In 2015, the State Council issued the Guiding Opinions on Actively Promoting the “Internet + Initiative,” explicitly proposing to leverage the internet's role as an innovation driver and promote its deep integration with all economic and social sectors, including healthcare. This laid the policy foundation for applying internet technology to chronic disease prevention and control. In 2017, the State Council further formulated the “China Chronic Disease Prevention and Control Medium- and Long-Term Plan (2017–2025),” emphasizing the importance of chronic disease prevention and control. Guided by the two core policies, this study systematically identifies national policy documents published from 2015 to 2025 by searching the Peking University Law Database and Chinese government websites using the core query: (“chronic disease” OR “chronic diseases”) AND (“Internet +” OR “smart healthcare” OR “digital health”). The search results were screened in strict accordance with the following criteria: clearly excluding provincial and below local policies, and only including national level planning, opinions and other normative documents issued by the State Council and its ministries and commissions, which are highly relevant to the theme, to ensure that the research focuses on the national top-level design and ensure the comparability of the samples, policy retrieval process diagram (Figure 1).

To ensure the objectivity of the scoring, two researchers independently evaluated 18 policies using PMC indicators, and calculated the Cohen's kappa coefficient to be 0.727. According to Landis and Koch's criteria (44), this indicates a high degree of consistency among raters.

For policy text analysis, this study employed ROST CM 6 and Gephi. The process began with the ROST CM 6 tool processing the 18 policy texts to perform word frequency statistics. Subsequently, a co-occurrence matrix was constructed based on the top 100 high-frequency words, applying a co-occurrence threshold of ≥3 to filter out generic administrative terms and retain associations between core concepts. This matrix was then imported into Gephi for semantic network analysis and thematic clustering. In Gephi, an edge weight filter of ≥5 was applied. For network visualization, the Fruchterman Reingold layout algorithm (with parameters set to Speed = 1 and Gravity = 12) was employed, with its parameters optimized to produce a clear representation.

The Polynomial Modeling and Comparison (PMC) Index model, originally developed by Estrada, focuses on quantifying policy internal consistency and implementation effectiveness through a systematic evaluation framework. This model first decomposes policy objectives into quantifiable key variables and establishes corresponding evaluation indicators for each variable, thereby forming a comprehensive policy assessment system. This study employs the PMC Index model to conduct a quantitative evaluation of chronic disease management policy texts. The PMC index model construction process comprises four main steps:

Assign binary values to secondary variables using (Equations 1, 2).

Then, this study referenced existing policy evaluation grading standards (42) to classify policy quality into five tiers: unacceptable, acceptable, good, excellent, and perfect. The criteria for tier classification are as follows in Table 1: when the PMC index falls between 9 and 10, policy consistency is highest, classified as “perfect”; scores between 8 and 8.99 are rated “excellent”; scores between 6 and 7.99 are classified as “good”; scores between 4 and 5.99 indicate weaker policy coherence and are rated “acceptable”; and scores between 0 and 3.99 denote very weak policy coherence, rated “poor.”

Through systematic screening, 18 policy documents related to smart healthcare for chronic disease prevention and control were collected and subjected to quantitative analysis and evaluation using the PMC index model. Detailed information including policy release years, issuing authorities, issue numbers, and policy titles is presented in Table 2.

The word frequency analysis reflects keywords frequently mentioned by policymakers. High-frequency terms extracted from policy texts facilitate deeper understanding of the priorities and objectives outlined in chronic disease management policies. By removing irrelevant and generic terms from this study, high-frequency words were extracted for word frequency analysis (Table 3), and visually analyze the relevant high-frequency words (Figure 2).

The co-occurrence network clustering diagram of high-frequency words serves as a visual tool in text analysis, intuitively presenting lexical associations and thematic structures. It represents high-frequency words in the text as “nodes,” uses “lines” to indicate co-occurrence relationships between words, and forms “clusters” from closely connected nodes, which correspond to the core themes of the text. Words within the same cluster are highly semantically related due to frequent co-occurrence, collectively pointing to a specific theme. Overall, this helps researchers quickly identify core discussion directions and intrinsic logical connections from massive amounts of text. From the extracted high-frequency words, we obtained a co-occurrence matrix vocabulary. Based on this, we generated a network co-occurrence diagram for high-frequency words in Figure 3.

This framework comprises 10 primary variables and 48 secondary variables. Primary variables include: Type of policy (X1), Policy timeliness (X2), Policy areas (X3), Policy tool (X4), Policy focus (X5), Policy level (X6), Policy theme (X7), Policy guarantee (X8), Policy evaluation (X9), and Policy disclosure (X10). Each primary variable is subdivided into several secondary variables to ensure comprehensive inclusion of relevant variables without omitting potential independent variables. All secondary variables are equally weighted and assigned binary values (involved = 1, not involved = 0). The complete evaluation indicator system is presented in Table 4.

The word frequency and thematic clustering analysis of 18 intelligent chronic disease management policy texts not only systematically reveals the current policy priority direction, but also highlights the potential tension between macro narrative and micro implementation of policies, providing empirical evidence for understanding the internal logic and optimization path of policy frameworks in this field. Word frequency analysis shows that “health,” “service,” “Internet,” “data,” “construction,” “management,” “promotion” and other words constitute the core of the policy text. Among them, the high-frequency co-occurrence of “health” and “service” clarifies the value orientation of policies with the fundamental goal of improving the health level of the entire population and optimizing service supply as the core path, which is consistent with the core concept of “putting people's health at the center” in the “Healthy China 2030” strategy. “Internet” and “data” have been repeatedly emphasized as key enabling tools, which confirms that digital technology is regarded as the core engine driving the innovation of chronic disease management model. At the same time, the prominent use of dynamic terms such as “development,” “promotion,” and “construction” collectively outlines a proactive policy framework that emphasizes technology driven, system building, and continuous improvement.

Further theme clustering extracts five core themes from the semantic content of the text, including “Empowering with Intelligent Technologies and Fostering Industrial Innovation,” “Chronic Disease Prevention and Control with Health Service System Development,” “Information Services and System Mechanism Development,” “Resource Optimization and Service Capacity Enhancement”, and “Policy Implementation and Execution.” These themes are interrelated, forming a complete policy narrative chain from “top-level strategic design (technology empowerment, system integration)” to “middle-level support construction (information mechanism, resource optimization)”, and then touching on “grassroots implementation.” This structure clearly indicates that the current policy design has distinct systematic and comprehensive characteristics, aiming to promote smart chronic disease management through multi-dimensional collaboration.

However, meticulous text analysis also exposes the potential disconnect between policy intentions and actual implementation environments, pointing to key optimization spaces: the frequent emergence of the term “construction” and the theme of “information services and institutional mechanism construction,” collectively indicating a high tilt of policy resources toward infrastructure such as information platforms and data systems. In contrast, statements related to service evaluation, quality supervision, and multi-party governance are relatively weak. This implies that there may be a risk of policies that prioritize upfront investment over sustainable operation and efficiency governance, which could lead to difficulties in converting construction achievements into sustainable service capabilities. Although verbs such as “promote” and “enhance” clearly indicate policy direction, there is insufficient detailed description of the specific “implementation process,” “division of responsibilities,” and “collaborative mechanism.” As an independent theme, the text intensity and specificity of “policy implementation and execution” are significantly inferior to other targeted themes. This reflects the gap between setting grand visions and providing clear and actionable implementation roadmaps in policies, which can lead to inconsistent understanding and difficulty in coordination among the executive level, resulting in implementation gaps. Although concepts such as “artificial intelligence” and “big data” are widely embedded in various topics, there is a lack of in-depth discussions specifically addressing governance issues such as “algorithm ethics,” “data security,” “privacy protection,” and “digital inclusion.” This indicates that in the process of promoting the integration and application of technology, the corresponding adaptive governance framework and ethical norms have not been deepened synchronously, which may become a potential bottleneck for future technology implementation and public trust.

Examining the above findings within the global digital health governance landscape reveals both consensus and distinctive features. China's policy framework emphasizes the use of data to enhance service accessibility and governance efficiency, which is highly consistent with the World Health Organization's (WHO) Global Digital Health Strategy (2020–2025) and the core principles of digital health governance advocated by the Organization for Economic Cooperation and Development (OECD) (49, 50). However, China's path shows a unique top-level design feature of “national leadership and systematic promotion.” It deeply integrates chronic disease management into the overall situation of medical reform through national strategies such as “Internet plus+health care,” and its policy is particularly holistic and coordinated. Compared with more focused strategies such as Germany (focusing on data privacy and security frameworks) or Tanzania (focusing on community empowerment and basic access) (51), China's current policies tend to systematically address broad challenges such as “inclusive services” and “basic empowerment.” This comprehensive path has advantages in rapid layout and the formation of economies of scale, but it also puts higher demands on the precision of policies, the meticulousness of implementation, and the depth of special governance.

Therefore, the policy layout revealed in this study not only confirms China's practical path of “technology as the support and service as the purpose,” but also points out the future optimization direction-while maintaining the advantages of top-level design, we need to focus on strengthening the implementation of monitoring and multiple co treatment mechanisms, so as to build a more flexible and inclusive chronic disease treatment system.

Through Pearson correlation analysis of policy years and PMC index, the results showed no significant correlation between the two. This indicates that the average PMC index of policies does not show a significant linear time trend, and the index fluctuates greatly in each year, reflecting that under this evaluation system, the consistency of policy quality in different years still needs to be strengthened, and there is still room for improvement in the scientific and normative design of policies. This phenomenon may be due to the different dimensions of policy emphasis in different years (such as technology application, service processes, guarantee mechanisms, etc.), resulting in fluctuations in PMC scores without forming a sustained upward trend.

The comprehensive performance analysis of the policy text shows that there are obvious systematic differences in the scores of each dimension, which deeply reflects the internal logic and focus of the current intelligent chronic disease management policy system. The dimensions of policy tool (X4 = 0.96), policy theme (X7 = 0.91), policy guarantee (X8 = 0.81), and policy evaluation (X9 = 1.00) generally received high scores, jointly outlining a standardized, comprehensive, and multi tool macro policy framework, reflecting the strong systematic construction ability at the top-level design level. In contrast, the scores of the two dimensions of policy focus (X5 = 0.58) and policy timeliness (X2 = 0.61) are significantly lower, constituting a shortcoming in the overall evaluation. The former indicates that while the policy text covers a wide range of fields, the prioritization of core challenges and the in-depth explanation of key measures are insufficient; The latter reveals that the connection and dynamic adjustment mechanism between short-term actions, medium-term planning, and long-term strategies is still unclear.

The clustering distribution of high and low score dimensions is not accidental. It points to a noteworthy tendency in the policy-making process: investing more in pursuing completeness of form and breadth of content, while relatively weak in focusing on strategic depth and dynamic coordination of planning. High scoring dimensions tend to focus on describing what elements policies should include and how to provide support, reflecting strong supply side planning thinking; the low score dimension is precisely related to how policies determine key breakthroughs in complex and changing environments, and maintain resilience and adaptability over time. These are the key process factors that affect the ultimate effectiveness of implementation.

Therefore, the overall pattern of performance scores reveals that the advantage of existing policy texts lies in constructing a standardized and widely covered macro framework, laying a solid foundation for the development of intelligent chronic disease management. However, in order to translate the policy blueprint into solid results, future optimization directions need to focus on addressing existing shortcomings, shifting from emphasizing the comprehensiveness of the framework to simultaneously enhancing strategic focus, coordination of action sequences, and flexibility for cross period adaptation. This requires policy formulation not only to focus on the completeness of static content, but also to embed dynamic process management thinking, thereby promoting the entire system to move from having a good foundation to achieving precise and efficient operation.

Among these policies, P14, P2, P17, P4, and P15 were rated “excellent,” reflecting common characteristics of high-quality policies. First, they are underpinned by top-level strategic support. P2 and P4 serve as national-level guiding documents, positioning “health big data” and “internet-based healthcare” as core pathways for chronic disease management, ensuring policy direction and stability. Second, they deeply focus on technology empowerment and integrated innovation. P14 actively promotes the deep integration of internet technology with healthcare services. P17 provides data support for intelligent chronic disease management. High-frequency terms like “data interoperability,” “platform development,” and “technical standards” appear significantly more frequently in these two policies, reflecting a strong emphasis on technological convergence. Third, emphasizing specialized precision measures, P15 stands as the sole specialized policy, ensuring coordinated and mutually reinforcing policy components.

The remaining 13 policies achieved a “good” rating, indicating that China's smart chronic disease management policy framework has taken shape but still holds room for improvement. First, while policies cover a broad scope, they lack sufficient specialization (e.g., P6, P18). Second, interdepartmental coordination mechanisms are nascent with limited innovation-driving capacity (e.g., P7). Third, foundational standards have taken shape but supporting infrastructure remains weak (e.g., P5). Compared to “excellent” policies, these “good” policies fall short in three main areas: insufficient specialized policy supply (only P16 is specialized among the 13); lack of specific application scenarios for technologies (e.g., P1); and absence of incentive and safeguard mechanisms (e.g., P9).

P14 received the highest score of 8.5 and was rated as an “excellent” policy. P9 received the lowest score of 6, yet was still classified as “Good.” The disparity likely stems from the following factors: In terms of policy positioning, P14 possesses a more macro strategic orientation and comprehensive policy framework. Regarding technological innovation, P14 demonstrates stronger leadership, encouraging new service models and the deep application of emerging technologies. In contrast, P9 primarily focuses on migrating traditional medical insurance services online. Regarding implementation safeguards, P14 establishes more comprehensive supporting mechanisms and demonstrates broader radiating effects across multiple sectors. The key reason P9 maintains a “good” rating lies in its resolution of the public pain point of healthcare service accessibility.

This study employs a combined approach of text mining and PMC index analysis. The PMC Index scores for the 18 policies ranged from 6 to 8.5, with an average score of 7.31, reflecting overall advantages in policy design and implementation.

However, compared with existing literature, this study reveals potential problems in the implementation of chronic disease management policies in China. Xue et al. (52) pointed out systemic deficiencies in China's chronic disease management policies, such as weak integration, unclear communication mechanisms in medical institutions, and lagging information platform construction. The analysis in this study shows that China's chronic disease management policies score higher in the dimensions of “policy tools” and “policy content,” but the score for “policy timeliness” is low. This highlights shortcomings in policy design and affects overall effectiveness, which is consistent with Xue et al.'s findings.

Li et al. (53) noted that the lack of long-term planning and flexibility in policy adjustments could affect sustainability and adaptability. This study echoes that point. In terms of “policy focus” and “policy timeliness,” the depth and flexibility of the policies are insufficient. Only 10 of the 18 policies have a long-term plan of 5 years or more, reflecting weaknesses in long-term sustainability and dynamic adjustment. Additionally, the lack of depth and focus in policy priorities poses further challenges.

Therefore, we propose the following suggestions:

Firstly, strengthen differentiated long-term planning and dynamic adjustment mechanisms. Based on the regional development level, formulate a long-term plan with hierarchical classification for 5 years or more, and clarify the phased goals: the developed eastern region focuses on the functional upgrading and deep application of smart chronic disease management platforms, the central region focuses on the construction of regional level chronic disease management centers and cross institutional data exchange, and the western region prioritizes the construction of smart infrastructure for grassroots medical institutions to fill gaps. At the same time, establish a normalized policy evaluation mechanism, rely on a stakeholder alliance composed of health administrative departments, medical institutions, research institutes, and patient groups, build regular feedback channels, and dynamically optimize planning content based on regional implementation effectiveness and people's livelihood needs.

Secondly, focus on key policy areas and clarify the precise implementation path of core technologies and services. Anchor the scenario application direction of core technologies such as AI, big data and IoT: support the research, development and promotion of AI algorithm in risk stratification and complication warning of chronic diseases such as hypertension and diabetes; Promote the popularization of IoT wearable devices in real-time monitoring of physical signs and medication reminders for chronic disease patients at home, especially for the older adults in rural areas, and develop low-cost and easy-to-use device adaptation solutions. Build a service system covering high-risk population screening, standardized diagnosis and treatment, and long-term health management around the entire process of chronic disease prevention, diagnosis, treatment, and rehabilitation, and connect it with medical insurance payment policies to include eligible smart chronic disease management services in the reimbursement scope.

Thirdly, establish a multi departmental collaborative governance mechanism to promote resource integration and benefit coordination. Establish a joint conference system led by the health department and involving multiple departments such as medical insurance, industry and information technology, finance, and civil affairs, clarifying the responsibilities of each department: the health department is responsible for policy standard formulation and clinical pathway standardization, the medical insurance department is responsible for payment method reform and cost control, the industry and information technology department is responsible for technical standards and industry support for smart medical equipment, and the finance department is responsible for special funding support for central and western regions and grassroots institutions. At the same time, we will establish a diversified collaboration model of “government guidance+market participation,” encourage private capital to participate in the construction and operation of smart chronic disease management platforms, promote medical institutions, research institutes, and technology enterprises to form industry university research innovation alliances, and accelerate the transformation of technological achievements.

Finally, improve the policy evaluation and supervision system that takes into account multiple dimensions. Construct a comprehensive evaluation framework that covers policy effectiveness, social benefits, economic efficiency, and implementation feasibility, incorporating core indicators such as regional adaptability, infrastructure coverage, patient satisfaction, and local fiscal affordability, fully considering political and economic factors such as economic development levels and medical resource endowments in different regions. Establish a dual supervision mechanism of “government supervision and third-party evaluation,” introduce independent third-party institutions to carry out policy implementation effect evaluation, and at the same time, open up public supervision channels to ensure transparency and fairness in the policy implementation process.

This study also has several limitations. Firstly, the PMC analysis method has technical limitations: the binary scoring method used for quantitative analysis is a necessary simplification of the complexity of policy provisions and may not reflect the differentiated importance of policy texts; Although the equal weight setting of analysis dimensions is based on the consideration of structural balance, its optimality needs further verification in subsequent research; In addition, the evaluation results lack external correlation testing with the actual implementation effect data of policies, mainly reflecting the characteristics of text content. Secondly, this study focuses on national level policies. Although it helps to grasp the macro strategic framework, it excludes provincial and local policies, which may introduce selection bias and make it difficult to fully reflect the huge differences in disease spectrum, resource endowment, and cultural customs in different regions of China. These factors profoundly affect the local adoption and implementation effectiveness of policies.

Future research can be improved in three areas: (1) Expand the scope to incorporate local policies into the analytical framework, examining the interplay between national and local policies in light of regional disease characteristics and economic-cultural variations to uncover local insights; (2) Adopt mixed-methods research combining qualitative approaches (e.g., surveys, in-depth interviews) with medical big data analysis to construct a “text-practice-outcome” three-dimensional evaluation model; (3) Establish dynamic tracking mechanisms using policy simulation techniques to predict policy evolution trends and develop intelligent monitoring systems for real-time implementation evaluation. It is recommended to prioritize multidisciplinary research integrating medicine, policy science, and data science methodologies to construct a Chinese-specific theoretical framework for intelligent chronic disease governance, thereby providing decision support for tiered diagnosis and treatment reforms.