Abstract

Objective

The aging population represents a formidable global challenge, with China experiencing an accelerated demographic shift. While previous research has established a directional link between mental health literacy, social participation, and active aging, the moderating effect of socioeconomic status (SES) on these associations remains underexplored. This study sought to address this gap by employing moderated network analysis, in contrast to the total score approaches commonly used in prior literature.

Methods

A cross-sectional design was implemented, involving 1,032 Chinese individuals aged 60 and above, who completed the Mental Health Literacy Questionnaire, Active Aging Scale, Social Participation Questionnaire, and Socioeconomic Status Index Scale. The moderated network method was applied to explore the bidirectional relationships between mental health literacy, social participation, and active aging, while examining the moderating role of SES.

Results

The analysis identified the strongest bidirectional relationships between attitudes toward mental illness and the development of spiritual wisdom. Seven interaction terms emerged involving mental health literacy, social participation, and active aging, moderated by SES. Notably, the most significant interaction terms were found between attitudes toward mental illness and engagement in active learning.

Conclusions

These results contribute novel insights into the bidirectional relationships among mental health literacy, social participation, and active aging, as well as the role of SES in moderating these relationships. The findings highlight the need for targeted policy interventions to address socioeconomic inequalities in later life, thereby fostering active aging.

Introduction

Population aging is a global trend and a core element of China鈥檚 modernization trajectory. According to the seventh China population census, individuals aged 60 and above constitute 18.7% of the population, reflecting a 5.44% increase compared to the previous census. Projections suggest that by 2030, China will transition into a moderately aging society [1]. As the 鈥淎ging Giant,鈥� China is home to the world鈥檚 largest elderly population [2], characterized by the phenomenon of aging before achieving substantial wealth, accompanied by a rapidly increasing elderly population and a heightened dependency burden. This demographic shift presents significant societal and economic challenges [3]. In response, 鈥淗ealthy China鈥� and 鈥淎ctively Addressing Population Aging鈥� have become key national strategies. The World Health Organization defines active aging as optimizing health, participation, and security to improve the quality of life as people age [4]. However, the incidence of successful aging among Chinese older adults remains low [5], and the relationships between overall and positive perceptions of active aging, as well as the underlying mechanisms or personal resources explaining these associations, have yet to be fully examined [6].

Mental health literacy serves as a critical asset in promoting mental health, encompassing the knowledge, attitudes, and behaviors that individuals cultivate to improve both their own and others' mental health and to manage mental illnesses [7]. In comparison to Western nations such as Switzerland, Germany, and Australia, mental health literacy in China remains lower, with a notable decline observed in older age groups [8, 9]. Older individuals in China encounter substantial barriers to utilizing e-health resources, demonstrate poorer recognition of mental illness, and are less likely to seek professional medical treatment [10,11,12]. These disparities may be partly attributed to cultural stigma, as mental illness is still heavily stigmatized in Chinese society [13], and the traditional Chinese medical model of mental health continues to dominate [14]. In contrast, research in Western countries has consistently shown that higher mental health literacy correlates with improved active aging outcomes and healthier behaviors, including early detection of mental illness and reduced internalized stigma [15, 16]. Older adults with low mental health literacy face a cascade of negative consequences, such as increased susceptibility to depression, anxiety, and stress, along with a diminished ability to seek help鈥攆actors that significantly impede active aging [17, 18]. However, the impact of culturally specific factors on mental health literacy and active aging among Chinese older adults remains underexplored.

Bidirectional relationship between mental health literacy, active aging, and social participation

Social participation encompasses the exchange of individual resources through active involvement in collective pursuits, including religious gatherings, recreational clubs, sports teams, and cultural or political events [19]. The activity theory posits that sustaining engagement in middle-aged activities and attitudes for as long as possible promotes well-being across physical, psychological, and social domains [20].

Specifically, the relationship between social participation and mental health literacy appears to be bidirectional. Existing studies indicate that social participation can predict mental health literacy [21]. A large-scale study conducted in China reported similar findings [22]. Older adults who regularly participate in various social activities tend to gain significant social support, which in turn enhances their utilization of mental health services and increases formal help-seeking behaviors [23, 24]. On the other hand, social participation鈥攃onsidered a health behavior鈥攎ay also stem from mental health literacy. According to social cognition theory, individuals' beliefs about health behaviors influence both their intentions and actual engagement in such behaviors [25]. For instance, a meta-analysis revealed that older adults with adequate health literacy were more likely to engage in physical activity for鈥夆墺鈥�5听days per week compared to those with insufficient health literacy [26]. Moreover, perceived stigma has been shown to predict levels of social participation [27].

Social participation and active aging exhibit a reciprocal relationship, creating a self-reinforcing cycle. Among Chinese seniors, high-frequency and long-term participation in leisure sports has proven to be an effective strategy for promoting active aging [28]. Additionally, physical activity has been shown to alleviate death anxiety in older adults with previous COVID-19 exposure, with improvements in psychological well-being serving as a mediating factor [29]. Within the active aging framework, participation in community activities has been associated with more positive attitudes toward aging in both China and Japan [30, 31]. Conversely, negative perceptions of aging correlate with heightened depressive symptoms [32]. A bidirectional relationship has also been observed. A Chinese study demonstrated that while social participation alleviated depressive symptoms, depressive symptoms could, over time, negatively affect the extent of social participation in middle-aged and older Chinese adults [22].

Existing evidence suggests a link between mental health literacy, active aging, and social participation; however, the majority of studies have examined only unidirectional relationships between two of these variables. Few investigations have explored the potential bidirectional interactions among health literacy, active aging, and social participation.

Socioeconomic status as a moderator

Education, income, and employment prior to retirement consistently emerge as key determinants of risky health behaviors in older adults [33]. However, some researchers think that only education moderates the link between health literacy and self-management capabilities, with income playing a less significant role [34]. Socioeconomic status (SES), a multidimensional construct encompassing income, education, occupation, and wealth, collectively influences individuals' capacity to attain specific objectives [35]. A growing body of research highlights significant associations between SES and mental health literacy [7, 36], SES and active aging [37], and SES and social participation [38, 39]. Furthermore, SES is likely to moderate the relationships among these variables.

The Andersen model asserts that health behaviors mediate the impact of predisposing and enabling factors on health outcomes [40]. SES, as a predisposing factor, may moderate the relationship between mental health literacy and active aging. Psychological resilience plays a significant role in fostering successful aging [41]. Recent studies have revealed that the relationship between mental health literacy and psychological resilience is moderated by SES [42]. According to the fundamental cause theory, SES-related resources鈥攕uch as knowledge, financial means, and social capital (e.g., social participation)鈥攕erve to prevent health risks and reduce the impact of poor health [43]. Consequently, individuals with higher SES are more likely to maintain better health and face substantially lower odds of experiencing physical, psychological, and cognitive multimorbidity compared to those with lower SES [39, 44]. Furthermore, a longitudinal study demonstrated that lower SES was linked to deteriorating trajectories of active and healthy aging, as well as reduced quality of life over a 14-year follow-up period [45]. In conclusion, SES likely moderates the bidirectional relationship between mental health literacy, active aging, and social participation. Higher SES increases the likelihood that individuals with strong mental health literacy or higher levels of social participation will experience active aging. Conversely, high SES may elevate the likelihood that those already engaged in active aging will exhibit higher mental health literacy or greater social participation.

Current study

Several limitations in prior research merit further consideration. Although evidence suggests a strong interrelationship between mental health literacy, social participation, and active aging, no network analysis has simultaneously addressed these factors. Additionally, the moderating role of SES, a critical predisposing variable, remains unexplored. This study employs network analysis to elucidate the complex bidirectional relationships between mental health literacy, social participation, and active aging, while also identifying potential moderation effects. Network analysis facilitates the identification of key components (nodes) and their interrelationships (edges), adjusting for their interactions [46]. Each node represents a variable or symptom, and each edge indicates the partial correlation between two nodes, with thicker edges denoting stronger associations. However, when moderators are considered, traditional symptom networks may prove insufficient. To address this limitation, moderated network analysis, utilizing nodewise regression, can be applied. This method examines how the interaction between two symptoms or variables is moderated by other variables, offering deeper insights into the network鈥檚 structure [47,48,49]. In contrast to the traditional moderated mediation model, network analysis better captures the dynamic relationships among symptoms [50]. It represents an effective tool for investigating moderating variables [51].

This study proposes four hypotheses: (1) A bidirectional relationship exists between mental health literacy, active aging, and social participation; (2) SES acts as a moderator in the relationship between mental health literacy and active aging; (3) SES moderates the relationship between mental health literacy and social participation; and (4) SES moderates the relationship between social participation and active aging.

Methods

Participants and procedure

This study adopted a cross-sectional design, employing convenience sampling and snowball sampling to distribute questionnaires between July and December 2021. Trained interviewers conducted face-to-face interviews in participants' homes. Sichuan Province, located in southwest China, served as the study area. A total of 1,032 responses were collected, of which 919 were valid, yielding an effective response rate of 92.24%. Participants who were not residents of Sichuan, under 60听years of age, or who submitted incomplete questionnaires were excluded according to predefined criteria. The final sample consisted of individuals aged 60 to 91听years (mean age 70.73鈥壜扁��6.365), representing 21 cities across Sichuan Province.

Prior to obtaining informed consent, participants were provided with a detailed overview of the study's objectives, procedures, and time requirements. For those with limited literacy or physical impairments, researchers completed the questionnaire based on verbal responses. Informed consent was obtained from all participants. Ethical approval was granted by the Ethics Committee of Chengdu Medical College (Chengdu, China).

Measures

Mental Health Literacy (MHL)

The Mental Health Literacy Questionnaire [52] consists of 35 items across six dimensions: (1) knowledge about mental health, (2) knowledge about mental illnesses, (3)attitudes about mental health, (4) attitudes about mental illnesses, (5) behaviors regarding mental health promotion, and (6)behaviors regarding mental illness. Two sub-questionnaires on knowledge and perception were judgmental, while four sub-questionnaires on attitudes and behaviors were rated on a 5-point scale. A higher total score indicated a greater level of mental health literacy. The Cronbach's 伪 coefficient for this study was 0.876.

Active Aging Scale (AAS)

The active aging level of older adults was assessed using the AAS [53], a 36-item instrument comprising seven dimensions: (1) being self-reliant, (2) engaging in active learning, (3) building up financial security, (4)developing spiritual wisdom, (5) maintaining a healthy lifestyle, (6) being actively engaged with society, and (7) strengthening family ties to ensure care in later life care. Response options ranged from 1 (not at all true) to 4 (very true), with higher scores reflecting better active aging outcomes and lower scores indicating poorer outcomes. The AAS was specifically designed for evaluating active aging in older Chinese adults. In this study, the Cronbach's 伪 coefficient was 0.935.

Social Participation (SP)

The Social Participation questionnaire, adapted from the China Health and Retirement Longitudinal Study (CHARLS) [54], is modified to capture various forms of public engagement. It comprises six dimensions: activity type (social interaction, physical exercise, volunteer public welfare, cultural and intellectual activities, family activities), number, and frequency, rated on a 5-point scale. The total score reflects social engagement, with higher scores signifying greater social participation. In this study, the Cronbach's 伪 coefficient was 0.85.

Socioeconomic Status (SES)

SES was evaluated using three indicators: education level, average monthly personal income, and primary occupation before retirement (intellectual work predominates鈥�=鈥�1, physical work predominates鈥�=鈥�2) [39]. Education level was categorized in accordance with the International Standard Classification of Education [55] into low (less than primary education鈥�=鈥�1), medium (primary education completed鈥�=鈥�2), and high (secondary education or above鈥�=鈥�3). Average monthly income was classified based on the 2021 per capita disposable income of Sichuan residents, reported by the Sichuan Provincial Bureau of Statistics and the Sichuan General Survey Team of the National Bureau of Statistics as RMB 2423/month [56], the average monthly personal income was categorized as low (below RMB 2000/month) and high (above RMB 2000/month). All indicators were transformed into standard scores and analyzed through principal component analysis, which identified a principal factor with an eigenvalue greater than 1 (0.647), with factor loadings of 0.792, 0.816, and 0.805, respectively. The SES score was calculated as [ (0.792鈥壝椻�塟_education鈥�+鈥�0.816鈥壝椻�塟_occupation鈥�+鈥�0.805鈥壝椻�塟_income/0.647]. Higher scores reflected a higher SES.

Data analyses

Descriptive statistics were computed using SPSS (version 21.0), while R (version 4.2.3) was employed for statistical analyses. The symptom network approach presumes all outcomes (i.e., nodes, questionnaire items) follow a multivariate or univariate normal distribution, achieved through joint estimation of model parameters via the standardized inverse covariance matrix (i.e., the partial correlation matrix). The optimal penalty coefficient for shrinkage was determined using the extended Bayesian information criterion (EBIC) [57] and the graphical least absolute shrinkage and selection operator (LASSO) method [58, 59], resulting in a sparse network model based on node correlations and model selection.

The symptom network approach becomes impractical when moderators, covariates, or higher-order interaction terms are included [49]. In such cases, nodewise regression [47, 60], offers a viable alternative. This method utilizes a graph-theoretic technique called neighbourhood selection [61], which estimates the network structure through a series of univariate regression models, each representing a separate variable. The resulting coefficients are combined to define the final multivariate network model (MNM), where variables are categorized as either predictors or outcomes [49].

Symptom network models can be represented as MNMs based on a consistent methodological framework, though interpretations of the results may vary, and they may not always correspond to a normalizable joint distribution [48]. In contrast to strength centrality, the Expected Influence (EI)鈥攄erived by summing all raw edge weights connected to a node鈥攐ffers a more robust measure of a symptom's significance in a network containing both positive and negative edges [62].

Results

Descriptive statistics and correlation analysis

The means, standard deviations, and correlations of the variables were detailed in Table听1 and Fig.听1.

Table 1 Descriptive statistics for the network variables

Fig.听1

Heatmap of the Spearman correlation matrix. The number in each box is the Spearman correlation coefficient. In the upper triangular matrix, only significant coefficients are presented after Bonferroni correction test (p鈥�&濒迟;鈥�0.05), and the blank box indicates coefficients do not survive this correction

Network analyses for the whole study sample

To mitigate spurious relationships and reduce the risk of false-positive associations, hierarchical LASSO was used to select variables and estimate a sparse network structure, as shown in Table听2 and Figs.听2 and 3. The nodewise adjacency matrix for the MNMs was provided in Table听3, while Table听4 presented the interaction term matrix for network analysis. Figures听4 and 5 visualize the MNMs based on the nodewise and interaction term matrices.

Table 2 Variables were selected through the varSelect function selection procedure

Fig.听2

Plot showing the point estimates from each model, along with confidence intervals based on the estimated standard errors

Fig.听3

Plot the confidence intervals associated with interaction terms based on the MNMs

Table 3 The nodewise regression matrix for the network model (p鈥�&濒迟;鈥�.05)

Table 4 The matrix of interaction terms associated with the moderated network (p鈥�&濒迟;鈥�.05)

Fig.听4

A indicates MNMs of the network. Green lines indicate a positive relationship between two variables. Red lines indicate negative relationships at p鈥�<鈥�.05. The dashed line indicates that only one of the interactive relationships between two nodes is moderated by the SES variable at p鈥�<鈥�.05 (i.e., AND rule). Conversely, the solid line indicates that both nodewise regressions between two nodes are significant at p鈥�<鈥�.05 (i.e., AND rule). The light blue shading represents the R² for each nodewise regression

Fig.听5

Moderated network model with the exogenous moderator plotted with the AND rule. Green lines indicate a positive relationship between two variables at p鈥�<鈥�.05. Red lines indicate negative relationships at p鈥�<鈥�.05. The dashed line indicates that both of the interactive relationships between two nodes are moderated by the SES variable at p鈥�<鈥�.05 (i.e., AND rule). Conversely, the solid line indicates that both nodewise regressions between two nodes are significant at p鈥�<鈥�.05 (i.e., AND rule). The square represents SES along with arrows directly denoting its links to the other variables

Table 3 revealed the mutually significant bidirectional relationships between mental health literacy, social participation, and active aging. Nodewise regression analysis identified the three strongest associations: MHL4 (attitudes about mental illnesses) with AAS4 (developing spiritual wisdom), MHL4 with AAS3 (building up financial security), and MHL2 (knowledge about mental illnesses) with AAS4. Additionally, Table听4 and Fig.听2 highlight significant interaction terms with SES (socioeconomic status), including: MHL1 (knowledge about mental health) with SP4 (cultural and intellectual activities), MHL4 with SP1 (social interaction activities), MHL4 with AAS2 (engaging in active learning), MHL6 (behaviors regarding mental illness coping) with AAS2, SP2 (physical exercise activities) with AAS7 (strengthening family ties for later life care), SP4 with AAS3, and SP5 (family activities) with AAS3.

Figure听6 illustrated the variations in network structure at different SES levels. Thresholded plot analysis identifies seven consistent relationships among MHL, SP, and AAS across SES levels: (a) MHL4 with AAS2, (b) SP5 with AAS3, (c) MHL1 with SP4, (d) SP1 with MHL4, (e) MHL6 with AAS2, (f) SP4 with AAS3, and (g) AAS7 with SP2. Four of these relationships (second, third, sixth, and seventh) exhibited relative stability across SES levels, with edge strengths of [0.15, 0.21; 0.39, 0.40; 0.18, 0.11; 鈭�0.50, 鈭�0.56]. In contrast, the first and fifth relationships displayed a significant decrease with increasing SES, as indicated by the interaction effects modulating edge strength [0.45, 0.16; 鈭�0.33, 鈭�0.14]. The fourth relationship, however, showed a significant increase with higher SES, as reflected by the interaction effects [ 鈭�0.35, 鈭�0.61].

Fig.听6

Conditional network models at the values of the mean鈥壜扁�塖D of SES. Plotted with the AND rule and a significance threshold of p鈥�&濒迟;鈥�0.05

Figure听7 depicted the relationship between mental health literacy, social participation, and active aging across different SES levels. As expected, at high SES levels, MHL4 had a positive effect on AAS2, and vice versa. Similarly, SP5 positively influenced AAS3 at high SES levels. In contrast, MHL1 exerted a negative impact on SP4 at high SES levels. SP1 also had a negative conditional effect on MHL4 at high SES levels. Additionally, MHL6 negatively affected AAS2 at high SES levels, while SP4 had a detrimental effect on AAS3. Finally, AAS7 negatively impacted SP2 at high SES levels.

Fig.听7

Plots of conditional marginal effects. Each plot represents average values across both relevant interaction terms. In (A), the plot shows the conditional marginal effects of MHL1鈥壝椻�塖ES on SP4. In (B), the plot shows the conditional marginal effects of SP1鈥壝椻�塖ES on MHL4. In (C), the plot shows the conditional marginal effects of MHL6鈥壝椻�塖ES on AAS2. In (D), the plot shows the conditional marginal effects of MHL4鈥壝椻�塖ES on AAS2. In (E), the plot shows the conditional marginal effects of AAS2鈥壝椻�塖ES on MHL4. In (F), the plot shows the conditional marginal effects of SP4鈥壝椻�塖ES on AAS3.In (G), the plot shows the conditional marginal effects of SP5鈥壝椻�塖ES on AAS3.In (H), the plot shows the conditional marginal effects of AAS7鈥壝椻�塖ES on SP2

Centrality, stability, and accuracy

Figure听8 illustrated that the node with the highest values for betweenness, closeness, and strength in the MNMs is AAS7, while MHL4 showed the highest expected influence value. The stability of edge weights and centrality estimates was quantified using the case-dropping subset bootstrap procedure, as shown in Fig.听9. The final correlation stability (CS) coefficients were provided in Table听5. Analysis of Table听5 revealed a CS of 0.75 for pairwise edge weights, indicating that when 75% of the total sample was excluded, at least 95% of the subsamples retain pairwise edge weights correlated with those of the original sample at a value exceeding 0.70. For interaction estimates, the CS was 0.44, signifying that the largest sample drop size where a correlation of 0.70 is maintained across 95% of subsamples was 44%. Overall, the dataset performed well across most metrics, with EI coefficients exceeding 0.25 for all pairwise weights and interaction estimates, suggesting robust stability of the MNMs.

Fig.听8

The Expected Influence (centrality) value (standardized Z scores) of the MHL, SP, and AAS

Fig.听9

Average correlations were plotted with 95% coverage intervals for each subsample size. There were 1000 interactions for cease-dropping

Table 5 Correlation stability coefficients

Discussion

This study analyzed data from older adults to explore the relationships between mental health literacy, social participation, and active aging, while also investigating the moderating role of SES. It marks the first use of moderated network analysis in this context, providing novel and insightful findings.

Current status of mental health literacy and active aging

The results revealed that the average scores for mental health literacy (26.89, SD 9.8) and active aging (50.92, SD 18.29) among older adults in western China were lower than the national averages reported by Jiang et al. [7] and Tian et al. [37]. This discrepancy may be attributed to cohort differences, as this study specifically focused on this population. Socioeconomic disadvantage has a well-established global impact on active aging [15]. As a developing country with vast regional disparities, China鈥檚 remote and rural areas, particularly in the West, face significant economic underdevelopment. Many seniors in these regions live on incomes below one dollar per day and lack access to basic resources, such as clean water, arable land, and adequate healthcare and insurance, which contributes to substantial health inequities [5]. Furthermore, lower educational attainment in this population limits the acquisition of mental health knowledge, competencies, and skills. Traditional Chinese beliefs, which often associate mental illness with spiritual possession, improper childbearing, or ancestral wrongdoing, also contribute to a more pronounced stigma surrounding mental health [63, 64]. Strong evidence links stigma to treatment avoidance and delays in seeking care for mental health issues [65]. Thus, targeted health services and interventions are critical to improving mental health literacy and fostering active aging among these individuals.

Bidirectional relationship between mental health literacy, social participation, and active aging

This research identified a bidirectional relationship between mental health literacy, social participation, and active aging, offering partial support for Hypothesis 1. Specifically, within this network, the strongest bidirectional link was observed between "Attitudes about mental illnesses" and "Developing spiritual wisdom" in the context of mental health literacy and active aging. Stigma and discrimination toward mental illness have remained prevalent from 1976 to 2014 [66]. Positive attitudes toward mental illness are crucial for encouraging help-seeking behavior, disclosure, and rehabilitation, and they go beyond simply the absence of negative views [67]. Wisdom, a personality trait, encompasses spirituality and openness to new experiences鈥攓ualities that are adaptive and can evolve with age and life experience [68, 69]. Thus, a shift from negative to positive attitudes toward mental illness reflects an openness to new experiences and the development of wisdom in older adults. Consistent with this, previous research has shown that as age increases, open-mindedness and a pro-integration attitude toward individuals with mental illness also rise [70]. Wiser older adults often demonstrate better self-care and are more likely to engage in health-promoting behaviors, particularly those that enhance mental well-being [71]. Empirical studies have consistently found that wisdom is significantly associated with mental health outcomes such as life satisfaction, happiness, resilience, and optimism [69, 72, 73]. Thus, the role of wisdom in transforming attitudes toward mental illness in older adults is clear.

SES moderated-relationship between mental health literacy, social participation, and active aging

The moderating role of SES in the relationship between mental health literacy and active aging demonstrated that higher SES levels amplified the mutually reinforcing relationship between attitudes about mental illnesses and engaging in active learning, thereby supporting Hypothesis 2. Previous research indicates that individuals with lower SES are more likely to harbor negative stigma toward mental illness [74]. A study in China further revealed that individuals with higher education tended to perceive that most people held relatively positive and tolerant attitudes toward those with mental illness [75]. The Theory of Planned Behavior [76, 77] posits that attitude is a key determinant in the adoption of target behaviors. Older adults with higher SES typically exhibit more positive attitudes toward mental illness and are more inclined to engage in active learning, which in turn contributes to the destigmatization of mental health issues. However, as SES rises, behaviors related to coping with mental illness may inhibit engagement in active learning. This may be due to the fact that higher SES is often linked to better access to mental health resources and greater perceived social support. Older adults with higher SES and stronger social support tend to hold more favorable attitudes toward seeking mental health services [78], increasing the likelihood of seeking professional help from specialized mental health institutions rather than general healthcare centers [79]. Such adaptive coping strategies may shift the focus away from engaging in active learning about mental illness.

In line with Hypothesis 3, SES moderated the bidirectional relationship between mental health literacy and social participation. Unexpectedly, higher SES levels moderated the inverse relationship between social interaction activities and attitudes about mental illnesses, as well as between knowledge about mental health and cultural and intellectual activities. This contrasts with previous studies suggesting a positive influence of social participation on mental health literacy [24]. The discrepancy may be explained by the unique circumstances of the COVID-19 pandemic, during which face-to-face interactions were restricted due to social distancing measures. A mixed-methods study indicated that eHealth literacy among older Chinese adults increased during the pandemic [80]. With the widespread use of smartphones and other technologies, older adults with higher SES were more likely to access timely health information and other resources through strong social networks [81]. As a result, they had greater opportunities to engage with mental health-related content, such as online lectures and social media articles, thereby improving their attitudes toward mental illnesses.

In the context of the bidirectional relationship between social participation and active aging, moderated by SES, cultural and intellectual activities were found to negatively impacted building financial security, while family activities had a positive impact, thus supporting Hypothesis 4. As SES levels increased, cultural and intellectual activities were associated with a lower likelihood of building financial security. Financial security entails preparing for later life, including funeral expenses, and ensuring sufficient resources for daily living [53]. In this study, cultural and intellectual activities encompassed activities such as chess and mah-jong, reading books and newspapers, engaging in the University of the Third Age, and other intellectual pursuits aimed at enriching leisure time, mitigating cognitive decline, alleviating loneliness, and promoting health behaviors [82,83,84]. In China, participation in these activities often requires payment, limiting their direct economic value. However, as SES increases, family activities contribute to enhanced financial security. Older adults in China frequently face income instability and rely on financial support from their children, with family playing a central role in their post-retirement life [85]. Shaped by traditional Chinese norms of filial piety [86], the intergenerational support model emphasizes reciprocity, reinforcing family bonds while providing financial assistance to older adults [87]. Elderly individuals engaged in housework or caregiving are more likely to receive financial, emotional, and domestic support from their children. As a result, increased financial support from children alleviates the financial burden of older adults, thereby strengthening their financial security.

Furthermore, as SES levels increase, strengthening family ties to ensure care in later life predicted fewer physical exercise activities. This relationship can be attributed to the fact that older adults with higher SES, particularly those living with their children, often take on primary household responsibilities and may also care for grandchildren. Consequently, they allocate much of their time to family-related activities to strengthen family ties, which limits the availability of discretionary time and personal resources for physical activity [88].

Limitations

Several limitations must be considered. First, the cross-sectional nature of the data precludes the assessment of causality in the relationship between mental health literacy and active aging. Future research could utilize longitudinal data and a cross-lagged tracking design to better understand the cause-and-effect dynamics between these variables. Second, this study relies on current monthly income as a sole indicator of economic status. Future investigations could incorporate dynamic, life-course indicators of economic status to provide a more comprehensive understanding of an individual's financial situation at various life stages. Third, while this study employs moderated network analysis to explore symptom-level relationships, further validation of the interactions between different dimensions of mental health literacy and social participation is needed to clarify their effects on active aging.

Conclusion

This study examined the symptom-level relationship between mental health literacy, social participation, and active aging, with SES serving as a moderating factor. Results indicated that mental health literacy among older adults in western China was notably low. A bidirectional relationship was observed between mental health literacy, social participation, and active aging. Furthermore, SES moderated the bidirectional relationship between mental health literacy and active aging, as well as between social participation and active aging. This exploratory analysis deepens the understanding of these symptom-level dynamics and highlights the need for policy interventions targeting socioeconomic disparities to promote active aging in later life.