Evaluating Social Resilience in Super-Aged Urbanism: A Cultural Dimension-Based Framework for Cluster Living Service Models

Abstract

As global urban centers transition into “Super-Aged Societies,” the paradigm of urban sustainability has shifted from mere housing provision to the development of Sustainable Care Retirement Communities (SCRCs). This study addresses a critical gap in the urban aging literature: the lack of culturally sensitive frameworks for social resilience in non-Western contexts. By integrating Hofstede’s Cultural Dimensions Theory, this research investigates how national culture influences the prioritization of community attributes within the “15 min city” framework. Methodologically, a hierarchical evaluation framework comprising 4 dimensions and 26 indicators was established. It employed the Fuzzy Delphi Method (FDM) to achieve expert consensus among stakeholders in Taiwan’s Long-term Care 3.0 ecosystem. Analysis using Double Triangular Fuzzy Numbers identified the “Charging Model,” “Staff-to-Resident Ratio,” and “Zoning with Care Continuity” as the highest-priority factors ($G^i$ ≥ 7.8). These results indicate that in cultures with high uncertainty avoidance, institutional financial stability and human-centric staffing are perceived as the structural bedrock of social resilience. Furthermore, the study highlights the emergence of AI-driven “Active Sensing” environments as a pivotal component of technical resilience, mitigating the loneliness epidemic while maintaining institutional efficiency. The findings suggest that social resilience in SCRCs is not merely a product of physical accessibility but is theoretically inferred by experts to be deeply rooted in the synergy of Bonding and Bridging Social Capital, rather than being a directly measured outcome. This research provides urban planners and policy-makers with a robust, evidence-based toolkit to design inclusive, resilient, and culturally aligned aging-in-place environments in the face of unprecedented demographic challenges.

1. Introduction

1.1. The Global Emergence of Super-Aging Urbanism

As of 2026, the global demographic landscape has reached a historic tipping point. With the population aged 65 and above exceeding 850 million, many developed nations—including Japan, Germany, and Taiwan—have officially transitioned into “Super-Aging Societies.” The strategic pivot from the National Ten-year Long-term Care Plan 2.0 (LTC 2.0) to the upcoming Long-term Care Plan 3.0 (LTC 3.0) in 2026 reflects Taiwan’s systemic adaptation to this demographic pressure. Official evaluations by [1] report that LTC 2.0 achieved a service satisfaction rate of 90% and successfully established a three-tier community-based care network (Tiers A, B, and C) serving over 756,000 recipients by 2025. Building on these achievements, LTC 3.0 focuses on the deep integration of home, community, and institutional services to enhance care continuity. This study’s evaluation of SCRC models serves as a direct academic response to these official mandates for dignified, localized aging-in-place environments. Taiwan presents a critical and urgent case study for global super-aging research due to its unprecedented aging velocity—transitioning from an ‘aged’ to a ‘super-aged’ society in just eight years, a pace significantly faster than many European nations. Furthermore, empirical data underscores the structural vulnerability of Taiwan’s urban aging: over 50% of the elderly population in metropolitan areas resides in multi-story walk-up apartments without elevators, creating severe spatial exclusion and ‘vertical isolation’. While the Ministry of Health and Welfare reports a 90% satisfaction rate for Long-term Care 2.0, the system currently lacks a validated evaluative framework for the emerging ‘Cluster Living’ models prioritized under the 2026 LTC 3.0 mandate. This research fills this gap by providing an evidence-based toolkit tailored to these specific socio-spatial and cultural pressures, offering a strategic blueprint for other rapidly aging East Asian societies sharing similar demographic and cultural trajectories.

This demographic transition aligns with the [2] Global Age-Friendly Cities framework, which identifies “Housing” and “Social Participation” as critical domains for healthy aging. However, as noted in the recent urban gerontology literature (e.g., [3]), traditional urban fabrics often fail to accommodate the complex socio-spatial needs of the oldest-old, leading to “environmental proactivity” gaps. The compromise of social resilience in Taiwan’s super-aged urbanism is not merely a demographic byproduct but a result of three intersecting systemic failures. First, architectural and spatial barriers—specifically the prevalence of multi-story walk-up apartments—create ‘vertical hollowing-out,’ forcing seniors into involuntary confinement and eroding their environmental proactivity. Second, the fragmentation of traditional social capital due to rapid urbanization has replaced intergenerational bonding with a reliance on professional but socially thin care systems, fueling a ‘loneliness epidemic’. Finally, a cultural mismatch in current residential models, which often replicate Western-centric institutional designs, ignores the deep-seated fear of ‘institutional abandonment’ in East Asian societies. These factors collectively impair the adaptive capacity of communities, necessitating a transition toward SCRC models that prioritize socio-technical integration and cultural alignment. While the community resilience discourse [4] emphasizes social capital as a primary adaptive capacity, there is a lack of comparative housing policy studies exploring how these capacities are culturally mediated in East Asian contexts. Consequently, this study moves beyond normative claims of “inclusive design” to provide an analytical evaluation of how cultural dimensions dictate the operational success of SCRCs. By 2050, the global elderly population is projected to reach 1.6 billion. In OECD countries, where over 90% of older adults prefer “Aging-in-Place,” the demand has evolved from “fixed housing” to diverse models emphasizing autonomy, social connectivity, and Ambient Assisted Living (AAL) [5]. Modern urbanism must now integrate generative AI for real-time risk mitigation and automated care records to manage this scale.

Traditional retirement models, however, often prioritize clinical safety at the expense of social connectivity, inadvertently fueling a “loneliness epidemic.” This research posits that Social Resilience—the capacity of a community to adapt to and recover from age-related stressors through social capital—is the bedrock of sustainable aging. Consequently, Sustainable Care Retirement Communities (SCRCs) have emerged as integral socio-technical systems within the “15 min city” framework, where healthcare, commerce, and social participation are accessible within a short walkable radius.

1.2. Social Resilience and Resilient Aging-in-Place

The theoretical foundation of this study is constructed upon a tripartite framework that integrates urban gerontology, social psychology, and community resilience. First, we adopt the [4] Resilience Theory, which posits that social resilience is a ‘network of adaptive capacities’ encompassing social capital and community competence. Second, this research is anchored in Hofstede’s Cultural Dimensions Theory, utilizing it as a lens to explain the variance in residential preferences between Western individualistic models and Eastern ‘connected autonomy’. Finally, the framework is grounded in the [2] Age-Friendly Cities criteria, ensuring that the 26 indicators are empirically linked to the functional needs of the oldest-old.

To ground the evaluation of the built environment in a rigorous spatial-gerontological context, this study incorporates the [6] Ecological Theory of Aging. This model posits that the well-being and functional autonomy of older adults are determined by the dynamic balance between an individual’s ‘competence’ (biological health and sensory-cognitive capacity) and ‘environmental press’ (the demands or stressors imposed by the physical and social setting). Within the SCRC framework, social resilience is achieved by strategically reducing environmental press through spatial interventions—such as Universal Design and AI-driven ‘Active Sensing’—which enable residents with declining competence to maintain a high level of ‘Person-Environment (P-E) Fit’. By optimizing this fit, SCRCs mitigate the risk of ‘environmental docility’ and empower seniors as active agents within the 15 min city fabric.

Social resilience in the context of urban aging transcends the physical “hardening” of the environment. Rather than a static state, this study conceptualizes social resilience as a dynamic “network of adaptive capacities,” a perspective pioneered by [4]. This theoretical lens posits that community resilience is not merely the durability of physical infrastructure but fundamentally hinges on the organic integration of information flow, community competence, and social capital. By adopting this analytical approach, we move beyond a descriptive definition to evaluate how these adaptive networks are operationalized within SCRCs. It encompasses the social and operational dimensions that allow residential clusters to maintain core functions amidst internal transitions, such as population aging and health decline.

While traditional discourse often frames older adults as a “vulnerable population,” resilience theory highlights them as active agents possessing rich life experience and local knowledge [7]. From a gerontological perspective, strengthening resilience requires a dual approach to the aging process:

• Intrinsic (Normative) Aging: Addressing natural biological progression through supportive spatial design.
• Extrinsic Aging: Mitigating health decline caused by unhealthy lifestyles or physical inactivity by optimizing service systems.

To operationalize social resilience, this study identifies three critical dimensions of social capital that sustain the SCRC framework:

• Bonding Social Capital: Strong ties within families and peer residents (e.g., neighborhood mutual aid). In modern SCRCs, this manifests as horizontal support networks that provide immediate emotional and functional assistance [8].
• Bridging Social Capital: “Weak ties” that cross geographical or organizational boundaries, such as links to NPOs or external community hubs. These networks facilitate information flow and prevent social isolation [9].
• Linking Social Capital: Vertical connections between residents and institutional power, such as government agencies or healthcare systems. In the context of Taiwan’s “Smart Rainforest” initiative, this determines the efficiency of resource delivery, including AI-driven subsidies and digital health records [10,11].

1.3. Bridging the Cultural Gap in SCRC Models

A significant limitation in the current literature is the Western-centric nature of Continuing Care Retirement Community (CCRC) models. Empirical evidence from the East Asian context suggests that the direct transplantation of Western-style gated CCRC models often encounters significant operational and social challenges. Studies on the early implementation of high-end retirement communities in China and Japan report lower-than-expected occupancy rates and social isolation, primarily due to a cultural dissonance with Western-centric ‘age-segregated’ designs. Specifically, the Western emphasis on individualistic autonomy often conflicts with the deeply rooted value of ‘connected autonomy’ in filial-piety-oriented societies, where seniors perceive total segregation as institutional abandonment. These implementation gaps justify the need for a localized framework that prioritizes intergenerational synergy and uncertainty avoidance, as evidenced by our expert consensus on these dimensions. To develop a model responsive to Taiwan’s sociocultural context, this research utilizes Hofstede’s Cultural Dimensions Theory. While Hofstede’s framework is primarily a macro-level construct developed for analyzing national cultures, it remains a robust baseline for identifying broad sociocultural variations that influence residential preferences in non-Western contexts [6]. We acknowledge the limitation that macro-dimensions may oversimplify subcultural nuances at the individual level; however, this study intends to operationalize these concepts at the community scale through a ‘theoretical-to-empirical’ translation. By mapping macro traits—such as Uncertainty Avoidance and Collectivism—onto specific micro-indicators like ‘institutional transparency’ and ‘intergenerational inclusion,’ we utilize the Fuzzy Delphi Method (FDM) to bridge the gap between national theory and residential-scale application. This ensures that the resulting 26 factor indicators are not merely cultural labels but functional drivers of social resilience within the SCRC model. National culture dictates the “perceived importance” of community attributes:

• Collectivism vs. Individualism: Unlike Western models that prioritize absolute autonomy, East Asian SCRCs often rely on “Filial Piety” and intergenerational connectivity for success.
• Uncertainty Avoidance: In societies like Taiwan, “Institutional Transparency” and “Medical Continuity” are identified as high-priority factors for resident security.

A critical review of the existing urban gerontology literature reveals two significant theoretical gaps. First, while current community resilience models (e.g., [4]) emphasize social capital, they often remain abstract and lack localized evaluative indicators tailored to the non-Western sociocultural fabric. Second, established residential models like the CCRC are fundamentally grounded in Western-centric ‘individualism,’ which fails to account for the ‘connected autonomy’ valued in East Asian societies.

The innovation of this study lies in its dual-layered approach to ‘Cultural Translation.’ Unlike previous descriptive studies, this research provides a methodological breakthrough by operationalizing macro-level cultural dimensions (Hofstede) into micro-level residential indicators through the Fuzzy Delphi Method (FDM). By integrating the National Culture Dimension as a core evaluative pillar alongside Management, Care Services, and Physical Environment, this study establishes the first culturally sensitive social resilience framework specifically designed for Taiwan’s Long-term Care 3.0 ecosystem. This enables a transition from a ‘Universalistic’ design paradigm to a ‘Culturally Aligned’ resilience strategy. Based on the identified gaps in cross-cultural urban gerontology, this study aims to establish a culturally sensitive and resilience-oriented evaluation framework for SCRC service models tailored to the sociocultural context of Taiwan. To achieve this objective, the research addresses the following three research questions:

RQ1:

What key factor indicators constitute a social resilience-oriented cluster living service model for the elderly within the context of Taiwan’s “Long-term Care 3.0” policy?

RQ2:

To what extent does national culture—specifically Hofstede’s dimensions—influence the prioritization of SCRC community attributes among diverse stakeholders in non-Western contexts?

RQ3:

Which service indicators achieve professional consensus through the Fuzzy Delphi Method (FDM), and what are the strategic implications for resilient urban aging policy?

2. Cluster Living Patterns of the Elderly

2.1. Elderly Care Needs and Residential Matching

The evolving self-care needs of older adults are the primary determinants of residential service models. Rather than static requirements, these needs shift dynamically across a health continuum: from a healthy stage emphasizing social participation to a disability stage requiring intensive nursing. This trajectory directly informs the Care Services Dimension (D2) of our evaluation framework. Specifically, the transition from functional independence to chronic health decline necessitates the “Professional Medical Integration” and “Psychological & Social Support” indicators to mitigate the social isolation often associated with physical fragility. Furthermore, the identification of the “sub-healthy” stage through frailty assessments serves as the rationale for the Physical Environment Dimension (D3). To foster social resilience, the living environment must shift from reactive clinical settings to preventive, supportive spaces. This necessitates the integration of “AI-Driven Active Sensing” and “Universal Design” indicators, which ensure real-time risk mitigation and maintain the functional independence of residents within the “15 min city” framework. By strategically matching residential services to self-care capacities—ranging from modular smart products for independent living to revitalized social housing with 24 h nursing—SCRCs can ensure a seamless care trajectory that preserves both resident dignity and institutional efficiency.

Table 1. Comparative analysis of elderly care service models in the United Kingdom, the United States, the Netherlands, and Japan.

Country	Model	Policy Orientation	Core Service Philosophy	Strategic Approach
UK	Age UK	Privatization & PbR (Payment by Results)	Preventive well-being; shrinking resource optimization.	Community-centered; data-driven identification of sub-healthy elders.
USA	On Lok (PACE)	Integrated Long-term Care & Medical Services	All-inclusive capitation; cultural & linguistic sensitivity.	“Aging in place” via multidisciplinary assessment and home-based support.
Netherlands	Buurtzorg	Universal coverage; focus on mental & physical disability.	Bureaucracy-free; nurse-led professional autonomy.	“Onion Model” centered on client self-management and informal networks.
Japan	CCRC	Public LTC Insurance (Ages 40+)	Regional revitalization; active “second life” promotion.	Multi-generational co-living; revitalizing vacant housing and community stores.

maps these health statuses to corresponding service models within the SCRC framework.

2.2. Global Service Models: Comparative Synthesis and Cultural Foundations

The evolution of elderly residential models reflects distinct strategies rooted in localized industrial structures and sociocultural values. A comparative analysis of leading international models—such as the privatization-oriented ‘Age UK’, the all-inclusive ‘On Lok PACE’ (USA), the decentralized ‘Buurtzorg’ (The Netherlands), and the regionally embedded Japanese CCRC—reveals that service delivery is fundamentally shaped by each nation’s resource allocation philosophy.

These variations demonstrate that the operational success of “Aging-in-Place” is mediated by national cultural dimensions. For instance, cultures characterized by high uncertainty avoidance (e.g., Japan) prioritize medical continuity and institutional trust, whereas individualistic cultures (e.g., USA/UK) place a higher valuation on home-based autonomy and service flexibility. This cross-cultural synthesis provides the primary theoretical rationale for integrating the National Culture Dimension (D4) into our evaluation framework. By analyzing how cultural values dictate residential preferences, this study ensures that the 26 factor indicators are culturally aligned to foster localized social resilience in non-Western contexts.

The fundamental value of the SCRC lies in its one-stop service continuum (independent living → assisted living → skilled nursing). However, older adults from diverse backgrounds exhibit varying psychological thresholds regarding “leaving home” versus “accepting institutional care” (

Table 2. Hofstede’s Cultural Dimensions disparities impact on cluster living preferences.

Country & Model	Dominant Cultural Dimension	Critical Selection Factors (Prioritization)
UK: Age UK	Low Power Distance	Emphasis on information transparency, autonomy, and individual choice in service navigation.
USA: On Lok	Individualism	High valuation of service flexibility and the preservation of home-based autonomy.
Netherlands: Buurtzorg	Femininity/Low Power Distance	Focus on the quality of nurse–patient relationships and the strength of community mutual-aid networks.
Japan: CCRC	High Uncertainty Avoidance	Prioritization of medical continuity stability, institutional trust, and a sense of social belonging.

).

3. Materials and Methods

3.1. Study Design and Data Sources

This study aims to identify the critical success factors of cluster living service models in fostering well-being and social resilience among older adults. To achieve this, we developed a comprehensive hierarchical framework based on an extensive literature review, which serves as the foundational screening criteria for the Fuzzy Delphi Method (FDM) questionnaire and subsequent validation.

The ultimate goal of this framework is to operationalize “Social Resilience” into measurable indicators that guide the construction of sustainable Care Retirement Communities (SCRCs). The established framework comprises 4 major dimensions and 26 factor indicators, as detailed below:

• Management Dimension: Evaluates the operational sustainability and efficiency of the service model. Key factors include organizational structure, financial feasibility, human resource allocation, quality management systems, and strategic marketing.
• Care Services Dimension: Focuses on the professional integration and personalization of care. This encompasses the synergy between medical intervention, preventive healthcare, and rehabilitation support to meet the dynamic health needs of older adults [12,13].
• Physical Environment Dimension: Emphasizes the spatial design of the living environment. This includes barrier-free infrastructure, the application of Smart Living Technologies (SLT), public space configuration, and safety management, aimed at supporting independent living and social participation.
• National Culture Dimension: Examines the socio-technical acceptance of residential models. This dimension analyzes how cultural values, social customs, family structures, and regulatory systems dictate elderly residential choices, ensuring that proposed strategies are aligned with the local sociocultural context.

This multidimensional analytical framework systematically defines and verifies the core factors affecting the successful operation of cluster living models. By integrating literature synthesis with FDM consensus, the study provides a robust empirical basis for future policy development and service optimization. The conceptual hierarchy of the study is illustrated in Figure 1.

3.2. Research Variables and Indicators

The hierarchical framework of this study is categorized into four dimensions: Management, Care Services, Physical Environment, and National Culture. The 26 indicators extracted in this study are derived from a ‘cultural translation’ of the eight domains within the [2] Global Age-Friendly Cities framework, integrated with the key urban gerontology criteria proposed by [3]. This synthesis ensures that the indicators possess both international dialogue capabilities and robust scientific anchoring. The D3 indicators are grounded in Lawton’s Ecological Theory of Aging, which emphasizes balancing ‘environmental press’ with resident competence through spatial interventions [14]. This is further aligned with the physical domains of the WHO Age-Friendly Cities framework [2], specifically focusing on the synergy between residential safety and urban accessibility.

Table 3. Dimensions and definitions of key factors in SCRC service models.

Dimension	Factor Indicator	Operational Definition	Citations for the Indicators
Management	1. Charging Model	Analysis of fee structures: bundled care vs. fee-for-service; security deposits vs. monthly rental models.	[8,9,11,15]
	2. Staff-to-Resident Ratio	Compliance with Taiwan’s LTC standards (e.g., 1:8 for care attendants; 1:20 for nurses; 1:80 for social workers).
	3. Capital Investment Model	Comparison between heavy-asset (new construction) and light-asset (renovating idle government spaces) approaches.
	4. Financial Cost Control and Management (Transparency)	Management of high-capital expenditures, including staffing, facility maintenance, and service planning to ensure ROI.
	5. Brand Awareness	The perceived reputation and trust in the operator (public vs. private) and their track record in elderly care.
	6. Geographic Location	Accessibility analysis: urban proximity, transportation connectivity, and environmental/scenic quality.
	7. Zoning with Care Continuity	Spatial division into Independent Living (IL), Assisted Living (AL), Skilled Nursing (SN), and Memory Support (MS).
Care Services	1. Professional Medical Integration	Provision of on-site medical personnel and equipment to ensure a seamless healthcare-to-housing link.	[8,12,13]
	2. Psychological & Social Support	Systems for counseling and “life companionship” to mitigate the emotional impact of physical decline.
	3. Nutritional Quality & Diversity	Diversity, hygiene, and nutritional balance of meal options tailored to geriatric dietary needs.
	4. Lifelong Learning Activities	Social engagement programs designed to enhance cognitive function and prevent social isolation.
	5. Maintenance & Hygiene	Protocols for regular disinfection, cleanliness of private/public spaces, and facility inspection.
	6. Personal Financial Planning Assistance	Services helping residents manage long-term living expenses, medical trusts, and end-of-life financial allocation.
	7. Emergency Notification System	Comprehensive notification infrastructure (e.g., pull-cords, wearable sensors) across all residential zones.
Physical Environment	1. Unit Configuration	Variety of residential options (single/double occupancy) with emphasis on private bathroom accessibility.	[10,11,15]
	2. Facility Diversification	Integration of multi-use spaces: dining, fitness, retail, beauty services, and medical clinics within the cluster.
	3. Universal Design & Wayfinding	Circulation planning utilizing barrier-free standards and intuitive directional signage for cognitive support.
	4. Security & Privacy	Balance between 24/7 security patrols/access control and the resident’s right to personal privacy.
	5. Architecture & Landscape	Aesthetic integration of the built environment with local topography and biophilic design elements.
	6. Fire Safety Compliance	Installation of fire suppression and evacuation systems exceeding statutory safety requirements.
	7. AI-Driven “Active Sensing”	Integration of IoT, Wave radar, and AI for unobtrusive fall detection and behavioral risk recognition.
National Culture	1. Power Distance	Preference for service models endorsed by medical authorities or formal government policy hierarchies.	[6,16]
	2. Individualism vs. Collectivism	Focus on intergenerational inclusion to mitigate the “stigma of abandonment” associated with institutional care.
	3. Masculinity vs. Femininity	Attraction to high-end facility status symbols and achievement-oriented marketing vs. nurturing care models.
	4. Uncertainty Avoidance	Demand for clear financial contracts, refund mechanisms, and guaranteed continuity of care.
	5. Long-term Orientation	Perception of SCRC entry as a strategic “long-term health asset” and preventive life planning.

defines the 26 factor indicators used to construct the Fuzzy Delphi questionnaire.

3.3. Fuzzy Delphi Method (FDM)

The traditional Delphi method seeks expert consensus through multiple rounds of anonymous questionnaires. However, to address the ambiguity of human perception and the high cost of repeated iterations, this study employs the Fuzzy Delphi Method (FDM). While grounded in the established principles of [17], FDM remains a predominant and robust methodological tool in contemporary research for screening critical indicators in complex urban systems and long-term care frameworks [18,19]. By integrating Fuzzy Set Theory with the expert opinion aggregation process, FDM uses Double Triangular Fuzzy Numbers to identify consensus and applies the Gray Zone Test to ensure convergence [17].

3.3.1. Expert Scoring Scale

To evaluate the importance of each indicator, a linguistic scale with a corresponding 0–10 numerical range was provided to the experts. To accommodate the vagueness of expert judgments, this study employs a linguistic scale ranging from 0 to 10. These linguistic variables are quantified into Triangular Fuzzy Numbers (TFNs), denoted as A = (l, m, u). In this notation, l represents the conservative value (lower bound), m represents the most likely value (membership degree = 1), and u represents the optimistic value (upper bound). The mapping between the linguistic preferences and their corresponding TFNs is presented in

Table 4. Mapping of linguistic variables to numerical evaluation scores (0–10).

Linguistic Variable	Numerical Range	Description
Very Unimportant	0.0–2.0	The indicator has a negligible impact on SCRC resilience.
Unimportant	2.1–4.0	The indicator has a minor or indirect influence.
Moderate	4.1–6.0	The indicator has a balanced or average impact.
Important	6.1–8.0	The indicator is a key driver for supply chain resilience.
Very Important	8.1–10.0	The indicator is essential and indispensable for SCRC.

.

The operational steps for the FDM are as follows:

3.3.2. Step 1: Establishing Cognitive Intervals

Experts provide a “most likely interval value” for each indicator. The minimum value represents the Conservative Cognition Value ($G^i$), and the maximum represents the Optimistic Cognition Value (Oⁱ).

3.3.3. Step 2: Statistical Analysis and Outlier Exclusion

Extreme values exceeding two standard deviations are excluded. For each item i, we calculate:

• Conservative Values: Minimum ($C_L^i$), Geometric Mean ($C_M^i$), and Maximum ($C_U^i$).
• Optimistic Values: Minimum ($O_L^i$), Geometric Mean ($O_M^i$), and Maximum ($O_U^i$).

3.3.4. Step 3: Constructing Triangular Fuzzy Numbers

Based on Step 2, two triangular fuzzy numbers are established for each item i:

• Conservative Triangular Fuzzy Number: $C^i$ = ($C_L^i$, $C_M^i$, $C_U^i$).
• Optimistic Triangular Fuzzy Number: $O^i$ = ($O_L^i$, $O_M^i$, $O_U^i$).

3.3.5. Step 4: Testing for Expert Consensus (Gray Zone Test)

The degree of consensus is determined by analyzing the overlap between $C^i$ and $O^i$ (Figure 2):

Case 1: No Gray Zone ($C_U^i\le O_L^i$)

If the maximum conservative value is less than or equal to the minimum optimistic value, a consensus range exists. The Consensus Importance Value ($G^i$) is calculated as:

$$G^i= {\displaystyle \frac{o_M^i+{ c}_M^i}{2}}$$

Case 2: Gray Zone with Convergence ($C_U^i>O_L^i$ and $Z^{i }$ < $M^{i }$)

A gray zone ($Z^{i }$=$C_U^i-O_L^i$) exists, but the uncertainty interval ($M^{i }$=$O_M^i-C_M^i$) is larger than the gray zone. This indicates a fuzzy overlap, but expert opinions remain statistically convergent. The $G^i$ value is determined by the intersection of the two fuzzy sets:

$$F^i\left(x_j\right)=\left\{\int x\left\{min\left[C^i\left(x_j\right),O^i\left(x_j\right)\right]\right\}\left.dx\right\}\right.$$

Specifically, it is calculated as:

$$G^i=\left\{x_j\left|\max u_{F^i}\left(x_j\right)\right.\right\}$$

Case 3: Gray Zone with Divergence ($C_U^i>O_L^i$ and $Z^{i }$ > $M^{i }$)

If the gray zone is larger than the uncertainty interval, expert opinions are significantly divergent. These items are returned to the experts for a second round of evaluation until Case 1 or Case 2 is achieved.

3.3.6. Justification for the Single-Round Protocol

While traditional Delphi protocols often require multiple iterative rounds to achieve convergence, this study adopted a single-round FDM. This approach is methodologically justified by FDM’s ability to capture the “cognitive uncertainty” of experts through interval values (Optimistic and Conservative) rather than single-point estimates [20,21]. By integrating Triangular Fuzzy Numbers and the Gray Zone Test, FDM can mathematically identify consensus and resolve ambiguity in a single step, thereby avoiding the high dropout rates and “opinion fatigue” associated with traditional iterative surveys [22]. In this study, the high expertise and seniority of the panel (averaging > 20 years) ensured that the initial cognitive intervals were well-defined, reaching statistical convergence without the need for additional rounds.

3.3.7. Threshold Setting

The final Consensus Threshold for this study is determined by the arithmetic mean of the geometric means across all items. Indicators with a $G^i$ value exceeding this threshold (e.g., $G^i$ > 7.8) are identified as the key factors for constructing the SCRC model.

3.3.8. Sensitivity Analysis of the Consensus Threshold

To verify the robustness of the indicator selection, a sensitivity analysis was conducted by shifting the threshold ($G^i$ = 7.80) by ±0.2.

Stringent Scenario (Threshold $G^i$ ≥ 8.0)

When the threshold is increased to 8.0, the selection process becomes more rigorous, focusing only on “highly critical” factors.

• Results Change: Under this scenario, indicators such as “Lifelong Learning Activities (7.86)” and “Uncertainty Avoidance (7.98)” would be excluded.
• Impact: While the model would become more streamlined, it might overlook the “Psychosocial Support” dimension, which experts identified as a significant trend in modern SCRC services. However, core safety and operational indicators (e.g., Fire Safety, Charging Model) remain unaffected, confirming their absolute necessity.

Relaxed Scenario (Threshold $G^i$ ≥ 7.6)

When the threshold is lowered to 7.6, the model adopts a more inclusive approach toward secondary factors.

• Results Change: Indicators such as “Emergency Notification System (7.69)” and “Geographic Location (7.20)” would be closer to or included in the selection.
• Impact: A lower threshold would introduce more environmental and logistical factors. However, this might dilute the focus of the SCRC model on “active sensing” and “resilience,” potentially introducing noise into the subsequent weight calculations in the ANP phase (

  Table 5. Sensitivity analysis of the indicator selection threshold.

  Scenario	Threshold	No. of Selection Factors	Factors Changes
  S1 (Relaxed)	${\mathrm{G}}^{\mathrm{i}}$ ≥ 7.6	20	Adds “Emergency Notification System (7.69),” increasing model complexity but lowering the overall average consensus.
  S2 (Baseline)	${\mathrm{G}}^{\mathrm{i}}$ ≥ 7.8	19	The current study’s choice. Maintains a balance between comprehensive coverage and expert consensus.
  S3 (Stringent)	${\mathrm{G}}^{\mathrm{i}}$ ≥ 8.0	13	Excludes 6 marginal indicators, such as “Nutritional Quality (7.98)” and “Lifelong Learning (7.86),” potentially narrowing the service scope.

  ).

The current threshold of 7.80 (the arithmetic mean) strikes an optimal balance. It ensures that all selected indicators fall within the “Very Important” linguistic range ($G^i$ > 8.0) or at the high end of “Important,” maintaining both the breadth of the CCRC ecosystem and the precision of the mathematical model. The core structure of the SCRC remains stable within a ±0.2 variance, demonstrating high model reliability.

4. Results

4.1. Analysis of Expert Consensus Using the Fuzzy Delphi Method (FDM)

4.1.1. Expert Selection and Composition

To ensure the objectivity and authority of the research indicators, this study employed the FDM to solicit insights from a highly specialized panel of experts. The selection process utilized purposive sampling based on the “Triple Helix” model, integrating three core dimensions: industrial practice, academic rigor, and policy feasibility. A total of seven experts were invited, and seven valid questionnaires were collected between 1 December 2025 and 31 January 2026 (100% response rate). In FDM and multi-criteria decision-making (MCDM) research, the professional depth and heterogeneity of the experts are prioritized over raw sample size for reaching mathematical consensus. The literature suggests that a focused panel of 5 to 12 experts with high domain authority is enough to achieve rigorous results [23]. For instance, recent studies in green building evaluation and smart senior care have successfully utilized panels of 6 to 7 experts, emphasizing that extensive seniority (minimum 10 years) is the primary determinant of data quality in small samples [21,24].

Consistent with these academic precedents, the experts invited for this study possess substantial seniority (

Table 6. Characteristics and Expertise of the FDM Panel Members.

No.	Category (N of Experts)	Expertise and Professional Background	Gender	Age	Year of Working Experience	Position	Organization Size
1	Industry (3)	Management and operation of a well-known CCRC and elderly wellness housing.	Male	60+	21–30	Head of Unit	101–500
2			Male	40–49	11–20	Professionals	101–500
3			Female	40–49	11–20	Professionals	101–500
4	Academia (2)	Research in elderly housing architecture, spatial planning, and co-living communities.	Male	60+	21–30	Head of Unit	101–500
5			Female	60+	21–30	Professionals	N/A
6	Government (2)	Central and local government officials in charge of urban development and housing policy.	Male	50–59	21–30	Head of Unit	501–100
7			Female	50–59	21–30	Head of Unit	101–500

), with the majority exceeding 20 years of professional experience. The panel represents a comprehensive microcosm of Taiwan’s senior housing ecosystem, structured as follows:

• Industry Practitioners (n = 3): Senior executives and chief operators of prominent Continuing Care Retirement Communities (CCRCs) and wellness housing in Taiwan. They provide critical insights into market demand, operational management, and the practical implementation of elderly service models.
• Academic Scholars (n = 2): Professors specializing in elderly housing architecture, spatial planning, and co-housing communities. Their participation ensures theoretical robustness and alignment with global “aging-in-place” and “super-aged urbanism” trends.
• Government Professionals (n = 2): High-ranking officials from the Ministry of the Interior and local governments responsible for urban development. They provide essential perspectives on regulatory compliance, land-use strategies, and alignment with Taiwan’s Long-Term Care (LTC) 3.0 policy.

This cross-sectoral composition ensures that the findings reflect the multifaceted reality of the Taiwanese context. Furthermore, all participants confirmed the absence of conflicts of interest, and the study adhered to strict ethical standards, ensuring total anonymity and non-identifiable data processing.

4.1.2. Data Collection Process

The questionnaire was conducted in one round. The experts were asked to provide their professional judgment on the importance of each secondary indicator using linguistic variables. These variables were subsequently converted into triangular fuzzy numbers to account for the subjectivity and uncertainty inherent in expert opinions.

Following the data compilation and analysis using the FDM, the threshold value for this study was established at 7.8, determined by the arithmetic mean of the geometric means across all indicators. Factors with a consensus importance value ($G^i$ ≥ 7.8) were retained as core elements of the SCRC model, while those falling below this threshold ($G^i$ < 7.8) were excluded.

The statistical analysis resulted in the exclusion of seven indicators, leaving 19 key factors (73.08% of the initial set) as the validated framework for social resilience-oriented cluster living. The detailed calculation results, including the geometric means and final $G^i$ values, are presented in Table 5, and the refined hierarchical structure is illustrated in Figure 2. In response to RQ1, the FDM analysis successfully validated a framework of 19 key factor indicators (73.08% of the initial set) that constitute the social resilience-oriented SCRC model.

4.2. Stability and Convergence Analysis

A critical measure of FDM reliability is the convergence of expert opinions. In this study, the stability of the identified indicators was verified through the evaluation of expert opinion intervals, which consistently resulted in values of −1 or 0 (see Table 7).

• No Gray Zone (Interval = −1): Indicates a clear consensus range among experts, where opinions exhibit high uniformity.
• Convergent Gray Zone (Interval = 0): Signifies a minor fuzzy overlap in judgments; however, the overlap does not lead to statistical divergence (Zi ≤ Mi), confirming that a stable consensus has been reached.

These results demonstrate a high level of agreement among stakeholders within Taiwan’s Long-term Care 3.0 ecosystem. Based on the consolidated consensus values ($G^i$) provided in Table 5, the indicators were filtered against a threshold of 7.80.

4.2.1. Analysis of Retained Core Indicators ($G^i$ ≥ 7.80)

The experts prioritized safety, operational stability, and policy alignment as the pillars of the SCRC service model. Fire Safety Compliance ($G^i$ = 8.56) and the Charging Model ($G^i$ = 8.49) emerged as the most critical factors, representing the foundational requirements for physical security and financial sustainability.

Furthermore, indicators such as Zoning with Care Continuity and Security & Privacy (both $G^i$ = 8.23) reflect a strong alignment with Taiwan’s Long-Term Care 3.0 policy, emphasizing the need for spatial layouts that respect resident dignity. The inclusion of AI-Driven “Active Sensing” ($G^i$ = 8.08) underscores the consensus that technological integration is no longer optional but a necessary component of modern geriatric spatial design.

4.2.2. Indicators Below the Threshold ($G^i$ < 7.80)

Indicators that failed to meet the consensus threshold were excluded to maintain the model’s precision. Personal Financial Planning Assistance ($G^i$ = 6.94) and internal logistics factors like the Capital Investment Model ($G^i$ = 7.42) were viewed as peripheral to the immediate quality of residential life.

Interestingly, in the National Culture dimension, experts deprioritized Power Distance and Masculinity vs. Femininity. This shift suggests that stakeholders in Taiwan’s aging context now value horizontal social capital and intergenerational inclusion over traditional hierarchical authority or symbolic institutional status.

• Note 1: The symbol “○” denotes that $C_U^i\le O_L^i$, indicating that the interval judgments of the experts have reached a consensus range without a gray zone. In such cases, the consensus importance value is calculated as $G^i$ = ${\displaystyle \frac{c_M^i+o_M^i}{2}}$.
• Note 2: Gray-shaded rows represent evaluation criteria.

Table 7. FDM Analysis table.

Dimension	Evaluation Criterion	$({\mathit{O}}_{\mathit{L}}^{\mathit{i}}$$,{\mathit{O}}_{\mathit{M}}^{\mathit{i}}$$,{\mathit{O}}_{\mathit{U}}^{\mathit{i}}$)			$({\mathit{C}}_{\mathit{L}}^{\mathit{i}}$$,{\mathit{C}}_{\mathit{M}}^{\mathit{i}}$$,{\mathit{C}}_{\mathit{U}}^{\mathit{i}}$)			${\mathit{M}}^{\mathit{i}}$	${\mathit{Z}}^{\mathit{i}}$	Interval	${\mathit{G}}^{\mathit{i}}$
Management	Charging Model	9	9.79	10	7	7.19	8	2.6	−1	○	8.49
	Staff-to-Resident Ratio	9	9.39	10	6	6.97	8	2.4	−1	○	8.18
	Capital Investment Model	7	8.91	10	5	5.93	7	3.0	0	○	7.42
	Financial Cost Control and Management (Transparency)	7	8.54	10	4	5.67	7	2.9	0	○	7.11
	Brand Awareness	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
	Geographic Location	7	8.72	10	4	5.67	7	3.0	0	○	7.20
	Zoning with Care Continuity	8	9.56	10	5	6.90	8	2.7	0	○	8.23
Care Services	Professional Medical Integration	8	9.36	10	6	6.97	8	2.4	0	○	8.17
	Psychological & Social Support	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
	Nutritional Quality & Diversity	8	9.17	10	6	6.79	7	2.4	−1	○	7.98
	Lifelong Learning Activities	8	8.96	10	6	6.76	8	2.2	0	○	7.86
	Maintenance & Hygiene	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
	Personal Financial Planning Assistance	7	8.52	10	3	5.36	7	3.2	0	○	6.94
	Emergency Notification System	7	9.31	10	4	6.07	7	3.2	0	○	7.69
Physical Environment	Unit Configuration	8	9.36	10	5	6.52	8	2.8	0	○	7.94
	Facility Diversification	8	9.36	10	6	6.76	8	2.6	0	○	8.06
	Universal Design & Wayfinding	9	9.39	10	6	6.97	8	2.4	−1	○	8.18
	Security & Privacy	8	9.56	10	5	6.90	8	2.7	0	○	8.23
	Architecture & Landscape	9	9.19	10	5	6.52	8	2.7	−1	○	7.85
	Fire Safety Compliance	9	9.79	10	6	7.33	9	2.5	0	○	8.56
	AI-Driven “Active Sensing”	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
National Culture	Power Distance	6	7.92	9	3	4.86	6	3.1	0	○	6.39
	Individualism vs. Collectivism	8	9.36	10	6	6.76	8	2.6	0	○	8.06
	Masculinity vs. Femininity	6	7.71	9	3	4.64	6	3.1	0	○	6.18
	Uncertainty Avoidance	8	9.17	10	6	6.79	7	2.4	−1	○	7.98
	Long-term Orientation	9	9.19	10	5	6.52	8	2.7	−1	○	7.85
Threshold Value		7.80

Table 7. FDM Analysis table.

Dimension	Evaluation Criterion	$({\mathit{O}}_{\mathit{L}}^{\mathit{i}}$$,{\mathit{O}}_{\mathit{M}}^{\mathit{i}}$$,{\mathit{O}}_{\mathit{U}}^{\mathit{i}}$)			$({\mathit{C}}_{\mathit{L}}^{\mathit{i}}$$,{\mathit{C}}_{\mathit{M}}^{\mathit{i}}$$,{\mathit{C}}_{\mathit{U}}^{\mathit{i}}$)			${\mathit{M}}^{\mathit{i}}$	${\mathit{Z}}^{\mathit{i}}$	Interval	${\mathit{G}}^{\mathit{i}}$
Management	Charging Model	9	9.79	10	7	7.19	8	2.6	−1	○	8.49
	Staff-to-Resident Ratio	9	9.39	10	6	6.97	8	2.4	−1	○	8.18
	Capital Investment Model	7	8.91	10	5	5.93	7	3.0	0	○	7.42
	Financial Cost Control and Management (Transparency)	7	8.54	10	4	5.67	7	2.9	0	○	7.11
	Brand Awareness	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
	Geographic Location	7	8.72	10	4	5.67	7	3.0	0	○	7.20
	Zoning with Care Continuity	8	9.56	10	5	6.90	8	2.7	0	○	8.23
Care Services	Professional Medical Integration	8	9.36	10	6	6.97	8	2.4	0	○	8.17
	Psychological & Social Support	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
	Nutritional Quality & Diversity	8	9.17	10	6	6.79	7	2.4	−1	○	7.98
	Lifelong Learning Activities	8	8.96	10	6	6.76	8	2.2	0	○	7.86
	Maintenance & Hygiene	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
	Personal Financial Planning Assistance	7	8.52	10	3	5.36	7	3.2	0	○	6.94
	Emergency Notification System	7	9.31	10	4	6.07	7	3.2	0	○	7.69
Physical Environment	Unit Configuration	8	9.36	10	5	6.52	8	2.8	0	○	7.94
	Facility Diversification	8	9.36	10	6	6.76	8	2.6	0	○	8.06
	Universal Design & Wayfinding	9	9.39	10	6	6.97	8	2.4	−1	○	8.18
	Security & Privacy	8	9.56	10	5	6.90	8	2.7	0	○	8.23
	Architecture & Landscape	9	9.19	10	5	6.52	8	2.7	−1	○	7.85
	Fire Safety Compliance	9	9.79	10	6	7.33	9	2.5	0	○	8.56
	AI-Driven “Active Sensing”	8	9.36	10	6	6.79	7	2.6	−1	○	8.08
National Culture	Power Distance	6	7.92	9	3	4.86	6	3.1	0	○	6.39
	Individualism vs. Collectivism	8	9.36	10	6	6.76	8	2.6	0	○	8.06
	Masculinity vs. Femininity	6	7.71	9	3	4.64	6	3.1	0	○	6.18
	Uncertainty Avoidance	8	9.17	10	6	6.79	7	2.4	−1	○	7.98
	Long-term Orientation	9	9.19	10	5	6.52	8	2.7	−1	○	7.85
Threshold Value		7.80

That fell below the threshold ($G^i$ < 7.80) and were subsequently deleted from the final framework to ensure the model’s precision and practical relevance.

• Core retained indicators (≧7.80):
  • Fire safety Compliance ($G^i$ = 8.56): This received the highest expert consensus value, indicating that experts regard safety as the foundation of long-term care facilities.
  • Charging model ($G^i$ = 8.49): Financial stability and transparency of the charging model are crucial to operation.
  • Zoning with care continuity ($G^i$ = 8.23) and Security & Privacy ($G^i$ = 8.23): These echo the emphasis on continuity of care in Long-Term Care 3.0 policy, highlighting the need for spatial layouts that respect elderly dignity.
  • Universal Design & Wayfinding ($G^i$ = 8.18) and Staff-to-resident ratio ($G^i$ = 8.18): These metrics are identified as the primary drivers of service quality and operational efficiency within the cluster living model.
  • AI-Driven “Active Sensing” ($G^i$ = 8.08): This aligns with the current project focus, and experts also recognize the necessity of AI in spatial design.
• Indicators below the threshold (to be considered for deletion or revision): Although these indicators are important, their relative priority was lower in the experts’ consensus:
  • Personal financial planning Assistance ($G^i$ = 6.94): Considered peripheral to the immediate residential service delivery.
  • Capital Investment Model ($G^i$ = 7.42) and Financial Cost Control and Management (Transparency) ($G^i$ = 7.11): Viewed as internal management logistics rather than direct determinants of resident well-being. The exclusion of ‘Financial Cost Control’ by experts does not imply a disregard for financial logic, but rather a prioritization of service quality and safety over internal administrative logistics.
  • Power distance and masculinity/femininity cultural indicators: The exclusion of these factors reveals a significant paradigm shift. This shows that, within the National Culture dimension, experts considered substantive intergenerational inclusion and uncertainty avoidance to be more important than pure institutional power or symbolic status. This suggests that social resilience in Taiwan’s aging context is increasingly rooted in horizontal social capital rather than vertical hierarchical authority.

4.3. Radar Chart of Expert Consensus

A radar chart plotted based on the expert consensus values ($G^i$) obtained from the FDM. This chart can intuitively present the relative importance of the four major dimensions and their specific criteria in constructing social resilience in Taiwan’s SCRC model.

• Axes: Each axis radiating outward from the center represents an evaluation criterion (e.g., charging model, fire safety, intergenerational inclusion, etc.).
• Consensus value ($G^i$): The farther the solid line point is from the center, the higher the expert consensus value for that criterion, and thus the higher its importance.
• Threshold dashed line (Red Dashed Line, T = 7.80): This is the screening benchmark established in this study. Outside the dashed line indicates that the criterion passed the screening and is a key success factor; inside the dashed line indicates that the criterion did not reach consensus in importance and may be removed from the model or regarded as secondary.
• Color blocks: Different colors represent the four major dimensions (Management, Care Services, Physical Environment, and National Culture), making it easier to observe overall performance across dimensions.

4.3.1. Graphic Distribution and Data Interpretation

The radar chart provides a spatial visualization of the 26 indicators, where the distance of each vertex from the center represents its perceived importance (G^i). Addressing RQ2, the graphic distribution confirms that national culture significantly dictates prioritization. The distribution reveals three distinct zones that define the resilience profile of Taiwan’s SCRC model:

A. Core Strength Zone (Outward Peaks)

The chart exhibits prominent outward expansion in the Physical Environment (e.g., Fire Safety Compliance, Security & Privacy) and Management (e.g., Charging Model) dimensions.

• Enabling Cornerstones: These peaks indicate that experts identify physical safety and economic viability as the foundational infrastructure required to sustain social resilience.
• Strategic Implication: Without a stable financial threshold, the delivery of psychosocial and cultural services becomes systemically fragile. A resilient community must guarantee fundamental security before layering more complex services.

B. Contraction Zone (Inward Recesses)

Significant inward recesses are observed within the National Culture dimension, specifically regarding Power Distance (G^i = 6.39) and Masculinity vs. Femininity (G^i= 6.18). These indicators fall significantly below the T = 7.80 threshold.

• Paradigm Shift: This contraction confirms a shift in contemporary Taiwanese elder care, where traditional hierarchical authority and status-driven consumption have been de-emphasized.
• Value Transition: Resilience is no longer defined by “prestige” (Economic Display), but by “stability” (Economic Security). Older adults no longer prioritize “deference to authority” as primary drivers when selecting a residential community.

C. Key Cultural Highlights

Despite lower scores across the broader cultural dimension, two indicators—Intergen erational Inclusion and Uncertainty Avoidance—successfully crossed the selection threshold.

• “Body and Spirit” Metaphor: These graphical highlights represent the functional purpose of SCRC resilience. While economic factors provide the “body” (structure), these indicators provide the “spirit” (purpose).
• Refined Need: This reflects a cultural transition toward “connected autonomy” rather than traditional “dependence.”

4.3.2. Consistency of Expert Opinion

The overall morphology of the radar chart boundary—specifically the relative smoothness among indicators that passed the threshold—reflects the degree of convergence among the expert panel. In this study, the majority of retained criteria fluctuate closely around the $G^i$ ≈ 8.0 mark. Within FDM research, this narrow variance across 19 different factors indicates an exceptionally high degree of consensus. This statistical stability lends strong persuasiveness to the study’s results, suggesting that the identified factors represent a unified vision of the future SCRC model among stakeholders in Taiwan.

The prominence of economic dimensions in the outward peaks does not diminish the role of social resilience; rather, it identifies the operational priority necessary to protect the community’s social integrity during long-term care cycles.

5. Discussion and Conclusions

5.1. Discussion

5.1.1. Management Dimension: From a “Real Estate Mindset” to a “Trust-Service Mindset”

• Key Indicators: Charging Model (8.49), and Zoning with Care Continuity (8.23).
• In-depth Analysis: The data reveal that expert concern for the charging model is second only to fire safety. This reflects a fundamental anxiety among Taiwanese older adults regarding long-term financial sustainability. While traditional retirement villages rely on high-entry deposits, a super-aged society demands asset liquidity and flexible allocation. This prioritization underscores that economic sustainability acts as the ‘enabling infrastructure’ for social resilience; without a viable financial vehicle, the community’s capacity to absorb social shocks—the core of social resilience—would be compromised.
• Policy Implications: An SCRC must evolve beyond a housing project into a financial vehicle. Integrating elder-care trust mechanisms can mitigate concerns regarding financial mismanagement or insolvency [25]. Furthermore, the high score for “Zoning with Continuity” underscores that “aging in place” is not just about the home, but about remaining within a familiar community even as care needs intensify.

5.1.2. Care Services Dimension: From “Physical Nursing” to “Whole-Person Support”

• Key Indicators: Professional Medical Integration (8.17); Psychological & Social Support (8.08).
• In-depth Analysis: In contrast to Western models that emphasize individualistic independence, our findings align with Eastern gerontology, where resilience is a collective property rooted in psychological companionship (8.08) and social capital.
• Academic Linkage: These findings suggest that non-medical support is a core determinant of well-being, echoing the “loneliness epidemic” cited in global gerontology. It should be clarified that the perceived synergy between social resilience and social capital identified here is a theoretical inference based on the high Gi values from expert consensus, rather than a directly measured empirical outcome. A resilient community must prioritize emotional support as highly as hygiene, requiring a multidisciplinary workforce that includes psychosocial experts alongside clinical staff.

5.1.3. Physical Environment Dimension: From “Barrier-Free” to “Ambient Sensing”

• Key Indicators: Fire Safety Compliance (8.56); Security & Privacy (8.23); AI-Driven “Active Sensing” (8.08).
• In-depth Analysis: AI-driven ‘Active Sensing’ (8.08) represents a transition from reactive to preventive resilience, where technology reduces the operational risk (Management) while simultaneously upholding the resident’s dignity (Social). AI investment is proposed as a risk-mitigation vehicle rather than a luxury amenity. In the context of Taiwan’s LTC 3.0, expensive sensor technologies serve to prevent catastrophic events (e.g., falls or fire fatalities), thereby ensuring the long-term economic and social viability of the SCRC model.
• Spatial Logic: Environmental design must now balance Active Sensing with privacy. By utilizing non-contact sensors (e.g., Wave radar), SCRCs can achieve real-time risk detection without the intrusive nature of cameras, thereby preserving the “Security & Privacy” (8.23) while maximizing life safety.

5.1.4. National Culture Dimension: From “Westernized Models” to “Cultural Translation”

• Key Indicators: Facility Diversification (8.06); Nutritional Quality & Diversity (7.98); Lifelong Learning Activities (7.86); Long-term Orientation (7.85).
• In-depth Analysis: This dimension represents the study’s most original finding. The high score for Intergenerational Inclusion challenges the Western CCRC tendency to create “elderly islands” (exclusive age-segregated enclaves). While previous research in the CCRC domain often prioritized ‘exclusive leisure’ as a mark of success, this study reveals a paradigm shift toward ‘connected autonomy.’ This divergence suggests that in East Asian contexts, social resilience is derived from multi-generational integration rather than age-segregated seclusion.
  • The “Voices of Children” Factor: Taiwanese older adults increasingly desire “connected autonomy”—the ability to live independently yet remain part of a vibrant, multi-generational social fabric where they can “see young people and hear children.” This reflects an Eastern preference for community connectedness over total seclusion [26].
  • Risk-Averse Pragmatism: The preference for robust financial/nursing guarantees over “facility luxury” aligns with high Uncertainty Avoidance. East Asian seniors prioritize long-term safety nets over hedonistic enjoyment.
  • Core Finding: The rejection of “Power Distance” and “Masculinity” suggests that contemporary seniors seek egalitarianism and inclusion rather than traditional Confucian hierarchical authority. This “Cultural Translation” is essential for localized policy design in Taiwan’s Long-term Care 3.0 era.

5.2. Conclusion and Policy Implications

5.2.1. Policy Recommendations and Practical Implications

Finally, answering RQ3, the professional consensus identified ‘Fire Safety’ and ‘Charging Models’ as foundational factors, leading to the strategic implications for LTC 3.0 policy detailed in this section. Based on the empirical results of this study, three strategic policy recommendations are proposed to guide the development of senior housing in a super-aged society:

• Standardizing “Smart Safety” with Privacy Integration: Since Fire Safety Compliance and Security & Privacy reached the highest consensus, the government should update building codes to mandate non-contact sensing technologies (e.g., Wave radar) as safety standards. This reduces the burden of manual inspections while addressing the profound need for “Dignified Monitoring” [27]. In this context, the retention of AI-driven ‘Active Sensing’ (Gⁱ = 8.08) represents a strategic shift from labor-intensive monitoring to technical efficiency, which directly addresses the high-priority ‘Staff-to-Resident Ratio’ (Gⁱ = 8.18) identified by the experts. By reframing these costs as ‘preventive resilience,’ policymakers can justify the initial expenditure as a means to avoid the extreme social and legal costs associated with elderly accidents. Furthermore, domestic manufacturers should be incentivized to develop AI-integrated interactive designs that foster intergenerational inclusion and personalized preventive care [28].
• Establishing Financial Trusts for Care Continuity: To mitigate high Uncertainty Avoidance, policy should foster a “Trust-Service” ecosystem. By integrating elder-care trusts with long-term care insurance and clear refund mechanisms, the psychological and economic barriers to entry can be lowered. This ensures that “Zoning Management” (moving from independent to assisted living) is supported by a stable financial lifecycle, reducing future pressures on public welfare expenditure [29].
• Embedding SCRCs within the 15-Minute City Framework: Urban planning must move beyond “Elderly Islands” by utilizing floor-area incentives to encourage mixed-use development. Embedding SCRCs within the existing social fabric—utilizing transport networks and green spaces—promotes intergenerational mutual support. Drawing from Japan’s experience, localized, small-scale, community-embedded care networks can transform senior housing into a driver of regional development and social vitality [30].

5.2.2. Future Research Directions

The findings of this study open several critical avenues for future academic inquiry within the evolving SCRC framework:

• Ethical and Functional Frontiers of Ambient Intelligence: Future research should investigate the long-term socio-psychological impact of transitioning from camera-based surveillance to ambient intelligence (e.g., sensors embedded in architectural surfaces), specifically exploring how Generative AI and Large Language Models (LLMs) can synthesize behavioral data into proactive emotional support [31]. Specifically, empirical studies are needed to determine the threshold where “unobtrusive monitoring” balances resident safety with the right to privacy.
• Socio-Behavioral Dynamics of Intergenerational Co-living: While this study highlights a cultural preference for intergenerational inclusion, more rigorous investigation is required to evaluate the success factors of de-institutionalized co-living models, such as senior–student rent-exchange programs. Research should focus on how shared communal values can be measured and maintained across diverse age cohorts to combat the “loneliness epidemic”.
• Service Innovations for ‘Solo Agers’: With the rise in single-person elderly households, future studies should analyze the effectiveness of “Surrogate Family Services”. This includes examining the integration of digital care coordinators with legal and financial safeguards, such as reverse mortgages, to provide a holistic safety net for those without traditional familial support.
• Psychological Impact of Wellness-Centered Environment Design: There is a significant research gap regarding the quantitative assessment of how “wellness-centered software”—such as sensory-friendly environmental design—affects the cognitive health of residents compared to traditional physical hardware (e.g., ramps and handrails).
• Resilience under Environmental Stress: As climate change intensifies, future research must address the adaptability of modular SCRC housing against extreme weather, incorporating climate-adaptive thermal control and decentralized power strategies to safeguard disabled populations [32]. Investigations should prioritize the development of backup thermal-control systems and medical equipment power strategies for disabled populations within high-density urban settings.

6. Limitations and Future Research

A primary limitation is that the identified link between social resilience and social capital represents a theoretical inference derived from expert consensus via the FDM, not a measured empirical outcome. While the results reflect the high perceived importance among stakeholders, future research is required to quantitatively test these theoretical links through longitudinal resident surveys or causal modeling. While this study systematically identifies the key factors of cluster living models through a multidimensional lens, several limitations must be acknowledged to contextualize the findings:

• Methodological Subjectivity: Although the FDM is highly effective for aggregating expert consensus and mitigating ambiguity, the inherent subjectivity of expert judgment may influence the final interpretation. Future research should incorporate quantitative frameworks, such as Multi-Criteria Decision Analysis (MCDA) or the Analytic Hierarchy Process (AHP), to further validate indicator weightings and enhance the robustness of the model.
• Geographic and Cultural Scope: This analysis focused primarily on the socioeconomic and regulatory landscape of Taiwan. Given the significant impact of cultural dimensions identified in this study, cross-cultural comparative research is necessary to examine how different global sociocultural contexts—such as those in Europe or North America—shape elderly residential preferences and the “translation” of the CCRC model.
• Absence of User-Centric Data: The current framework relies on the “Top-Down” perspectives of experts (practitioners, academics, and officials). Future studies should incorporate the “Bottom-Up” lived experiences and satisfaction levels of actual residents through longitudinal surveys or in-depth interviews. This would provide a more holistic evaluation of how “Social Resilience” is perceived by the end-users.
• Urban–Rural Disparity: Taiwan exhibits substantial disparities in healthcare infrastructure and social capital between metropolitan and rural areas. Further research is required to investigate how residential choices and service demands vary across these settings to ensure spatial equity in elderly care distribution.
• Technological Ethics and Evolution: As digital health and Ambient Assisted Living (AAL) technologies evolve rapidly, future studies must assess not only their functional effectiveness but also their ethical implications, particularly regarding data privacy and the potential for “technological isolation.”
• Optimization of the Care Continuum: Finally, as the SCRC model matures, further investigation is needed on how to optimize the physical and operational transitions between Independent Living (IL) and Assisted Living (AL). Research should focus on minimizing the “relocation trauma” often associated with declining health status within integrated residential models.

Based on the findings, we propose three methodological guidelines for future inquiry: (1) Transitioning from expert consensus to multi-criteria weight validation (e.g., AHP); (2) Integrating ‘Bottom-Up’ satisfaction metrics from diverse urban-rural cohorts; and (3) Conducting longitudinal cross-cultural comparisons to test the universal applicability of the SCRC framework in other East Asian or Western contexts.