight dimensions of sustainable water management, namely water management costs, water management for consumption and usage, water security for development, water consumption cost, water quality and environment management, water disaster management, watershed forest conservation, and water resources management, consisting of one indicator, VR26.

3.2.3. Social Dimension (SODm)

SODm reflects the quality-of-life status, enhancing safety and health promotion in society, consisting of 3 elements as follows:

• Health Service Standard (HSS) reflects the standards established to implement the health service system and to measure the success of quality management as per standards, including building and environmental standards, medical equipment, and behavioral health education implementation to the public to deliver quality services that are trusted by the public and society, consisting of six indicators, VR27–VR32. Within the HSS element, health resource management has the highest average of 4.56, implying that the experts consider health resource management as an important indicator.
• Social Security (SS) reflects the quality of social security affected by events that occur in society as caused by human behavior, thereby consequently affecting the safety of life and property, injury and death; for instance, crime problems, substance use, accidents, or incidents that may occur if a person in society behaves recklessly, affecting physical health, mental health, quality of life and public peace. It consists of five indicators; VR33–VR36. Among the four indicators, traffic accident mortality receives the highest average of 4.56, which means the experts significantly consider traffic accident mortality as an indicator.
• Health Promotion (HP) reflects the control performance promotion and basic personal health improvement in attaining ultimate well-being. Health promotion is a process that encourages people to attain self-improvement on health and self-care, such as via exercise, etc., by reducing the behavioral risk factors that may cause diseases. It consists of five indicators; VR37–VR41. Among the five indicators, universal health coverage service achieves the highest average of 4.67, implying that universal health coverage service is a significant indicator.

3.2.4. Economic Dimension (ECDm)

ECDm reflects the income distribution and economic growth; economics is an element in meeting human basic needs and well-being, and ECDm reflects the improved quality of life in terms of public wealth and healthy well-being. It consists of four elements as follows:

• City’s Employment (CEM) reflects economic power; weak economy, sluggish economy, people’s lack of income, poverty, and discontinuity of private expenditure may affect quality of life, and poverty problems usually resulting from the low employment rate, consisting of one indicator, VR42.
• Poverty Reduction (PR) reflects economic policy and humanity, aiming at eradicating poverty as it is a harmful obstacle to social development, physical health, and mental health, as well as shortening the life expectancy of people in society. Most of them are remedied by economic empowerment and humanitarian policy. Poverty alleviation can improve the quality of life. It consists of one indicator, VR43.
• Household Debt reflects the household debt burden to pay interest or principal sum to the creditors on the due date. The incremental household debt affects private consumption and suspends economic growth in the long term. Liabilities and debts can also affect the happiness of people in society, particularly mental health, which may influence and press an individual’s attitudes to the extent of suicide decisions eventually. It consists of one indicator, VR44.
• Economic Growth reflects economic expansion including land, labor, capital, and entrepreneurs. Economic growth has resulted from the people in society having been employed by the business sector, while the government sector receives more income tax to spend on public services in society. Economic growth can enhance the standard of living; that is, when the public’s income is higher, the greater spending power in various areas results, for example, health care, education, consumption, etc. It consists of one indicator, VR45.

3.3. Reliability and Sampling Adequacy Results

The number of respondents for the questionnaire is 442, with 54% having a graduate degree, 27% having an undergraduate degree and the other 19% having a doctoral degree. From the affiliation perspective, 58% work in the public sector, 34% work in the private sector, 6% work in state enterprise sand the last 2% work in other sectors. Looking at the roles, 36% of the correspondents work in health issues, 35% work in environmental issues, 20% in economic issues and 6% in social issues. When considering work experience, 49% of the correspondents had work experience of 10 years or more while the other 51% had work experience of less than 10 years. Based on residency, 23% of the correspondents live in the northeastern area, 22% live in Bangkok, 22% live in the central area and 12% live in the southern area. Prior to conducting FA, a reliability analysis was performed for the suitability of the data sets. In the SEM analysis, it was found that Cronbach’s Alpha was 0.87 > 0.70, indicating that the instrument and variables used in the research meet the criteria, and KMO and Bartlett’s Test (BT) showed the KMO was 0.88 > 0.50 and BT was 0.00 < 0.05, indicating that all variables were marvelous, with high suitability for the factor analysis [56] After the confidence and suitability test, the EFA process followed.

3.4. Exploratory Factor Analysis (EFA)

In EFA, 15 elements derived by calculation were used. The indicators in the groups shown in Table 2 were new variables, including VR01–VR05 to HS, VR06–VR09 to CDC, VR10–VR15 to NCDs, VR16–VR22 to HR, VR27–VR32 to HSS, VR33–VR36 to SS, VR37–VR41 to HP. As for the elements containing one indicator, the former name was used, including VR23–VR26 to ERM, APM, PNA, WM, VR42–VR45 to CEM, PR, HHD, ECG respectively. Next, element extraction was performed using the principal component analysis method. The number of factors was determined based on eigenvalue > 1 and factor rotation was performed using the Varimax method. The results showed that the commonalities value and factor loading of all variables were > 0.50, indicating that all variables were suitable for factor analysis and consistent with the experts’ variables grouping in Table 2. In the EFA analysis, all 15 elements were classified into four dimensions: HS, CDC, NCDs, HR to HEDm, ERM, APM, PNA, WM to ENDm, HSS, SS, HP to SODm, and CEM, PR, HHD, ECG to ECDm. After that, the elements were examined through the CFA, as shown in

Table 3. CFA Factor Loading Statistics.

The Standardized Factor Loading ( $L_i$)					Hypothesis Testing
Element	ECDm	HEDm	ENDm	SODm	Estimate	S.E.	C.r.	p
HHD	0.73				1.17	0.11	9.64	***
ECG	0.80				0.96	0.11	9.46	***
CEM	0.77				0.96	0.10	9.46	***
PR	0.74				1.09	0.12	9.43	***
CDC		0.75			0.95	0.09	10.02	***
HR		0.78			1.00	0.10	10.02	***
HS		0.67			1.03	0.09	8.92	***
NCDs		0.76			0.97	0.10	9.47	***
APM			0.67		1.02	0.08	12.58	***
ERM			0.68		0.98	0.08	12.58	***
WM			0.81		1.19	0.10	12.17	***
PNA			0.76		1.19	0.10	11.94	***
SS				0.69	0.86	0.10	8.62	***
HP				0.85	0.96	0.13	8.62	***
HSS				0.73	0.88	0.11	9.03	***

Notes: Standard Error (S.E.); Critical Ratio (C.r.); Unstandardized. p < 0.001 for all coefficients (***).

.

3.5. Confirmatory Factor Analysis (CFA)

The purpose of the EFA was to examine the structural relationship between variables in each group to reduce the number of variables and check the suitability of the variables in the structural model analysis; however, it could not test the hypothesis based on the structural relationship between variables. Thus, CFA was required to assert the elements’ relationships or hypotheses to see whether they met the criteria defined or not. The EFA analysis data were transferred from the SPSS program to the AMOS program. The analysis results are shown in Table 3. It was found that all elements of the measurement model in CFA were statistically significant. (***) corresponding to the EFA grouping, the critical ratio (C.r.) was ≥1, greater than the expected minimum value, indicating that the indicators are suitable to represent as dimension. In contrast, C.r. was <1, indicating that the variables were not suitable to represent the dimension [56]. Next, the model suitability testing is performed with test statistical values of two groups as shown in

Table 4. Measurement of model fit.

Index	Recommended Value	Estimated Value	Type	Ref.
CMIN/DF	$\le$3 good $\le$5 permissible	1.44 good	Absolute fit	[56]
RMSEA	$\le$0.05 good $\ge 0.05$–0.08 moderate $\ge$0.8 bad	0.033 good	Absolute fit	[58]
GFI	$\ge 0.95$ good $\ge 0.90$ acceptable	0.96 good	Absolute fit	[56]
AGFI	$\ge 0.90$ good	0.94 good	Absolute fit	[58]
RMR	Close to 0	0.025	Absolute fit	[56]
NFI	$\ge 0.95$ good $\ge 0.90$ acceptable	0.94 acceptable	Incremental fit	[56]
CFI		0.98 good	Incremental fit	[56]
TLI	$\ge 0.95$ good $\ge 0.80$ acceptable	0.97 good	Incremental fit	[56]
IFI	$\ge 0.90$ good	0.98 good	Incremental fit	[57]
P	$\ge 0.05$ significant	0.00	-	[57]
CMIN	0	105.3	-	[57]
DF	-	73	-	-

; it was found that Group 1 had Absolute fit, consisting of test statistics, CMIN/DF = 1.44, indicating that the model fit with the empirical data as whole, and RMSEA = 0.033, GFI = 0.96, AGFI = 0.94, RMR = 0.025, all test statistic values for Group 1 were of a good level. Group 2—Incremental fit, consisting of test statistics, NFI = 0.94, CFI = 0.98, TLI = 0.97, and IFI = 0.98, all statistical values for Group 2 were of a good level, except for NFI, which was of an acceptable level. To conclude, the model was a good fit with the empirical data; although the Chi-square p was 0.00 not significant, it can be explained that because Chi-square depends on the sample size, if the sample size is large, the Chi-square value is greater accordingly. In this study, the sample size was 400, with >200 being considered a large sample. Testing Chi-square may also conclude that the model is not a good fit with the empirical data. Therefore, a correction follows Bollen’s [57] recommendations; CMIN/DF < 3 was considered instead of the Chi-square value in testing the model fit with the empirical data.

Validity and Reliability in the CFA Analysis

Table 5. The HSM validity and reliability.

Dimensions/Element			$L_i$	$L_i^2$	$e_i$	C.R.	AVE.	C.A.
The HSM	--->	ECDm	0.77	0.59	0.41	0.85 *	0.59 *	0.87 *
	--->	HEDm	0.84	0.71	0.29
	--->	ENDm	0.76	0.58	0.42
	--->	SODm	0.70	0.49	0.51
ECDm	--->	HHD	0.73	0.53	0.47	0.85 *	0.58 *	0.78 *
	--->	ECG	0.80	0.64	0.36
	--->	CEM	0.77	0.59	0.41
	--->	PR	0.74	0.55	0.45
HEDm	--->	CDC	0.75	0.56	0.44	0.83 *	0.55 *	0.77 *
	--->	HR	0.78	0.61	0.39
	--->	HS	0.67	0.45	0.55
	--->	NCDs	0.76	0.58	0.42
ENDm	--->	APM	0.67	0.45	0.55	0.82 *	0.54 *	0.83 *
	--->	ERM	0.68	0.46	0.54
	--->	WM	0.81	0.66	0.34
	--->	PNA	0.76	0.58	0.42
SODm	--->	SS	0.69	0.48	0.52	0.80 *	0.58 *	0.79 *
	--->	HP	0.85	0.72	0.28
	--->	HSS	0.73	0.53	0.47

Note: the standardized factor loading ($L_i$); variance ($L_i^2$); the error variance 1 −$L_i^2$ ($e_i$); Composite (construct) Reliability (C.R.); Average Variance Extracted (AVE.); Cronbach’s Alpha (C.A.); Acceptable (*).

shows that the validity and reliability for both the model toward the four pillars and the four pillars towards the elements are of a satisfactory level, with the AVE less than the C.R., where C.R. must be more than 0.7 to be considered acceptable. As the AVE is more than 0.5 and the C.A. is more than 0.7, it can be concluded that the proposed HSM is valid and reliable. Then, the researchers conducted a study on the model structure by considering $L_i$ and reported in the order of maximum to miminum. Among the four dimensions in the HSM, HEDm has the largest $L_i$ with a value of 0.84. Looking into the HEDm dimension, it is found that the $L_i$ of HR, NCDs, CDC and HS are 0.78, 0.76, 0.75 and 0.67. Next, the dimension with the second highest $L_i$ is ECDm with a value of 0.77. Within the dimension, the $L_i$ of ECG, CEM, PR and HHD are 0.80, 0.77, 0.74 and 0.73, respectively. Then, the third highest $L_i$ is ENDm, with a value of 0.76. Within the ENDm, the $L_i$ of WM, PNA, ERM and APM are 0.81, 0.76, 0.68 and 0.67 accordingly. Lastly, in the social dimension, the $L_i$ of SODm is 0.70 with the $L_i$ of HP, HSS and SS are 0.85, 0.73 and 0.69. All indicators of each element are shown in Table 2; the structural model is illustrated as Figure 2.

SM presents a structural of the HSM’s measurement model as required to test the hypothesis for the full structural model, as shown in Figure 2, consisting of an assumption as follows:

• H1: The HSM structural model is consistent with empirical data (Accepted). Because this is a hypothetical testing of the structural model, various types of testing were adopted in the study. The statistics in Table 4 were used to analyze the data. Thus, H1 is accepted. The hypothesis testing can be summarized as a structural equation as shown in Figure 2 as follows:

$$\mathrm{The} \mathrm{HSM} =0.84{\mathrm{HEDm}}_{\mathrm{He}}+0.77{\mathrm{ECDm}}_{\mathrm{Ec}}+0.78{\mathrm{ENDm}}_{\mathrm{En}}+0.70{\mathrm{SODm}}_{\mathrm{Sc}}$$

(3)

$$\begin{array}{l}{\mathrm{HEDm}}_{\mathrm{He}}=0.76{\mathrm{v}}_{\mathrm{he}1}+0.67{\mathrm{v}}_{\mathrm{he}2}+0.78{\mathrm{v}}_{\mathrm{he}3}+0.75{\mathrm{v}}_{\mathrm{he}4}+{\mathsf{\epsilon }}_{\mathrm{he}}\\ {\mathsf{\epsilon }}_{\mathrm{he}}={ \mathrm{e}}_{\mathrm{he}1}+{\mathrm{e}}_{\mathrm{he}2}+{\mathrm{e}}_{\mathrm{he}3}+{\mathrm{e}}_{\mathrm{he}4}\end{array}$$

(4)

$$\begin{array}{l}{\mathrm{ECDm}}_{\mathrm{Ec}}=0.77{\mathrm{v}}_{\mathrm{ec}1}+0.80{\mathrm{v}}_{\mathrm{ec}2}+0.73{\mathrm{v}}_{\mathrm{ec}3}+0.74{\mathrm{v}}_{\mathrm{ec}4}+{\mathsf{\epsilon }}_{\mathrm{ec}}\\ {\mathsf{\epsilon }}_{\mathrm{ec}}={ \mathrm{e}}_{\mathrm{ec}1}+{\mathrm{e}}_{\mathrm{ec}2}+{\mathrm{e}}_{\mathrm{ec}3}+{\mathrm{e}}_{\mathrm{ec}4}\end{array}$$

(5)

$$\begin{array}{l}{\mathrm{ENDm}}_{\mathrm{En}}=0.67{\mathrm{v}}_{\mathrm{en}1}+0.68{\mathrm{v}}_{\mathrm{en}2}+0.81{\mathrm{v}}_{\mathrm{en}3}+0.76{\mathrm{v}}_{\mathrm{en}4}+{\mathsf{\epsilon }}_{\mathrm{en}}\\ {\mathsf{\epsilon }}_{\mathrm{en}}={\mathrm{e}}_{\mathrm{en}1}+{\mathrm{e}}_{\mathrm{en}2}+{\mathrm{e}}_{\mathrm{en}3}+{\mathrm{e}}_{\mathrm{en}4} \end{array}$$

(6)

$$\begin{array}{l}{\mathrm{SODm}}_{\mathrm{So}}=0.73{\mathrm{v}}_{\mathrm{so}1}+0.69{\mathrm{v}}_{\mathrm{so}2}+0.85{\mathrm{v}}_{\mathrm{so}3}+{\mathsf{\epsilon }}_{\mathrm{so}}\\ {\mathsf{\epsilon }}_{\mathrm{so}}={\mathrm{e}}_{\mathrm{so}1}+{\mathrm{e}}_{\mathrm{so}2}+{\mathrm{e}}_{\mathrm{so}3}\end{array}$$

(7)

Table 6. The validity test of the interrelated dimensions.

Relation between Dimensions			Cor.	MSV	Cov.	S.E.	C.R.	p
HEDm	<-->	SODm	0.71	0.50	0.29	0.03	8.88	***
ECDm	<-->	SODm	0.51	0.26	0.21	0.03	6.91	***
ENDm	<-->	SODm	0.53	0.28	0.28	0.04	7.12	***
ECDm	<-->	ENDm	0.67	0.45	0.38	0.04	9.17	***
HEDm	<-->	ENDm	0.61	0.37	0.36	0.04	8.58	***
ECDm	<-->	HEDm	0.68	0.46	0.3	0.03	8.86	***

Notes: Correlations (Cor.); Covariances (Cov.); Maximum Shared Variance (MSV); Standard error (S.E); Critical ratio (C.R.); Unstandardized. p < 0.001 for all coefficients significant (***).

shows that the P value of all tests are ***, interpreted as all dimensions being interrelated. Hence, all assumptions are accepted. Moreover, since MSV is less than AVE from Table 5 for all relations, we can conclude that the model is valid and reliable. The interrelations among the four dimensions and its illustration are shown in Figure 3.

Figure 3 presents interrelation among the four dimensions after testing an assumption on interrelation among each dimension. To verify the hypotheses, the researchers have set assumptions as follows.

• H:2 health dimension interrelates with social dimension (accepted)
• H:3 economic dimension interrelates with social dimension (accepted)
• H:4 environmental dimension interrelates with social dimension (accepted)
• H:5 economic dimension interrelates with environmental dimension (accepted)
• H:6 health dimension interrelates with environmental dimension (accepted)
• H:7 economic dimension interrelates with health dimension (accepted)

The causal relations can be depicted as below.

The HSM as shown in Figure 4 represents a four-dimensional sustainability model with each relationship overlapping and affecting human health directly and indirectly, including the dimensions of health, economy, environment, and society. Currently, sustainable urban development focuses on three main dimensions and often views health as small issues within SODm, while health is defined by medical services only. The concept of The HSM is that sustainable health determinants do not depend on SODm only. Thus, HEDm is separated from SODm. Being healthy is not just determined by health services, but all overlapping dimensions directly and indirectly affect health. Therefore, integration of the dimensions is needed to achieve overall health sustainability.

Health and well-being are linked to two implications. Firstly, the implications on health, including (1) physical health is the existence of a healthy body and being free from physical ailments, (2) mental health is the mental capacity, good mental health, and emotional stability to tackle difficulties and failure in life without contributions to stress, anxiety or depression causing deteriorating physical health. Thus, it can be concluded that a good physical condition and good mental health are the primary basis for normal happy lives in a society, and consequently, the feeling of social safety. Secondly, the implications on well-being, including (1) objective well-being—the basic physiological needs that human beings must have met, including food, health, shelter and safety and security needs from dangers, career stability, adequate emergency savings, etc., leading to the wider economic development; (2) subjective well-being—emotional need and responses, including belonging being loved by neighboring people and society (for example, having a family, being accepted and recognized by the society), being respected (esteem needs), including rank, position, competency; all these have affected the quality of life of the people.

Physical and mental well-being makes it possible for individuals to live their life to the fullest, consequently resulting in improved performance, a better quality of life, and improved economic stability when people in society are satisfied with their welfare. Community health development comprises three components: (1) health promotion, (2) health protection, including vaccinations, mental support counseling for quitting drugs, disease screening, etc., (3) health services, including nutrition counseling, treatment, and disease prevention information to the extent that quality is met adequately by the government to increase control and self-improvement on health, generally provided by the government from public health service investment to public health quality standards to support urban expansion appropriately as well as the equality and equity of access. Sustainability development in HEDm and ECDm at the same time is called “human development”.

Livable Development constitutes improving and adjusting the ecosystem into equilibrium, making changes such as a smaller number of patients and improved delivery of comprehensive public health services, clean living environment, such as air quality, water quality, and safety conditions, ability to coexist with nature, awareness and respect for all in the food chain, encouraging sharing the city, and developing a variable transportation option to support walking and cycling to reduce pollution, accidents and traffic congestion, and encourage urban people to share social activities, thus reducing the impact on the ecosystem as a whole (ecological footprint reduction).

Environmental wellness represents the state of physical well-being as a result of a pleasant environment supporting well-being and encourage interaction with nature. Sustainability can be achieved with the cooperation of the economic sector, which is a sector that directly impacts environmental health. The economic sector must be aware (environment awareness) of the environmental vulnerability and the importance of environmental protection, and implement environment policy that is designed to prevent and reduce the harmful impacts (carbon footprint reduction) generated by human activities on the environment; for example, educating the community on environmental protection, the control of chemicals, noise, pollution, waste, water resources, as well as energy-saving measures and reuse, etc. The HSM can be summarized as a diagram as shown in Figure 4.

4. Discussion

The research presents a model considering health sustainability. While precedent research does not consider the health dimension, our proposed model integrates the health dimension into the existing three sustainability dimensions. Based on the agreement from the experts and the factor analysis, the model contains 45 indicators, grouped into 15 elements and four dimensions, respectively. Model fit is tested by SEM and it was found that both the absolute fit index and the incremental fit index achieve a favorable result, consistent with the empirical data. Afterwards, the researchers conducted validity and reliability tests and found that the elements in all dimensions satisfied the preset criteria, consistent with the agreement from the experts. In the final step, the researchers tested the assumptions on interrelation among four dimensions and found that all dimensions have a positive interrelation. Hence, a change occurring in one dimension will cause a direct variation in other dimensions.

At present, the sustainability assessment model in Thailand is built only to measure at the national level. There has been no development of a concrete model to measure at the provincial level, leaving a space for the proposed model to implement. In the application, the model is crucial as it enables a city planner to formulate a strategy driving toward health sustainability and improving residents’ well-being in all dimensions. The HSM is designed to be a general assessment model, suitable for the context of Thailand to apply quantitative indicators and support the assessment with MCDM, and is expected to facilitate the assessment not only in any specific province, but in all 77 provinces of Thailand. For more details on the applicability of the HSM, the readers can refer to our research in Part 2: The Dynamic Evaluation Model of Health Sustainability under MCDM Benchmarking Health Indicator Standards [59].

Limitations and Future Research

• The research aims to propose a model to assess sustainability from four dimensions. While the integrity among the dimensions has been tested with the SEM, the relationships within elements and dimensions have yet been verified, leaving room for future research to conduct path analysis to test both direct and indirect effects. In future research, multiple regression analysis and descriptive statistics can be elaborated to provide in-depth understanding for each dimension and element.
• The research only presents an assessment model without any application with real data. Future research can conduct a sustainability assessment of a city by applying the model with MCDM.
• The sample mentioned in the study is gathered via a snowball sampling method. As a result, there is limitation in controlling the demographics of the sample. Future research should identify the demographics and expected composition of the sample to maintain its integrity.