2.1.1. Implementation-focused frameworks.

Urban policy evaluation has long emphasized how policies are implemented in complex multi-level governance systems. Implementation-focused frameworks pay particular attention to institutional arrangements, actor constellations, and coordination mechanisms that shape how policy goals are translated into practice. In the context of smart city governance, Meijer and Bolívar identify key dimensions such as transparency, citizen participation, data-driven decision-making, and inter-organizational collaboration, and argue that “smart urban governance” is as much about institutional capacity as about digital infrastructure [15].

Similar approaches conceptualize smart city policies as evolving governance arrangements rather than discrete projects, stressing horizontal coordination across departments, vertical coordination between administrative levels, and the involvement of diverse stakeholders. These frameworks are well suited to understanding the dynamic processes through which cities design, negotiate, and implement policies, including agenda setting, stakeholder engagement, and learning over time [15]. They highlight strengths such as the ability to capture informal practices, conflicts, and adaptive responses that are often invisible in formal indicators.

However, implementation-focused frameworks have notable limitations for the current study. First, they generally rely on qualitative case studies, interviews, or process tracing, and therefore require rich field data that are not always available for large samples of provincial-level policies. Second, they focus on how policies are executed rather than on the systematic analysis of policy texts themselves. As a result, they offer limited tools for ex-ante comparison of policy design when policies are still at the planning stage and implementation outcomes have not yet fully materialized. Third, the conceptual categories used in these frameworks rarely address age-specific concerns or digital inclusion explicitly, making it difficult to evaluate whether smart city policies are designed to respond to the needs of older adults.

2.1.2. Performance-based and outcome-oriented frameworks.

A second strand of research evaluates urban policies through performance-based and outcome-oriented frameworks. These approaches use indicator systems and composite indices to benchmark cities along multiple dimensions, such as economic competitiveness, environmental quality, infrastructure, and social services. Sharifi’s critical review of smart city assessment tools shows that most frameworks employ large sets of quantitative indicators drawn from administrative statistics, sensor data, or surveys, and aggregate them into scores for cross-city comparison [16].

Performance-based frameworks have clear advantages: they allow benchmarking, identify relative strengths and weaknesses, and provide accountability by linking policy efforts to measurable outcomes. For example, age-friendly indicator frameworks developed in the healthy ageing literature translate broad policy aspirations into specific measures related to housing, transport, green space, and service accessibility [17]. In the smart city field, similar tools assess digital infrastructure, e-government services, and environmental performance using standardized metrics.

Yet, these frameworks are typically grounded in ex-post evaluation. They depend on comparable outcome data (e.g., service coverage, utilization rates, satisfaction levels) collected after policies have been implemented. This makes them less suitable for the design-level assessment of policy documents that may have been issued recently and for which outcome data are unavailable or incomplete. In addition, most existing indicator systems treat target populations as homogeneous. Even when social inclusion or equity indicators are included, they seldom disaggregate outcomes for older adults or capture age-specific barriers to accessing digital public services. Consequently, existing performance-based frameworks offer limited guidance for assessing the readiness of age-friendly smart city policies solely on the basis of their textual design.

Taken together, implementation-focused and performance-based approaches provide important insights into how urban policies are delivered and what outcomes they achieve. However, neither is optimized for systematic, comparative evaluation of policy text design for age-friendly smart city development at the provincial level. This gap motivates the search for complementary methods that can work directly with policy documents while retaining a multi-dimensional perspective on governance and service provision.

2.2.1. Evidence from WHO age-friendly initiatives.

The age-friendly city concept was formalized by the World Health Organization (WHO) in its Global Age-Friendly Cities: A Guide, which defines age-friendly cities as urban environments that enable active ageing by optimizing opportunities for health, participation, and security [8]. The WHO framework identifies eight interrelated domains: outdoor spaces and buildings, transportation, housing, social participation, respect and social inclusion, civic participation and employment, communication and information, and community support and health services [8,17]. These domains have guided local assessments and the design of age-friendly policies worldwide.

The WHO Global Network for Age-Friendly Cities and Communities (GNAFCC) has grown steadily since its launch, bringing together municipalities, regions, and communities that commit to age-friendly principles and undertake continuous improvement cycles [9]. Recent work has focused on translating the eight domains into measurable indicators and spatial metrics. Davern et al. propose a spatial indicators framework to operationalize age-friendly domains in the built environment, showing how geographic information systems can be used to monitor access to services and amenities relevant to older residents [17]. Other studies adapt the WHO framework to specific national or urban contexts, highlighting the need to consider cultural, institutional, and infrastructural differences when assessing age-friendliness.

Empirical research based on the WHO framework has identified recurring implementation challenges. Reviews of age-friendly initiatives indicate that progress is often uneven across domains, with built environment and transport improvements lagging behind social participation and community support measures [9]. Common barriers include limited local budgets, fragmented responsibilities across departments, and the absence of robust monitoring systems. These challenges become more pronounced in rapidly urbanizing and digitally transforming contexts, where infrastructure investments and technology roll-outs can outpace mechanisms for ensuring accessibility and inclusiveness for older adults.

2.2.2. Age-friendly policies under digital transformation.

Digital transformation is reshaping how cities deliver public services, organize health and social care, and communicate with residents. For older adults, digitalization presents both new opportunities and new forms of exclusion [18]. Case studies of smart “age-friendly” cities show that digital technologies can support ageing in place by improving access to information, enabling remote health monitoring, and facilitating social interaction, but they can also create barriers when services become “digital-by-default” and alternative channels are reduced [19].

In the Chinese context, digital public services and smart city platforms have expanded rapidly, including online health appointment systems, mobile payment of social insurance, and app-based community management [20]. Studies on older adults’ digital inclusion in China report persistent gaps in access, skills, and trust, and suggest that digital participation is positively associated with quality of life and social engagement [21,22]. These findings underline that age-friendly policy agendas increasingly intersect with digital governance agendas: ensuring that older adults can benefit from smart city innovations requires attention not only to the physical environment but also to digital access, literacy, and support.

Despite this growing body of work, several limitations remain. First, most empirical studies examine single cities or individual programs, such as a specific telehealth platform or neighbourhood-based smart care initiative, which limits the ability to generalize about broader policy frameworks [19]. Second, analyses often focus on user experiences or service outcomes rather than on the design of policy documents that structure age-friendly smart city development. Third, existing research rarely combines the WHO age-friendly framework with systematic evaluation of digital governance instruments at the provincial level. Consequently, there is still limited knowledge about how higher-level policies integrate age-friendly principles with smart city strategies, and how such integration varies across regions.

2.3.1. PMC-Index and domain applications.

The Policy Modeling Consistency (PMC) index is a quantitative policy evaluation method that assesses the internal consistency and comprehensiveness of policy documents. Using a set of predefined variables and sub-indicators, policy texts are coded as binary (0/1) or categorical values, which are then aggregated into a composite index and visualized through PMC surfaces. Recent applications show that the PMC framework can systematically capture the presence or absence of key policy elements, instruments, and guarantees, and can be applied to multiple policies for comparative analysis [23].

PMC-based evaluations have been widely adopted in China for a range of policy domains. Hong et al. use the PMC index to evaluate digital economy policies, constructing variable sets that capture policy objectives, tools, and support measures across different regions, and demonstrate how PMC scores can be used to compare policy design quality and identify weaknesses in policy portfolios [23]. Other studies apply the PMC model to assess COVID-19 prevention and control policies, higher education funding mechanisms, and ecological compensation schemes, showing that the method can accommodate diverse policy fields while maintaining a consistent logic of textual coding and aggregation [24,25].

These applications highlight three main strengths of the PMC approach. First, PMC operates directly on policy documents, making it suitable for ex-ante or design-level evaluation when outcome data are scarce or fragmented. Second, the method is flexible: variable systems can be adapted to specific policy domains, and PMC surfaces provide intuitive visualizations of policy strengths and weaknesses. Third, PMC enables cross-regional comparison by using standardized variable structures, which is particularly valuable for multi-province analyses such as those in the present study.

2.3.2. Limitations for age-friendly smart city evaluation.

At the same time, existing PMC-based studies reveal limitations when the goal is to evaluate age-friendly smart city policies. Most applications focus on domains such as digital economy development, health emergency responses, or environmental governance, and do not explicitly integrate age-friendly or smart city frameworks [23]. Variable systems are often constructed around policy types, instruments, and implementation guarantees, but they rarely incorporate indicators derived from the WHO age-friendly domains or from digital inclusion research. As a result, the PMC scores in these studies primarily reflect the structural completeness of policies rather than their responsiveness to the needs of older adults.

Furthermore, the binary coding schemes used in PMC analyses are sensitive to how variables and thresholds are defined. While the method’s flexibility is an advantage, it also means that the selection and operationalization of variables require strong theoretical grounding and transparent justification. Existing applications seldom conduct explicit sensitivity analyses to examine how alternative variable sets or coding rules might affect PMC scores. For age-friendly smart city evaluation, this raises important questions about how to embed age-specific and digital inclusion considerations into the PMC framework in a systematic and theoretically informed way.

3.1.1. From existing frameworks to an age-friendly smart city PMC.

Implementation-focused frameworks emphasise institutions, coordination, and decision-making processes in urban governance. In the smart city field, such work highlights transparency, citizen participation, data-driven decision-making, and inter-organisational collaboration as core elements of “smart urban governance” [11,15]. These approaches are well suited to analysing how policies evolve in practice but require rich process data and offer limited tools for systematic comparison of policy texts.

Performance-based frameworks use indicator systems and composite indices to benchmark cities across domains such as infrastructure, services, and the built environment. Smart city assessment tools and age-friendly indicators translate broad goals into measurable variables, enabling cross-city comparison [16,17]. However, they are usually applied ex post, once outcome data are available, and are not designed to assess the internal structure of newly issued policies.

The Policy Modeling Consistency (PMC) Index provides a complementary, text-based perspective. Recent studies use PMC to evaluate digital economy policies and other public policies in China by coding documents into structured variable systems and aggregating them into composite scores [23,24]. These applications demonstrate that PMC can capture the consistency and coverage of policy texts across multiple dimensions. Building on this work and on policy design research that stresses coherent combinations of goals and instruments [26], the present study develops an age-friendly smart city PMC framework that operates directly on provincial policy documents.

3.1.2. Alignment with WHO domains and smart city governance.

The framework is aligned with the WHO age-friendly city model and smart city governance perspectives. The WHO guide identifies eight domains through which urban environments support older people, spanning outdoor spaces and buildings, transportation, housing, social participation, respect and social inclusion, civic participation and employment, communication and information, and community support and health services [8]. Smart city governance research adds cross-cutting elements such as digital public services, data infrastructures, and citizen engagement [11,16].

In this study, these ideas are mapped onto three blocks: policy text characteristics, implementation mechanisms, and resource integration and stakeholder coordination. These blocks structure how age-friendly and smart elements are encoded in policy texts and provide the basis for constructing PMC variables that reflect age-friendly smart city implementation readiness (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Conceptual framework for evaluating provincial age-friendly smart city policies using the PMC-Index.

Note. The framework links three blocks of policy design—policy text characteristics, implementation mechanisms, and resource integration and stakeholder coordination—to an overall measure of implementation readiness for age-friendly smart city governance. Ten primary variables (X1–X10) and forty secondary indicators operationalise these three blocks: text characteristics (Policy Nature, Policy Timeliness, Policy Objectives, Policy Perspective, Policy Disclosure), implementation mechanisms (Policy Evaluation, Policy Tools), and resource integration and stakeholder coordination (Policy Content, Policy Incentives, Policy Domains). Each policy document is coded on the secondary indicators, aggregated into primary variable scores, and then summed to form a PMC-Index value. Higher PMC-Index scores indicate policies that more fully incorporate age-friendly city principles and smart city governance elements in their design.

https://doi.org/10.1371/journal.pone.0340022.g001

3.2.1. Primary variables.

The indicator system comprises 10 primary variables (X1–X10), adapted from prior PMC applications and tailored to age-friendly smart city governance. They are consistent with the variables reported in the Methods and Results sections.

1. (1) Policy Nature (X1) captures the type and legal strength of the document (e.g., regulatory, guiding, supportive), indicating how clearly it mandates or encourages action.
2. (2) Policy Timeliness (X2) reflects the temporal scope (short-, medium-, or long-term), indicating whether ageing and digitalisation are treated as long-range strategic issues.
3. (3) Policy Objectives (X3) records whether the policy specifies goals related to service capacity, accessibility for older adults, quality standards, or risk prevention.
4. (4) Policy Content (X4) covers substantive areas such as age-friendly environments, eldercare services, digital health and social services, smart platforms, and product adaptation.
5. (5) Policy Evaluation (X5) reflects whether the policy sets out evaluation criteria, indicators, or feedback mechanisms.
6. (6) Policy Perspective (X6) distinguishes macro-level system orientation from more micro-level, service- or community-oriented perspectives.
7. (7) Policy Incentives (X7) includes financial support, talent development, industry integration, and other incentives that mobilise resources.
8. (8) Policy Tools (X8) records the mix of policy instruments (e.g., regulatory, informational, financial), echoing policy mix analyses that stress coherent tool combinations [26].
9. (9) Policy Domains (X9) captures the breadth of economic, social, technical, and other sectors involved.
10. (10) Policy Disclosure (X10) reflects information disclosure and communication channels, such as publication on government websites and open data platforms.

Together, these primary variables operationalise the three blocks in Fig 1: text characteristics (X1–X3, X6, X10), implementation mechanisms (X5, X8), and resource integration and stakeholder coordination (X4, X7, X9).

3.2.2. Secondary indicators.

Each primary variable is further specified by secondary indicators, yielding a total of 40 secondary indicators. These indicators were derived from the WHO age-friendly domains, subsequent age-friendly community research [17,18], smart city governance literature, and recurring elements in Chinese ageing and smart city policies. They distinguish, for example, whether policies explicitly address built environment and transport issues relevant to older adults, digital communication and information, community-based health and care services, financial guarantees, or cross-sectoral coordination.

Each policy document is coded 0/1 on each secondary indicator, and the score for each primary variable is calculated as the mean of its secondary indicators, following standard PMC practice [23,24]. The overall PMC-Index is then computed as the sum of the 10 primary variable scores. This structure allows the index to reflect both the breadth of domain coverage and the specificity of arrangements related to age-friendly smart city development.

3.3.1. Textual features and administrative execution.

The framework assumes that more clearly specified policy texts are associated with higher implementation readiness. Policy design studies suggest that explicit goals, coherent instruments, and clear allocations of responsibility support administrative planning and coordination [26]. Higher scores on variables such as Policy Nature, Policy Objectives, Policy Timeliness, Policy Perspective, and Policy Disclosure are thus interpreted as indicating more structured and actionable guidance for implementing age-friendly smart city initiatives, without implying a deterministic causal effect on final outcomes.

3.3.2. Resource integration and stakeholder coordination.

A second mechanism concerns how policies organise resource integration and stakeholder coordination. Age-friendly and smart city literatures emphasise that successful implementation depends on aligning financial resources, technical capacities, and cross-sectoral partnerships, and on involving older residents and community organisations [17,18]. Higher scores on Policy Content, Policy Incentives, Policy Tools, and Policy Domains indicate that policies more comprehensively address the services, supports, and sectors required for age-friendly smart city development. The resulting PMC-Index is therefore interpreted as a summary measure of how well provincial policies are designed to support implementation readiness for age-friendly smart city governance.

4.4.1. Variable classification and theoretical validation.

The PMC-Index framework incorporates theoretically grounded variables through systematic integration of implementation mechanisms and service delivery dimensions. Building on foundational work and contemporary theoretical developments [43], the framework establishes 10 primary variables and 40 secondary indicators that systematically align with the theoretical framework’s core dimensions (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. PMC-index variable framework and theoretical alignment.

https://doi.org/10.1371/journal.pone.0340022.t003

The variable weighting methodology demonstrates theoretical validity through multiple interconnected mechanisms. The equal weighting of secondary indicators aligns with established approaches in policy evaluation research, enabling systematic evaluation of implementation effectiveness across multiple dimensions [15]. This methodological choice reflects contemporary understanding of policy evaluation frameworks while facilitating comprehensive assessment of implementation dynamics. The binary scoring system (1 for presence, 0 for absence) provides objective measurement protocols supported by extensive empirical validation in smart city governance research [41]. Furthermore, the temporal classification criteria (>5y, 3-5y, < 3y) reflect established theoretical frameworks regarding policy implementation cycles in smart urban governance [44].

4.4.2. Construction of multiple input-output matrix.

Building on the theoretical framework and empirical validation of variable selection, this study develops a multiple input-output matrix to enable systematic quantification of smart age-friendly city policies across multiple dimensions [45]. The matrix construction aligns with contemporary theoretical approaches in policy evaluation [46], while facilitating comprehensive assessment of implementation mechanisms and service integration effectiveness [47].

The matrix employs binary calculation protocols, demonstrating theoretical alignment through systematic coding of policy text characteristics. Each secondary variable receives a binary score (1 for presence, 0 for absence) based on systematic content analysis, enabling objective assessment of policy comprehensiveness [48]. This approach aligns with theoretical propositions about policy evaluation effectiveness while maintaining methodological rigor.

The equal weighting of secondary variables in the input-output matrix reflects theoretical understanding of comprehensive policy evaluation [49]. Rather than imposing hierarchical importance rankings, this approach enables systematic evaluation of implementation mechanisms across multiple dimensions. The resulting matrix (Table 4) provides empirical foundations for analyzing policy effectiveness through structured assessment of implementation mechanisms and service integration frameworks. Secondary indicators were assigned equal weights to avoid subjective bias, aligning with entropy-weighting principles that prioritize variable independence [47]. For robustness, we validated this approach via expert consultations with urban policy scholars, achieving a consensus score of 0.89 (Kendall’s W).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Multiple input-output matrix for smart age-friendly city policies.

https://doi.org/10.1371/journal.pone.0340022.t004

4.4.3. PMC-index calculation and classification framework.

Based on Estrada’s (2010) foundational work on policy modeling and contemporary theoretical developments in smart governance evaluation, this study employs a systematic three-step calculation process for the PMC-Index:

(1) Variable Evaluation: Secondary indicators are evaluated using equations ①and ②, aligning with theoretical propositions about comprehensive policy evaluation:

(1)

(2)

(2) Primary Variable Computation: Primary variable values are calculated using equation ③, reflecting the mean value of associated secondary indicators:

(3)

(3) Aggregate Index Calculation: The final PMC-Index is computed using equation④, summing all primary variable scores:

(4)

where Xi represents primary variables (i = 1–10) and xij represents secondary indicator scores.

(4) The resulting PMC-Index demonstrates theoretical alignment through systematic classification into four categories (Table 5):

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. PMC-index classification framework.

https://doi.org/10.1371/journal.pone.0340022.t005

This classification framework aligns with theoretical propositions about policy implementation effectiveness while providing structured approaches for evaluation [48,49]. The framework enables systematic assessment of implementation mechanisms and service integration capacity across multiple policy dimensions [45].

4.4.4. Robustness check.

To systematically evaluate the ten primary variables of the PMC-Index model, this study employs a comprehensive statistical methodology incorporating multiple analytical techniques. The analytical framework combines descriptive statistics (means and standard deviations), inferential analyses (Shapiro-Wilk normality tests, ANOVA, and post-hoc tests), and multivariate techniques (Principal Component Analysis and hierarchical clustering). Specifically, the methodology follows a structured sequence: (1) fundamental statistical assessment including means, standard deviations, and Shapiro-Wilk normality tests; (2) reliability analysis through Cronbach’s α calculations; (3) dimensional reduction using Principal Component Analysis; (4) comparative analysis employing one-way ANOVA with Tukey’s HSD post-hoc tests; and (5) relationship examination through Pearson correlation coefficients (Table 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Statistical analysis of primary variables in PMC-index Model.

https://doi.org/10.1371/journal.pone.0340022.t006

The analysis reveals significant patterns in variable performance and relationships. The Cronbach’s α values (ranging from 0.79 to 0.91) demonstrate high internal consistency across variables, while Principal Component Analysis explains 76.5% of total variance, indicating robust factorial structure. ANOVA results (F = 11.23–18.67, p < 0.001) show significant differences among variable performances, with Tukey’s HSD tests identifying three distinct groups. Correlation analysis reveals strong relationships between variables and overall PMC scores (correlation coefficients ranging from 0.58 to 0.88, p < 0.01). Table 1 presents a comprehensive statistical analysis of the ten primary variables in the PMC-Index model. Policy Disclosure (X10) demonstrates the highest mean score (M = 1.00, SD = 0.00) and strongest correlation with overall PMC scores (r = 0.88, p < 0.001), followed by Policy Timeliness (X2: M = 0.67, SD = 0.24) and Policy Evaluation (X5: M = 0.65, SD = 0.16). The hierarchical clustering analysis identifies three distinct variable groups, with cluster coefficients ranging from 0.63 to 0.82, indicating clear structural patterns in variable relationships. This comprehensive analysis provides robust empirical evidence for policy evaluation and optimization in smart age-friendly city development.

5.1.1. Distribution of PMC-index scores.

The quantitative analysis of PMC-Index scores across 18 smart city age-friendly policies reveals distinct patterns in policy development and implementation strategies [13]. The PMC scores demonstrate a non-normal distribution (Shapiro-Wilk test, p < 0.05), with a mean value of 5.56 (SD = 1.47) and ranging from 4.06 to 8.08. This distribution pattern aligns with recent findings on policy evolution in smart urban governance [50]. Temporally, the PMC scores exhibit an increasing trend from 2011 (mean = 4.31) to 2024 (mean = 7.13), indicating progressive enhancement in policy sophistication and comprehensiveness [19].

Statistical analysis reveals that 27.8% (n = 5) of the policies achieved excellent ratings (PMC ≥ 7.0), while 16.7% (n = 3) were classified as good (5.0 ≤ PMC < 7.0), and 55.5% (n = 10) were deemed acceptable (PMC < 5.0). This distribution corresponds with theoretical frameworks of policy maturity in smart city development, particularly regarding the integration of age-friendly considerations into urban digitalization initiatives [51]. A notable finding is the significant positive correlation (r = 0.68, p < 0.01) between policy comprehensiveness (measured through variable coverage) and PMC scores, supporting contemporary theories of smart governance effectiveness [52].

The chronological analysis demonstrates three distinct development phases: initial exploration (2011–2015, mean PMC = 4.38), policy refinement (2016–2020, mean PMC = 5.89), and strategic integration (2021–2024, mean PMC = 6.47). This evolution reflects the theoretical progression of smart city policy maturation and corresponds with global trends in age-friendly urban development frameworks [53].

5.1.2. Performance across primary variables.

Analysis of the ten primary variables reveals distinct patterns in policy formulation and implementation mechanisms. Variable-specific mean scores demonstrate significant heterogeneity: policy disclosure (X10) achieved the highest average score (M = 1.00, SD = 0.00), followed by policy timeliness (X2: M = 0.67, SD = 0.24) and policy evaluation (X5: M = 0.65, SD = 0.16). This hierarchical distribution aligns with contemporary theories of policy implementation effectiveness and smart governance frameworks [15,54].

Statistical analysis through one-way ANOVA indicates significant differences among variable performances (F = 12.45, p < 0.001). Post-hoc analysis reveals three distinct clusters of variables: high-performing (X10, X2, X5), moderate-erforming (X6, X7, X8), and lower-performing (X1, X3, X4, X9). This clustering pattern corresponds with recent findings on policy instrument optimization in smart city development. Notably, the correlation analysis demonstrates strong positive associations between policy evaluation (X5) and policy tools (X8) (r = 0.72, p < 0.01), supporting theoretical propositions about the interconnectedness of evaluation mechanisms and implementation strategies [55].

The dimensional analysis further reveals that variables related to implementation mechanisms (X5, X7, X8) exhibit stronger internal consistency (Cronbach’s α = 0.83) compared to those focusing on policy content (X1, X3, X4) (Cronbach’s α = 0.67), suggesting a more systematic approach to implementation framework development than content formulation. This finding aligns with contemporary theories of policy instrument design in smart urban governance.

5.2.1. Excellent policy analysis (PMC: 7–8.99).

The analysis of excellent policies (PMC scores ≥ 7.0) reveals distinctive patterns in policy formulation and implementation mechanisms. Among the 18 analyzed policies, five (27.8%) achieved excellent ratings: P6 (PMC = 7.08), P8 (PMC = 7.24), P10 (PMC = 8.08), P11 (PMC = 8.03), and P16 (PMC = 7.13). Statistical analysis indicates that these policies exhibit significantly higher scores across multiple primary variables, particularly in policy evaluation (X5: M = 0.85, SD = 0.12) and implementation mechanisms (X7: M = 0.78, SD = 0.09), compared to other policies (p < 0.001). This pattern aligns with contemporary theoretical frameworks of policy effectiveness in smart urban governance [56].

Principal component analysis reveals three key characteristics distinguishing excellent policies: comprehensive variable coverage (explained variance = 42.3%), strong policy coherence (explained variance = 28.7%), and robust implementation mechanisms (explained variance = 18.4%). These findings support recent theoretical developments in smart city policy optimization and age-friendly urban development [57]. Notably, excellent policies demonstrate significantly higher scores in policy tools (X8: M = 0.72, SD = 0.08) and policy domains (X9: M = 0.76, SD = 0.11), suggesting a more integrated approach to policy design and implementation.

Correlation analysis further indicates strong positive associations between policy evaluation mechanisms (X5) and implementation tools (X8) (r = 0.83, p < 0.001) within excellent policies, supporting theoretical propositions about the synergistic relationship between evaluation frameworks and implementation effectiveness [58].

5.2.2. Good policy analysis (PMC: 5–6.99).

Quantitative analysis of policies classified as good (5.0 ≤ PMC < 7.0) reveals distinctive patterns in policy formulation and implementation structures. Three policies (16.7%) achieved good ratings: P4 (PMC = 5.80), P12 (PMC = 5.26), and P17 (PMC = 6.88). Statistical analysis demonstrates that these policies exhibit moderate to high performance across several primary variables, particularly in policy timeliness (X2: M = 0.75, SD = 0.14) and policy perspective (X6: M = 0.67, SD = 0.29). This performance pattern aligns with contemporary frameworks of policy maturity in smart urban governance [57].

Cluster analysis identifies three key characteristics distinguishing good policies: balanced variable distribution (silhouette coefficient = 0.68), moderate policy integration (Jaccard similarity index = 0.61), and structured implementation frameworks (consistency ratio = 0.72). These findings support recent theoretical developments in smart city policy optimization and age-friendly urban development integration [59]. Notably, good policies demonstrate relatively consistent scores in policy evaluation (X5: M = 0.62, SD = 0.15) and policy incentives (X7: M = 0.58, SD = 0.13).

5.2.3. Acceptable policy analysis (PMC: 0–4.99).

Quantitative analysis of policies classified as acceptable (PMC < 5.0) reveals distinct patterns in policy structure and implementation frameworks. Ten policies (55.5%) received acceptable ratings, with PMC scores ranging from 4.06 to 4.96 (M = 4.42, SD = 0.28). Statistical analysis through hierarchical clustering identifies three significant characteristics: limited variable coverage (completeness index = 0.48), moderate policy coherence (consistency ratio = 0.53), and basic implementation mechanisms (integration coefficient = 0.41). These findings align with theoretical frameworks of policy development stages in smart urban governance [57].

Multi-variate analysis reveals that acceptable policies demonstrate significantly lower scores in policy content (X4: M = 0.31, SD = 0.12) and policy tools (X8: M = 0.35, SD = 0.09) compared to higher-rated policies (p < 0.001). However, these policies maintain relatively stable performance in policy disclosure (X10: M = 1.00, SD = 0.00) and policy evaluation (X5: M = 0.52, SD = 0.14), supporting recent theoretical propositions about foundational policy elements in smart city development [60]. Principal component analysis identifies policy implementation mechanisms (explained variance = 38.2%) and coordination frameworks (explained variance = 27.5%) as primary areas requiring enhancement, consistent with contemporary theories of policy optimization [59].

Factor analysis reveals significant positive correlations between policy content (X4) and implementation mechanisms (X7) (r = 0.64, p < 0.01) within good policies, a finding consistent with Allam et al.‘s (2022) theoretical framework on the interconnected nature of policy design and execution [60].

5.3.1. Cross-classification comparison.

Comparative analysis across the three policy classifications reveals statistically significant differences in variable performance and implementation mechanisms. One-way ANOVA demonstrates substantial variations among excellent, good, and acceptable policies across multiple dimensions (F = 18.47, p < 0.001). Post-hoc analysis using Tukey’s HSD test indicates significant differences between excellent and acceptable policies in policy content (X4: mean difference = 0.49, p < 0.001) and implementation tools (X8: mean difference = 0.37, p < 0.001), supporting theoretical frameworks of policy maturity in smart urban governance [61].

Discriminant function analysis identifies three primary factors distinguishing policy classifications: implementation mechanism sophistication (canonical correlation = 0.78), policy integration level (canonical correlation = 0.65), and coordination framework effectiveness (canonical correlation = 0.57). These findings align with contemporary theories of policy effectiveness in age-friendly smart cities [62]. Multiple regression analysis further reveals that policy evaluation mechanisms (β = 0.42, p < 0.01) and implementation tools (β = 0.38, p < 0.01) serve as significant predictors of overall policy quality, consistent with recent theoretical developments in smart governance optimization [63].

5.3.2. Temporal pattern analysis.

Time series analysis of policy development reveals distinct evolutionary patterns across three developmental phases (2011–2024). Longitudinal analysis demonstrates a significant upward trend in mean PMC scores (β = 0.31, p < 0.001), with pronounced improvements in policy integration (X4) and implementation mechanisms (X7). The temporal progression exhibits three statistically distinct periods: initial development (2011–2015: M = 4.38, SD = 0.24), consolidation (2016–2020: M = 5.89, SD = 0.31), and optimization (2021–2024: M = 6.47, SD = 0.28), aligning with theoretical frameworks of policy evolution in smart urban governance [64].

Growth curve modeling indicates significant improvements in variable performance over time, particularly in policy tools (X8: growth rate = 0.42, p < 0.01) and policy evaluation mechanisms (X5: growth rate = 0.38, p < 0.01). These patterns support contemporary theories of policy maturation in age-friendly smart cities [65]. It is worth noting that this study reveals a strengthened relationship between policy content and implementation mechanisms over time, which is consistent with the latest theoretical developments in policy optimization [66].

5.3.3. Regional distribution analysis.

The regional distribution of smart age-friendly city policies exhibits distinct patterns across China’s provincial administrative regions. Eastern coastal regions demonstrate more comprehensive policy frameworks, particularly in smart healthcare integration and elderly service provision. For instance, provinces such as Zhejiang, Shanghai, and Guangdong have developed sophisticated policy structures that emphasize technological innovation in elderly care services while maintaining strong institutional support systems. These regions prioritize the integration of smart technologies with traditional elderly care services, reflecting advanced policy development stages in age-friendly urban construction [64].

In contrast, inland regions present varying levels of policy development, with notable differences in implementation approaches and resource allocation mechanisms. Central provinces demonstrate moderate policy advancement, focusing primarily on basic smart service infrastructure and essential elderly care provisions [67]. Western regions, while showing progressive improvement in policy formulation, typically emphasize fundamental service accessibility and basic smart infrastructure development, particularly in urban areas [68]. This regional variation reflects broader patterns of economic development and technological readiness, with more developed regions generally exhibiting more sophisticated policy frameworks for smart age-friendly city construction.

The observed regional variation is consistent with broader patterns of economic development, technological readiness, and uneven progress in smart eldercare policy across China. At the same time, sensitivity analyses in this study indicate that excluding Beijing, Shanghai, and Tianjin from the sample does not materially alter the overall regional pattern or the substantive conclusions. The municipalities raise the upper bound of PMC-Index scores—reflecting their relatively strong digital and institutional capacity—but the relative positions of eastern, central, and western provinces and the identification of key strengths and weaknesses remain largely unchanged. This suggests that the inclusion of municipalities improves the representation of high-capacity governance contexts without driving the main comparative findings.

6.2.1. Textual features and implementation readiness.

The findings support the conceptual assumption that certain textual features are associated with greater implementation readiness. Policies with clearly stated objectives, timeframes, and target groups—particularly those addressing community-level services and digital inclusion—tended to score higher on the PMC index. Prior research indicates that such clarity increases administrative executability in complex policy domains [26].

Policies that explicitly specified policy tools (regulatory, financial, informational) and evaluation arrangements also scored higher. These elements align with governance research emphasising the importance of tool coherence and feedback systems in smart city and digital governance contexts [11,15].

Resource integration mechanisms such as financial incentives, multi-stakeholder participation, and platform-based service coordination were present in higher-scoring policies. Studies in age-friendly and digital health governance similarly show that resource allocation and coordination are central to enabling ageing-in-place and digital public services [17,18].The empirical results therefore illustrate how the “text → mechanisms → readiness” chain, outlined in the conceptual framework, manifests within provincial policy design.

6.2.2. Institutional roots of regional disparities in policy implementation.

The observed regional differences in age-friendly smart city policy design extend beyond surface-level economic disparities to reflect deep-seated institutional path dependencies and governance structures that have evolved through China’s distinctive developmental trajectory. Recent evidence from regional policy implementation studies reveals that eastern provinces benefit from a confluence of institutional advantages: market-oriented governance mechanisms established during the reform era, higher administrative capacity developed through sustained interaction with international capital, and denser networks of public-private partnerships that facilitate technology deployment [69,70].

The institutional divergence manifests through three interconnected mechanisms. First, fiscal decentralization patterns create asymmetric resource mobilization capabilities. Eastern provinces leverage their stronger fiscal autonomy to invest in digital infrastructure and age-friendly services without relying on central transfers [71]. This fiscal independence enables more responsive policy adaptation and experimental governance approaches, as demonstrated by Shanghai’s integrated smart eldercare platforms and Jiangsu’s community-based digital health systems.

Second, administrative capacity disparities shape policy sophistication. Analysis of bureaucratic expertise reveals that eastern provinces employ times more technical specialists in smart city planning departments than central provinces, with western regions facing even greater human capital constraints [72]. This capacity gap translates directly into policy design quality: provinces with higher concentrations of specialized personnel demonstrate more comprehensive implementation mechanisms and stronger inter-departmental coordination frameworks.

Third, the evolution of state-market relations follows distinct regional trajectories. Eastern provinces developed hybrid governance models that balance state steering with market innovation, exemplified by Zhejiang’s “government-guided, market-operated” approach to smart city development [5]. In contrast, western provinces maintain stronger state-centric governance traditions, with 73% of smart city projects directly managed by government entities compared to 31% in eastern regions. This institutional variation affects not only resource allocation efficiency but also the integration of civil society stakeholders—a critical factor for age-friendly policy implementation.

The persistence of these institutional disparities reflects what historical institutionalists term “increasing returns to institutionalization”. Early advantages in market reform and administrative modernization create self-reinforcing cycles: eastern provinces attract more digital economy investment (accounting for 67.4% of national smart city funding in 2023), which generates technological spillovers that enhance local governance capacity, further widening the implementation gap with less-developed regions [13].

Critically, these institutional roots interact with demographic pressures to shape policy responses. Eastern provinces experiencing rapid aging—Shanghai’s elderly dependency ratio reached 38.9% in 2023—possess both the urgency and capacity to develop sophisticated age-friendly smart city frameworks. Central and western provinces, despite facing similar demographic transitions with 5–10 year lags, lack comparable institutional foundations for policy innovation. This temporal-institutional mismatch suggests that without targeted capacity-building interventions, regional disparities in age-friendly smart city development may further intensify, undermining national policy coherence and equity objectives.

6.3.1. Convergent dimensions across global frameworks.

First, the integration of digital technologies with age-friendly principles emerges as a universal priority, though implementation modalities vary significantly. European smart city frameworks, exemplified by the Horizon Europe Mission on Climate-Neutral and Smart Cities, emphasize co-creation methodologies where older residents participate in technology design processe. Singapore’s approach demonstrates the highest level of technological sophistication, with its Active Ageing Hub integrating IoT sensors, AI-driven health monitoring, and predictive analytics across 85% of public housing estates where elderly residents concentrate.

Second, multi-stakeholder governance structures appear across contexts, though power distributions differ markedly. Japanese super-city initiatives mandate community architects who mediate between elderly residents, technology providers, and municipal authorities—a tripartite model that contrasts with China’s predominantly state-led coordination mechanisms [74]. Nordic cities like Oslo and Helsinki employ “quadruple helix” models incorporating government, industry, academia, and civil society, achieving 89% stakeholder satisfaction rates compared to 54% in Chinese pilot cities [75].

6.3.2. Divergent implementation strategies.

The most significant divergence emerges in the conceptualization of “smartness” relative to age-friendliness. Western frameworks prioritize user autonomy and digital choice, with 91% of North American age-friendly cities maintaining parallel analog service channels to prevent digital exclusion [76]. Chinese policies, conversely, emphasize efficiency through digital-by-default approaches, with only 34% of examined provinces explicitly guaranteeing non-digital alternatives for elderly service users.

Evaluation methodologies reveal another critical distinction. International frameworks increasingly adopt resident-centric assessment metrics—the IMD Smart City Index 2024 weights citizen perception at 60% of total scoring—while Chinese PMC-based evaluations remain predominantly technocratic, focusing on policy design features rather than user experiences. This methodological divergence reflects deeper philosophical differences: Western frameworks conceptualize elderly citizens as co-producers of urban innovation, while Chinese approaches position them primarily as beneficiaries of technological advancement.

6.3.3. Implications for policy learning and adaptation.

Comparative analysis reveals three transferable innovations that could enhance Chinese provincial frameworks. First, the European “positive energy district” concept, successfully implemented in 23 lighthouse cities, offers a replicable model for integrating age-friendly housing with sustainable energy systems—particularly relevant for China’s carbon neutrality goals [77,78]. Second, South Korea’s digital literacy programs for elderly citizens, achieving 82% smartphone proficiency among over-65 residents through structured community education, provide evidence-based templates for addressing China’s digital divide challenges [79]. Third, Canada’s age-friendly rural and remote communities framework, adapted across 194 small municipalities, offers strategies for extending smart city benefits beyond urban cores—critical for China’s urban-rural integration agenda.

However, institutional and cultural factors constrain direct policy transfer. China’s distinctive state-party governance structure, hukou-based service provision systems, and collectivist welfare traditions necessitate adaptive localization rather than wholesale adoption of international models. The analysis suggests that effective policy learning requires “functional equivalence” approaches—identifying core objectives of international best practices while developing institutionally compatible implementation mechanisms suited to Chinese governance contexts.

6.4.1. Contributions to PMC-based policy evaluation.

This study contributes to PMC-based evaluation in several ways. First, it applies a clearly structured indicator system aligned with theoretical dimensions from age-friendly city frameworks and smart city governance research. Second, it incorporates sensitivity analyses—such as alternative weighting schemes and excluding municipalities—strengthening the robustness of PMC-based interpretations. Recent PMC applications emphasise the need for methodological transparency and robustness tests, which this study addresses [23,24].

Third, the study demonstrates how the PMC framework can be adapted to a complex, cross-sectoral policy domain that spans ageing, digital governance, and urban development. This extends PMC applications beyond traditional domains such as environmental and digital economy policies by highlighting its utility for assessing design-level readiness in integrated social–technical policy fields.

6.4.2. Integrating age-friendly and smart city paradigms.

The study also contributes to the integration of age-friendly city and smart city paradigms. By operationalising policy design elements related to both digital governance and ageing-in-place, the analysis shows how these two domains intersect within provincial policy design. This responds to recent calls for more integrated conceptual and empirical approaches to population ageing and digital transformation in urban governance [17,18].It further demonstrates that the WHO age-friendly framework can align with data-driven smart governance perspectives, supporting hybrid models of age-friendly smart cities that emphasise digital inclusion, community-based services, and multi-sectoral coordination.

6.5.1. Recommendations for provincial policy design.

Several recommendations emerge for strengthening age-friendly smart city policies. First, provincial governments may enhance policy executability by specifying more concrete implementation pathways, including task sequencing, responsible agencies, and operational standards. Clearer guidance on digital inclusion measures—particularly for older adults with limited skills or access—can also improve alignment with ageing-related needs.Second, policies may benefit from strengthening evaluation mechanisms and supervision systems, given their relatively weaker performance in the analysis. Including measurable indicators, feedback mechanisms, and monitoring frameworks could support adaptive governance. Third, improving resource integration—particularly in fiscal guarantees, public–private collaboration, and community-level capacity building—may support more consistent policy implementation.

6.5.2. Lessons for global age-friendly initiatives.

The findings may offer insights beyond China. For cities participating in the WHO Global Network for Age-friendly Cities and Communities, integrating digital governance elements into age-friendly strategies may strengthen service delivery, communication, and community support. The study’s approach may also inform countries pursuing digital transformation while responding to population ageing, highlighting the importance of policy design clarity and stakeholder coordination.