Population aging has become a significant challenge to global socioeconomic development [1]. According to the United Nations’ 2024 World Population Prospects, by the end of the 21st century, the global population aged 65 years and older is expected to reach 2.2 billion, surpassing the population of children younger than 18 years, and this proportion will continue to rise. China’s aging process is particularly rapid, with the population aged 65 years and older accounting for 15.6% of the total in 2024, and this share is expected to accelerate in the coming decades. In this context, the health—particularly the mental health dimension, such as depression—of middle-aged and older adults is not only crucial for individual well-being but also profoundly impacts health care resource allocation, the stability of social security systems, and sustainable economic development.

In recent years, improvements in social security systems and advancements in medical technologies have somewhat enhanced the physical health of middle-aged and older adults, but depression has still not received adequate attention [2]. Depression is one of the most common psychological disorders among older adults. Studies show that the global prevalence of major depression is as high as 9.7% [3], while the depression rate for middle-aged and older adults in China is 24.1% [4], with severe depression accounting for 2.3% [5]. Depressive symptoms not only increase individual medical expenditures and social health care burdens [6] but may also impair the social functioning of middle-aged and older adults, reduce their quality of life, and further affect social welfare systems and macroeconomic development [7]. However, current health evaluation systems often underestimate the long-term socioeconomic impacts of depression, leading to its neglect in public policy and resource allocation [8]. In this context, how to effectively reduce depressive symptoms among middle-aged and older adults has become a key issue for advancing healthy aging.

Social networks are considered one of the most important factors affecting the depression status of middle-aged and older adults. As retirement, children’s migration for work, and the increased mobility of social networks take place, interactions with the outside world may decrease, leading to an increased sense of social isolation, which exacerbates depressive symptoms [9]. Social networks not only include direct face-to-face interactions but also extend to remote social connections facilitated by digital technology [10]. Existing studies show that middle-aged and older adults with lower levels of social networks are more prone to anxiety and depression, while actively maintaining social relationships helps reduce depressive symptoms [11]. The promotion of digital infrastructure development and the widespread availability of the internet have provided middle-aged and older adults with new social channels, alleviating social isolation to some extent. The application of digital technology can help reduce depression by enhancing social networks and increasing social support [12]. On the one hand, the internet allows middle-aged and older adults to maintain closer contact with their children and friends, reducing social alienation and thus lowering feelings of loneliness and risk of depression [13]. On the other hand, the use of social media and web-based interaction platforms expands their social networks, enhancing their sense of social participation and helping to alleviate depressive symptoms [14].

However, there is still controversy regarding the impact of digital technology on depression among middle-aged and older adults. One view holds that digital technology can compensate for the social interaction and participation deficits of middle-aged and older adults, thus reducing depressive symptoms [15]. Another view, based on the “substitution effect of presence,” argues that excessive reliance on digital technology may reduce real-world social interactions, weakening the emotional support provided by face-to-face communication, and thus negatively affecting depression outcomes [16]. Moreover, excessive use of social media may also lead to issues such as information overload and internet fraud, increasing anxiety and psychological stress [17]. Although these studies provide valuable insights into the relationship between digital technology and depression, there is still no consensus in the academic community regarding how digital infrastructure impacts the depressive symptoms of middle-aged and older adults. Current research mainly focuses on internet use at the individual level, with less attention paid to the systemic impact of digital infrastructure development as an exogenous variable. Additionally, the role of social networks in this process remains underexplored. Therefore, examining how digital infrastructure development influences the depressive symptoms of middle-aged and older adults through the enhancement of social networks is not only of significant academic value but also has practical implications for policy-making on healthy aging.

Based on this, this paper uses the quasi-natural experiment of the “Broadband China” pilot policy to systematically examine the effect of digital infrastructure development on depressive symptoms among middle-aged and older adults and explore the mechanisms through social networks. Compared with existing studies, the main contributions of this paper are as follows: First, rigorous identification leveraging policy shocks. Existing research has focused on the correlation between individual internet use and depression, making it difficult to overcome endogeneity issues. This paper uses the “Broadband China” pilot, applying the difference-in-differences (DID) method to estimate the effect of digital infrastructure development on depressive symptoms, providing more externally valid empirical evidence for policy optimization. Second, it reveals the role of social networks. Although social networks are considered an important factor influencing depression, few studies have focused on how digital infrastructure reshapes the social patterns of middle-aged and older adults and their implications for depressive symptoms. This paper focuses on how digital infrastructure enhances family connection, close social interactions, and social participation, improving social interaction levels, thereby alleviating loneliness and reducing depressive symptoms, filling a gap in the literature. Third, expanding the perspective of health inequality. This paper further examines the heterogeneous effects of digital infrastructure development across different sexes, ages, and socioeconomic groups and, by combining the Kakwani relative deprivation index, reveals how the digital divide affects inequality in depression outcomes among middle-aged and older adults. This analysis not only deepens the understanding of the social effects of digital infrastructure development but also provides empirical support for the government to develop targeted healthy aging policies.

The data for this study are drawn from the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative survey conducted by the National School of Development at Peking University. Since 2011, CHARLS has carried out follow-up surveys every 2 to 3 years, covering 150 districts and counties nationwide through a multi-stage, probability-proportional-to-size sampling design [41,42]. The survey repeatedly collects extensive information on respondents—including demographic characteristics, socioeconomic conditions, health status, and family structure—ensuring comparability across waves.

This study uses 5 survey waves (2011, 2013, 2015, 2018, and 2020). The dependent variable, depressive symptoms (measured by the Center for Epidemiologic Studies Depression Scale [CES-D] 10 scale), is consistently available across all waves. Time-varying individual-level covariates—such as age, sex, marital status, education, family size, pension status, insurance coverage, and chronic disease status—are also updated in each wave. The key explanatory variable, digital infrastructure development, is constructed by linking city-level “Broadband China” pilot policy data to respondents based on their place of residence in each survey year.

The original pooled dataset contained more than 60,000 person-wave observations. To handle missing data, we applied listwise deletion, excluding respondents with incomplete information on the dependent variable or any covariates. The final analytic sample comprises 56,211 person-wave observations of adults aged 45 years and older. Because the proportion of missing cases was modest and not systematically patterned, the resulting sample remains broadly representative, and the risk of bias due to missing data is minimal.

The CHARLS project was reviewed and approved by the Institutional Review Board of Peking University (IRB00001052-11015). All participants, including middle-aged and older adults, provided written informed consent prior to participation. This study used deidentified, publicly available data, and all procedures complied with the ethical standards of the Declaration of Helsinki. Participants’ privacy and confidentiality were maintained throughout the study. No further ethics approval was required for this secondary data analysis.

The core explanatory variable of this study is digital infrastructure development. According to the “Broadband China” Strategy and Implementation Plan issued by the State Council in 2013, the core goals of the strategy include enhancing broadband network coverage, optimizing network service quality, and promoting the development of digital applications to improve regional digital infrastructure. The implementation of the “Broadband China” strategy not only represents the government’s systematic investment in digital infrastructure but also has created significant regional infrastructure disparities, providing a suitable policy experiment environment for this study. Therefore, following the approaches of Peng et al [43], He et al [44], and Jia and Li [45], this study uses the “Broadband China” pilot policy as a proxy for digital infrastructure development. If middle-aged and older adult individual i in city c was included in the “Broadband China” pilot program in year t, the binary variable did is assigned a value of 1; otherwise, it is assigned a value of 0. For individuals who changed residence during the observation period, the assignment of did was based on the policy status of their current city in each survey wave, thereby dynamically capturing the actual local policy environment experienced by each respondent. Since the proportion of cross-city movers in the sample was very small, the potential influence of such cases on the estimation results is negligible.

The dependent variable in this study is depressive symptoms among middle-aged and older adults. Based on the work of Ma et al [46], depressive symptoms were measured using the CES-D. This scale includes 10 questions related to core symptoms and common manifestations of depression, such as: “I get upset over little things,” “I have trouble concentrating,” “I feel downhearted,” “I feel that things are an effort,” “I am hopeful about the future,” “I feel frightened,” “I have trouble sleeping,” “I feel happy,” “I feel lonely,” and “I feel like I can’t continue my life.” CHARLS asks respondents about their feelings and behaviors during the previous week, with response options ranging from “Rarely or none of the time (<1 day)” to “Most of the time (5-7 days).” Each item is scored from 0 to 3, with 2 items, “I am hopeful about the future” and “I feel happy,” being reverse-coded to align with the other items. The total score for these 10 items is the CES-D score, which serves as the depression outcome variable in this study, with a range from 0 to 30. A higher score indicates worse depressive symptoms. The CES-D has been widely applied in studies of middle-aged and older adults, and previous research has demonstrated its good reliability and validity [47,48].

The mechanism variable in this study is the social network of middle-aged and older adults, which includes family connections, close social interactions, and social participation. According to social network theory, the network is not only characterized by the breadth of relationships but also by their intimacy [30]. Therefore, the 3 dimensions selected in this study reflect varying degrees of intimacy in social relationships. Family connection, as the most intimate dimension of social contact, is measured by the frequency of contact with children. This frequency is based on the response to the question “How often do you see your children when you don’t live with them?” The response scale covers multiple time frames, ranging from “almost every day,” “two to three times per week,” “once a week,” “once every two weeks,” “once a month,” “once every three months,” “once every six months,” and “once a year,” to “almost never.” Thus, the measure captures the regularity of parent-child contact across daily, weekly, monthly, and annual intervals. Values range from 1 to 9, where a higher score indicates less frequent contact and reflects weaker family connections. For middle-aged and older adults with multiple children, the mean frequency across all children is used. Close social interactions represent looser social relationships and are measured by the response to “Have you visited friends or socialized in the past month?” A response of “Yes” is assigned a value of 1, and “No” is assigned a value of 0. Close social interactions reflect the frequency of interaction with nonfamily members, which typically offers emotional support and psychological relief, while indicating interaction with a more distant social circle. Social participation reflects broader social connections and indicates the extent of involvement in public and external social arenas. Social participation is measured by responses to whether, in the past month, the individual participated in activities such as “playing mahjong, chess, cards, or attending community events,” “joining social or volunteer activities,” or “attending educational or training courses.” A response of “Yes” is scored as 1, otherwise 0. The social participation score is the sum of these 4 activities, ranging from 0 to 4, indicating the individual’s involvement in social interaction and public affairs.

This study also controls for other factors that may affect the depressive symptoms of middle-aged and older adults, including individual characteristics (such as residence, sex, age, education level, ethnicity, pension insurance, health insurance, and chronic diseases) and family characteristics (such as marital status and household size). Furthermore, the study controls for time fixed effects and city fixed effects. All data management and statistical analyses were performed using Stata (version 18.0; Stata Corp). Descriptive statistics for the relevant variables are shown in Table 1, with additional descriptive statistics by survey wave reported in Multimedia Appendix 1.

In economic and social research, the DID method is commonly used to assess policy effects. The fundamental approach involves treating the policy pilot as an exogenous “quasi-natural experiment.” The implementation of the Broadband China pilot policy may generate pre- and postimplementation differences within the pilot regions, as well as differences between pilot and nonpilot regions at the same point in time. By performing regression analysis on these 2 types of differences, the net effect of the policy on the depressive symptoms of middle-aged and older adults can be effectively identified. Therefore, this study treats the Broadband China policy as a quasi-natural experiment and uses the DID method to examine its impact. The baseline regression model is specified as follows:

Where the subscripts i, c, and t represent individual, city, and time, respectively. Depression_ict is the dependent variable, indicating the level of depressive symptoms of middle-aged and older adult individual i in city c at year t. DID_ct is the core explanatory variable, representing whether city c implemented the “Broadband China” pilot in year t. It is the interaction term between the city-level treatment group indicator and the pre and postpilot implementation dummy variables for the “Broadband China” initiative. Controls_ict represents the individual and household-level control variables. γ_c and δ_t are the city fixed effects and time fixed effects, respectively. ε_ict is the random error term.

The validity of the estimation in equation 1 relies on the parallel trend assumption, which asserts that, in the absence of policy intervention, the dependent variable trends of the treatment and control groups should exhibit similar patterns. To test this assumption and explore the dynamic effects of the policy, the study adopts an event study approach, following the methodology of Beck et al [49]. The model for the dynamic effect analysis is as follows:

Where β_t represents the estimated values corresponding to the years 2011-2020. The definitions of other variables are consistent with those in equation 1.

The Kakwani relative deprivation index [50] effectively quantifies disparities in depressive symptoms. According to the theory of relative deprivation, individuals with poorer health experience greater health deprivation and higher levels of health inequality. This study follows the methodologies of Li et al [41] and Turguttopbaş [51] and uses the Kakwani index to measure inequality in depressive symptoms. Specifically, since our data are panel in nature, the index was computed on a wave-by-wave basis (2011, 2013, 2015, 2018, and 2020). In each wave, the depressive symptom levels of middle-aged and older adults were ranked, and the relative deprivation value of each individual was calculated within that year’s reference group. As a result, every respondent has a time-varying relative deprivation value, reflecting their relative health position in each survey year. Formally, let Y represent the reference group in year t, with a sample size of n, and depressive symptom distribution vector y=(y₁, y₂, y₃, y_n–1, y_n), where y₁ ≤ y₂ ≤ y₃ ≤ y_n–1 ≤ y_n. Therefore, the relative deprivation index RD (y_j, y_i) of middle-aged and older adult individual i compared with individual j is defined as:

Based on equation 4, the average relative deprivation index of depressive symptoms for individual i in year t is calculated as follows:

Where μ_y is the average depression score of all samples in the reference group Y, n_yi⁺ is the number of middle-aged and older adults in the reference group Y whose health exceeds the health level of individual i (denoted as y_i), and μ_yi⁺ is the average health score of middle-aged and older adults in the reference group Y whose health exceeds y_i.

To address potential multicollinearity, we calculated variance inflation factors (VIF) for all explanatory variables. All VIF values were well below the conventional threshold of 10 (mean VIF=1.17), suggesting that multicollinearity is not a concern. Model fit is evaluated using adjusted R² (within), reported in each regression table.

Table 2 presents the baseline regression results on the impact of digital infrastructure development on the depressive symptoms of middle-aged and older adults. In Model 1, which includes only the core explanatory variables, the coefficient of the interaction term is negative and statistically significant. Models 2 and 3 include individual and household characteristics, respectively, and the coefficient of the interaction term remains significant, indicating the robustness of the results. Model 4 further incorporates time and regional fixed effects. The coefficient of the interaction term is –0.3349 and is statistically significant at the 1% significance level. These results suggest that digital infrastructure development is associated with a significant reduction in depressive symptoms among middle-aged and older adults, thus supporting Hypothesis 1.

To explore the mechanisms through which digital infrastructure development affects the depressive symptoms of middle-aged and older adults, this study introduces social network as a mediator, dividing it into 3 dimensions: family connection, close social interactions, and social participation. The regression results in Table 5 show that digital infrastructure development significantly promotes these 3 dimensions of the social network, thereby further reducing depressive symptoms among middle-aged and older adults, confirming the mediating effect.

To test the significance of the mediation effect, this study uses the Bootstrap method. If the CI of the indirect effect does not contain zero, it indicates that the effect is significant. Table 6 shows that the indirect effects of family connection (95% CI –0.0256 to –0.0038), close social interactions (95% CI –0.0162 to –0.0030), and social participation (95% CI –0.0290 to –0.0029) are all significant. These results are consistent with the previous regression analysis, indicating that digital infrastructure development effectively improves psychological well-being by mitigating depressive symptoms through strengthened social networks in various dimensions.

Given the considerable differences in socioeconomic characteristics, health conditions, and levels of digital technology acceptance among middle-aged and older adults, the impact of digital infrastructure development on depressive symptoms is likely to vary across subgroups. Table 7 reports the heterogeneous effects. The results indicate that the policy significantly reduces depressive symptoms among the middle-aged (Model 1) and individuals with lower levels of education (Model 3), with coefficients of –0.4084 (P<.01) and –0.3300 (P<.01), respectively. In contrast, the effects for older adults (Model 2) and those with higher education (Model 4) are weaker and statistically insignificant, with coefficients of –0.2599 (P>.1) and –0.3769 (P>.05), suggesting that the policy’s benefits for depressive symptom reduction are not robust in these groups.

Turning to sex, digital infrastructure development significantly reduces depressive symptoms among both women and men, though the effect is stronger for women (–0.3794, P<.01) than for men (–0.2931, P<.05). Taken together, these findings highlight the differentiated impacts of the “Broadband China” policy, showing that middle-aged adults, less-educated individuals, and women derive relatively greater reductions in depressive symptoms, thereby underscoring the policy’s potential to mitigate health inequalities.

Table 8 presents the impact of digital infrastructure development on health inequality among older adults. Based on the Kakwani relative deprivation index, this study first reversely encodes the depressive symptoms variable, where higher scores indicate better health. Regression analysis is then conducted. Model 1 does not include control variables, while Models 2 and 3 introduce individual and family characteristics, respectively. Model 4 further considers time and regional fixed effects. In Model 4, the coefficient for the core explanatory variable did is –0.0089 (P<.01), indicating that digital infrastructure development significantly reduces inequalities in depressive symptoms among middle-aged and older adults. This finding confirms the role of digital infrastructure development in alleviating health inequality among middle-aged and older adults.

This study uses panel data from CHARLS 2011-2020 and adopts the DID method, based on the “Broadband China” pilot policy, to assess the impact of digital infrastructure development on the depressive symptoms of middle-aged and older adults, analyze its mechanisms, and explore its effects on health inequality in this group. The findings are as follows:

First, digital infrastructure development has been shown to reduce depressive symptoms of middle-aged and older adults. The findings of this study remain robust even after controlling for selection bias and unobserved factors, providing further evidence of the association effect of digital infrastructure development on the depressive symptoms of older adults. This result is consistent with existing literature, which highlights that digital technologies can effectively alleviate depression [54], a conclusion that applies to various populations, such as children [55], adolescents [56], and older adults [24]. However, the innovation of this study lies in its departure from the traditional focus on individual or household-level perspectives, instead examining the impact of digital infrastructure development on depressive symptoms from a macro policy perspective. This approach reveals the positive role of government-driven policy interventions aimed at expanding digital technology access in enhancing the well-being of the older adult population, further emphasizing the critical role of policy orientation in improving health outcomes.

Second, digital infrastructure development helps to alleviate depressive symptoms among middle-aged and older adults by promoting the improvement of their social networks. This finding is consistent with social network theory, which highlights the critical role of social connections in individual psychological health [57]. Existing research has confirmed that social networks provide emotional support, reduce feelings of loneliness, and enhance social identity [58]. However, while many studies focus on the impact of social networks on psychological health, few explore how digital infrastructure development can promote the formation and enhancement of social networks at a macro level. This study fills this gap by revealing the role of digital infrastructure development in strengthening the social networks of middle-aged and older adults, thereby mitigating depressive symptoms. It further underscores the systematic effect of policy interventions on promoting psychological well-being.

Further analysis shows that improvements in depressive symptoms among middle-aged and older adults are primarily achieved through 3 social network pathways: enhancing family connection, close social interactions, and social participation. The strengthening of these pathways effectively reinforces the social support system, ultimately leading to reduced depressive symptoms. Strengthening family connections is likely due to the reduction in communication costs provided by digital technologies, enabling more frequent and intimate interactions between middle-aged and older adults and their children or relatives [59]. Regarding close social interactions, the proliferation of digital platforms, especially social media, provides a more convenient channel for older adults to stay in touch with friends and expand their social circles, thereby alleviating feelings of loneliness and enhancing social support [60]. Digital technologies have also significantly promoted the social participation of middle-aged and older adults, particularly in community activities, web-based learning, and public affairs, which not only increases their sense of social role but also strengthens their sense of belonging [33]. These findings indicate that digital infrastructure development not only represents an investment in material resources but also has a profound influence on social relationships, highlighting the critical role of digital policies in reducing depressive symptoms among middle-aged and older adults.

Third, the effects of digital infrastructure development on depressive symptoms vary significantly across different subgroups. Specifically, the positive association between digital infrastructure development and improved psychological well-being is more pronounced among women, middle-aged adults, and those with lower education levels. Consistent with existing research, the health improvements in these groups may stem from the promotion of social welfare and individual lifestyle changes facilitated by digital infrastructure development [61,62]. However, this study further reveals the underlying reasons why these groups benefit more.

The significant benefits observed in middle-aged adults suggest that the impact of digital infrastructure development on this group is closely related to the awakening of their health awareness and changes in behavior. Middle-aged adults are typically at a critical transitional stage in life, facing pressures related to health management and the dual burdens of work and family. Digital infrastructure can provide this group with more convenient access to health information and services, which helps mitigate depressive symptoms. Therefore, health interventions targeted at middle-aged adults should focus more on the application of digital tools to enhance health awareness and promote behavior change.

The significant benefits observed among individuals with lower education levels reveal the key role of education in digital technology acceptance and use. Lower-educated groups often face higher health risks, and the simple and user-friendly health tools and educational content provided by digital technologies can effectively reduce cognitive and technological barriers [63]. These groups’ ability to access health information and manage their health is enhanced through digital infrastructure. Therefore, digital health interventions targeted at lower-educated groups should focus more on designing health tools that align with their cognitive abilities and technical skills to ensure effective participation and benefit.

At the same time, this study also identifies significant sex differences. Digital infrastructure development has a positive impact on reducing depressive symptoms for both men and women, but the effect is more pronounced among women. This finding may be linked to the multiple responsibilities and pressures that women often bear in both family and social roles [64]. The improvement of digital infrastructure not only provides more opportunities for social interaction but also facilitates access to health resources and psychological support, thereby helping to alleviate women’s emotional distress and mental burden to some extent. Therefore, in advancing digital infrastructure development, greater attention should be directed toward addressing the specific needs of women, making use of digital tools to enhance their health and psychological support, and ultimately promoting sex equity in health outcomes.

In contrast, the effects on older adults and those with higher education levels are relatively weaker, which may be closely related to their existing health resources and methods of accessing information. Older adults often face multiple challenges with their health, but they may have higher barriers to adapting to new technologies [65]. Therefore, although digital infrastructure can provide convenience for older adults, issues such as technological adaptability and the learning curve may affect their health benefits. Additionally, urban residents and those with higher education levels are generally more easily able to access medical services and health information, meaning the marginal benefits of digital infrastructure development are lower in these groups.

Fourth, digital infrastructure development has alleviated health inequality in depressive symptoms among middle-aged and older adults to some extent. Previous studies have shown that health inequality is influenced by various factors, including education level [66] and geographic location [67]. However, most existing literature focuses on how traditional socioeconomic factors exacerbate health inequality, with less emphasis on how digital infrastructure can bridge gaps among different social groups by optimizing the allocation of public resources and improving the accessibility of services related to depression. This study finds that digital infrastructure development has had a significant positive effect on disadvantaged groups, demonstrating a certain “compensatory effect.” A possible explanation is that traditional health services are often limited by spatial, temporal, and financial constraints, whereas digital infrastructure can provide remote medical care, health education, and information access, significantly reducing the geographical and economic barriers faced by disadvantaged groups and improving their access to health services. Moreover, the ease of use of digital technologies and the convenience of information delivery help reduce the cognitive burden faced by groups with lower health literacy during health decision-making, thereby enhancing their ability to manage depressive symptoms independently. This “compensatory effect” suggests that under the promotion of digital infrastructure development, traditionally disadvantaged groups can better access health services and improve their health outcomes. This not only helps reduce health disparities among groups but also demonstrates the enormous potential of digital technology in promoting health equity.

This study has several strengths. First, it innovatively identifies the associations between digital technology and depressive symptoms. Based on the natural experiment of the “Broadband China” pilot policy, the study uses the DID method to address endogeneity issues and provides external validity evidence of the effects of digital infrastructure development on the depressive symptoms of middle-aged and older adults, filling the gap in existing research regarding the impact of macro digital policy interventions on depression. Second, it explores the social network mechanism in depth. The study reveals how digital infrastructure promotes family connections, close social interactions, and social participation to mitigate depressive symptoms among middle-aged and older adults, enriching the theoretical framework of the interaction between digital infrastructure and social networks. It provides empirical evidence for policymakers to improve the well-being of middle-aged and older adults while promoting social network development. Third, it expands the perspective of health inequality. By measuring disparities in depressive symptoms and conducting heterogeneity analysis from multiple perspectives, the study explores the “compensatory effect” of digital infrastructure development on low-resource groups, providing strong empirical support for healthy aging policies and emphasizing the tremendous potential of digital technology in reducing health disparities.

However, the study has some limitations. First, the data used in this study were primarily collected and assessed through self-reports or proxy reports, which may introduce recall bias and subjectivity, especially concerning depressive symptoms and social network-related issues. Nevertheless, the CHARLS survey follows standardized protocols, ensuring the reliability and comparability of cross-wave data, thus effectively reducing potential biases. Second, due to limitations in data updates, this study was unable to obtain the CHARLS database after 2020. Therefore, future research could extend the time frame and further explore the long-term effects of digital infrastructure on depressive symptoms among middle-aged and older adults. Third, this study did not analyze other potential factors that may affect depression, such as the expansion of long-term care pilot projects and the improvement of health literacy. These factors may influence depressive symptoms to some extent. Future research should incorporate more diversified data sources to further assess the associations between these factors and depression among middle-aged and older adults.

This study, based on panel data from CHARLS 2011-2020, uses a quasi-natural experiment based on the “Broadband China” pilot policy to evaluate the impact of digital infrastructure development on the depressive symptoms of middle-aged and older adults. It also explores the role of social networks and subgroup heterogeneity. The key conclusions are as follows:

First, digital infrastructure development significantly improves the depressive symptoms of middle-aged and older adults, confirming the effectiveness of digital technology in reducing depression. The study analyzes the positive role of government-driven digital infrastructure development from a macro policy perspective, emphasizing the critical role of policy guidance in promoting reductions in depressive symptoms and improving psychological well-being. Second, digital infrastructure development alleviates depressive symptoms by strengthening the social networks of middle-aged and older adults. The study finds that digital infrastructure effectively promotes family connection, close social interactions, and social participation, thereby enhancing social support systems. This indicates that digital infrastructure not only impacts material aspects but also has profound effects at the social relationship level. Third, the impact of digital infrastructure development on depressive symptoms across different subgroups exhibits significant heterogeneity, particularly for women, middle-aged adults, and low-education groups. This suggests that digital infrastructure development helps these groups overcome barriers to health information access and health management, thereby reducing depression levels. Fourth, digital infrastructure development plays an important role in alleviating health inequality. By increasing the accessibility of health services, it significantly reduces depressive symptoms among low-resource groups, thereby improving their psychological well-being, demonstrating the “compensatory effect” of digital infrastructure.

Based on these conclusions, the study proposes the following policy recommendations. First, digital infrastructure should be embedded in long-term national strategies, with coordinated governance and financial investment to reduce regional disparities. Such efforts resonate with the World Health Organization’s Global Strategy on Digital Health and highlight the importance of aligning national initiatives with global frameworks. Second, improving digital literacy across the life course is equally crucial. Lifelong, tailored training—especially for women, middle-aged adults, and low-education groups—should combine web-based platforms with community-based programs, drawing on international experiences such as the EU’s Active and Healthy Ageing initiatives. Third, strengthening social networks through digital platforms can further alleviate depressive symptoms. Public-private partnerships that encourage family interaction, community participation, and intergenerational support have proven effective in other contexts, such as Japan and Europe, and can be adapted locally. Finally, digital health services should be fully integrated into primary care systems. Expanding reimbursement for telemedicine, embedding digital depression-related services into community health centers, and prioritizing underserved regions, as demonstrated in Australia’s telehealth reforms, can promote inclusivity and reduce health inequalities.