Impact statement
This study is based on the middle-aged and older population in China, assessing the relationship between CLAs and depressive symptoms, and exploring the mediating role of pain. The results showed that higher levels of CLA participation were associated with lower risk of depression symptoms, suggesting that brain- and mind-activating leisure activities may help maintain the mental health of middle-aged and older people. We also found that pain played a mediating role in this association, suggesting that CLAs may partially alleviate depressive symptoms by reducing pain perception. Although this mediating effect is relatively small, it provides new clues for understanding the interaction between pain and emotional issues. For China and other countries experiencing population aging, this study provides a reference for preventing mental health problems among older adults. In addition to traditional medical interventions, older people are encouraged to maintain regular CLAs in communities and families, which may become a simple and low-cost strategy for preventing depression symptoms and improving physical and mental well-being. This work has certain reference value for optimizing public health policies and service systems for older people.
Introduction
China has the world’s largest middle-aged and older population, and is experiencing a rapid aging process. The physical health problems of the Chinese population are becoming increasingly prominent, and their mental health problems are also becoming increasingly apparent. According to epidemiological research, the prevalence of reported depressive symptoms among people aged 45 years and over was 24.1% (Fan et al., Reference Fan, Guo, Ren, Li, He, Shi, Zha, Qiao, Zhao, Li, Pu, Liu and Zhang2021). Depressive symptoms not only significantly reduce their life quality and increase family caregiving burdens, but also are closely related to a higher risk of various chronic diseases, such as cognitive decline and cardiovascular disease (Zhou et al., Reference Zhou, Ma and Wang2021; Huang et al., Reference Huang, Xu, Dai, Li and Tu2024). It is worth noting that when depressive symptoms persist and reach a certain severity, they can develop into severe depressive disorder, which is the main cause of suicide in older people. Symptoms revealed based on diagnostic scales range from 5.37% to 56% (Obuobi-Donkor et al., Reference Obuobi-Donkor, Nkire and Agyapong2021), posing severe challenges to the allocation of societal healthcare resources.
In research on maintaining mental health in middle-aged and older adults, the function of leisure activities has gradually gained academic attention. Cognitive leisure activities (CLAs) refer to activities that individuals engage in outside of daily life for enjoyment and well-being. These activities are believed to stimulate cognitive and intellectual development, and positively influence older adults’ cognitive function and mental health (Hallgren et al., Reference Hallgren, Owen, Stubbs, Zeebari, Vancampfort, Schuch, Bellocco, Dunstan and Trolle Lagerros2018; Zou et al., Reference Zou, Herold, Cheval, Wheeler, Pindus, Erickson, Raichlen, Alexander, Muller, Dunstan, Kramer, Hillman, Hallgren, Ekelund, Maltagliati and Owen2024; Zhang et al., Reference Zhang, Chen, Yu, Li, Zou, Mavilidi, Green, Owen, Hallgren, Raichlen, Lu, Alexander, Paas and Herold2025b). Previous studies have shown that the participation of middle-aged and older people in CLAs is associated with depressive symptoms, but this association may exhibit significant heterogeneity across different cultural and socioeconomic backgrounds. In Western countries, relevant research mostly focuses on structured cognitive activities such as reading and puzzle games, with their negative association with depressive symptoms generally confirmed (Nygaard et al., Reference Nygaard, Mosing, Finkel, Franz, Kremen, Martin, Plassman, Reynolds, Sachdev, Thalamuthu, Walters, Christensen, Pedersen, Gatz, McGue and Consortium2025). In the Chinese population, common local activities, such as Tai Chi, calligraphy and chess and card games, are the main forms, which, to some extent, lead to cultural specificity in the mechanism and intensity of their effects. Although a negative correlation between CLA and depression symptoms is established in middle-aged and older people in China (Zhang et al., Reference Zhang, Sun and Jin2025a), the relevant evidence is still relatively limited, especially based on a nationwide large sample cohort and long-term follow-up, and there is still a lack of systematic research on this relationship.
Pain, as a common physical complication, is prevalent in middle-aged and older populations and is associated with the presence of depressive symptoms. Chronic pain can increase the risk of depression symptoms in middle-aged and older people through various pathways, such as disrupting sleep, inducing neuroendocrine disorders, restricting daily activities and affecting the function of emotion-related brain areas (Li et al., Reference Li, She, Zhu, Li, Liu, Luo and Ye2025a). Research has shown that middle-aged and older people with chronic pain have a 1.32- to 3.38-fold increased risk of developing depression (Marcuzzi et al., Reference Marcuzzi, Skarpsno, Nilsen and Mork2022; Zeng et al., Reference Zeng, Liao, Lin, Zou, Chen, Liu and Zhou2024). It is worth noting that CLA is suggested to alleviate pain perception through mechanisms such as distraction, emotion regulation and improved social interaction (Cournoyer Lemaire and Perreault, Reference Cournoyer Lemaire and Perreault2024). A cross-sectional study showed a negative correlation between the frequency of cognitive activities (such as listening to music and reading) and pain intensity (Baday-Keskin and Keskin, Reference Baday-Keskin and Keskin2025). These findings indicate a correlation between depressive symptoms, CLAs and pain.
However, it is worth noting that most of the evidence on CLAs, pain and depression mentioned above comes from cross-sectional studies, which mainly reflect correlations at the same time point and lack longitudinal tracking at multiple time points, making it difficult to confirm the time-series relationships between variables in the study. Longitudinal tracking cohort studies can reveal the temporal sequence and dynamic changes between these factors through long-term follow-up, thereby more effectively identifying key time windows where pain and depression interact, and evaluating the long-term outcomes of individual depressive symptoms. Therefore, to fill the research gap, this study utilizes representative longitudinal cohort data from the CHARLS from 2011 to 2018. It aims to explore the longitudinal relation between CLA participation and depressive symptoms in middle-aged and older adults, and further study the mediating role of pain. Compared with previous studies, this study focuses on the longitudinal relationships among the three factors, using a large sample size and medium- to long-term follow-up in China, providing a scientific basis for developing more context-specific strategies to promote mental health among middle-aged and older people.
Methods
Data source and study sample
This study utilized data from the CHARLS database. The CHARLS database comprises high-quality microdata from Chinese households and individuals aged 45 years and above. It serves as a critical resource for examining demographic aging and fostering multidisciplinary studies in this field (https://charls.pku.edu.cn/). The data are publicly available, and the data collection process was approved by the Peking University Institutional Review Board. Given that written informed consent was acquired from every participant, the study was exempt from ethical approval. This study utilized survey data from 2011 to 2018 to examine the association between the CLA Score (CLAS) and the incidence of depression among the elderly (n = 17,705). Individuals aged under 45 years (n = 391), with missing data on CLAS-related variables (n = 1,556), with depressive symptoms in baseline (n = 6,421), with missing depressive symptoms data (n = 4,561) and with other covariates missing (n = 640) were excluded. There were a total of 4,136 participants in the final analysis. The specific covariates missing are as follows: sex (n = 4), education (n = 1), sleep (n = 8), BMI (n = 609), diabetes (n = 2), hypertension (n = 4) and dyslipidemia (n = 19). Figure 1 illustrates the process of sample selection.
Inclusion and exclusion criteria.
Variable selection
CLA/CLAS measurement
In this study, CLA/CLAS was the independent variable. CHARLS interviewers inquired about participants’ engagement in the provided social activities during the past month. The second response option (“playing mahjong, chess, cards or attending a community club”), the ninth option (“stock investment”) and the tenth option (“going online”) were all identified as CLAs (Sheng et al., Reference Sheng, Chen and Qu2024). Each CLA was scored based on participation frequency as Not participating (score = 0), Not regularly (score = 1), Almost every week (score = 2) or Almost daily (score = 3). Adding up the three CLA scores yields the CLAS, with a total score on a scale from 0 to 9 (Supplementary Table S1). Due to the generally low scores among participants, CLAS was categorized into four groups for statistical analysis: Group 1 (0 points), Group 2 (1 point), Group 3 (2 points) and Group 4 (> 2 points).
Measurement of depressive symptoms
Depressive symptoms were the primary outcome variable in this study. In the CHARLS survey, depressive symptoms were assessed using the CES-D scale. This scale includes 10 items, each with four response options scored from 0 to 3. The total score ranges from 0 to 30, with lower scores indicating fewer depressive symptoms. Scores ≥10 were defined as indicative of depression (Andresen et al., Reference Andresen, Malmgren, Carter and Patrick1994). The assessment of depressive symptoms was based on the response questionnaire of each follow-up survey, and the CES-D-10 scale only reflects the degree of depressive symptoms assessment rather than the diagnosis of clinical depression. The assessment of depressive symptoms in this study was based on a summary of questionnaire responses from three follow-up surveys conducted in 2013, 2015 and 2018, with an average follow-up duration of 5.7 years. The cumulative incidence of depressive symptoms in the analyzed sample was 42.9%.
Covariates
Covariates in this study included age (years), sex (female/male), education level (illiterate/elementary and below/middle and above) (Chen et al., Reference Chen, Zhao, Zhang, Lin, Shen, Wang, Jia and Yang2024), marital status (single/married), sleep (hours), smoking (no/yes), alcohol drinking (no/yes), BMI (kg/m2), hypertension (no/yes), diabetes (no/yes) and dyslipidemia (no/yes). Marital status was categorized as “married” for individuals currently cohabiting with a spouse or married but temporarily separated due to reasons such as work, and as “single” for those who were separated, divorced, widowed or never married. Hypertension was defined as meeting any of the following criteria: (1) systolic/diastolic blood pressure ≥ 140/90 mmHg, (2) self-reported physician diagnosis or (3) recent use of antihypertensive medication (Zhang et al., Reference Zhang, Zhao, Lu, Meng and Zhou2023). Diabetes was defined as meeting any of the following conditions: (1) fasting blood glucose ≥126 mg/dL, (2) glycated hemoglobin level ≥ 6.5%, (3) self-reported physician diagnosis or (4) recent use of glucose-lowering medication (Zhang et al., Reference Zhang, Zhao, Lu, Meng and Zhou2023). Dyslipidemia is considered if any one of the criteria is met: (1) total triglycerides ≥150 mg/dL, total cholesterol ≥240 mg/dL, high-density lipoprotein cholesterol <40 mg/dL or low-density lipoprotein cholesterol ≥160 mg/dL; (2) self-reported physician diagnosis; or (3) recent use of lipid-lowering medication (Zhang et al., Reference Zhang, Zhao, Lu, Meng and Zhou2023).
Mediating variable
Pain served as the mediating variable in this study. In CHARLS, participants were asked, “Do you often feel physical pain?” Those who answered “yes” were then asked, “Which part of your body is experiencing pain? Please list all the body parts that you are currently experiencing pain in.” Participants could choose from multiple body parts (15 parts: head, shoulders, arms, wrists, fingers, chest, abdomen, back, waist, hips, legs, knees, ankles, toes and neck), and finally summarize the number of parts with pain for mediation analysis (Ye and Wang, Reference Ye and Wang2024).
Statistical analysis
In this study, baseline characteristics of the study population were presented by categorizing participants into four groups based on their CLAS. Continuous variables were expressed using means and standard deviations, while categorical variables were presented as frequencies (n) and percentages (%). Group comparisons were made by applying analysis of variance to continuous variables and the chi-square tests to categorical variables. Due to the outcome variable in this study being defined as the first occurrence of depressive symptoms during the follow-up period, each participant only contributed one endpoint event, and the data structure did not have hierarchical features of repeated measurements or changes over time. Consequently, a multilevel model was not used. Instead, multivariate logistic regression was used for analysis, and random effects were not introduced. Multivariate logistic regression was employed to assess the relationship between CLAS and depressive symptoms, as well as to explore regional differences between CLA and depression across China. Differences in the association between CLAS and depressive symptoms risk across age and sex groups were evaluated by subgroup analyses. In mediation analysis, pain (as a continuous variable) was treated as the mediator variable, and its effect was tested using the bootstrap method with a repeated sampling of 1,000 times. R.4.1.3 was used to conduct statistical analyses. Baseline characteristics were generated using the “tableone” package, logistic regression analyses were conducted with the “stats” package, subgroup analyses and interaction tests were conducted using the “jstable” package and mediation analysis was carried out with the “mediation” package. P < 0.05 was considered statistically significant.
In addition, we conducted a sensitivity analysis to evaluate the impact of confounding factors such as physical activity (PA) and cognitive function on the research results. PA is obtained through standardized tools based on the International Physical Activity Questionnaire (IPAQ). Participants reported their weekly frequency (days/week), duration and intensity of PA (classified as vigorous, moderate or mild). Given that CHARLS only provides time intervals for each type of PA, we have converted these responses into average time intervals to more accurately quantify the duration of PA. Subsequently, the metabolic equivalent task (MET) score was calculated based on the time allocated to each participant within a week. The formula for calculating the MET multiplier is as follows: MET score = 8.0 × intense PA time per week + 4.0 × moderate PA time per week + 3.3 × mild PA time per week (Li et al., Reference Li, Guan and Wang2025b). Cognitive function assessment includes two dimensions: episodic memory and mental state. Episodic memory consists of immediate memory and delayed memory. Participants were asked to repeat the 10 Chinese words just read to them in any order (immediate memory) and recall these words after 5 min (delayed memory), with 1 point awarded for each correct answer. The average score of immediate memory and delayed memory was calculated as the episodic memory score, ranging from 0 to 10 points. The assessment of mental state included two parts: the Telephone Interview of Cognitive Status (TICS-10) and graphic drawing. The telephone interview test requires the interviewee to report the date of the day (including month, day, year and season) and the day of the week (0–5 points), and to subtract 7 from 100 for 5 consecutive trials (0–5 points). In the graphic drawing test, participants judge and draw the correct shape based on the displayed overlapping pentagon images (0–1 points). The overall cognitive score is the sum of the episodic memory score and the mental state score, with a score range of 0–21 points (Jiang and Jiang, Reference Jiang and Jiang2025). Covariate data with multiple missing imputations in the sensitivity analysis were reanalyzed using the “mice” package.
According to the definition provided by the National Bureau of Statistics of China (2011), Mainland China was divided into four regions (Eastern, Central, Western and Northeastern) to explore geographical differences in the risk of depression among populations with varying CLAS levels.
Results
Baseline characteristics
Table 1 shows the baseline characteristics of the population in this study. The final sample comprised 4,136 individuals, with a mean age of 56.8 ± 7.8 years. Specifically, respondents with higher CLAS levels were relatively older and had higher BMI values. They also showed higher proportions of males, higher education (middle and above), smoking (yes), alcohol drinking (yes), dyslipidemia (yes) and absence of depressive status, along with fewer reports of pain (P < 0.05).
Baseline characteristics
Note: Categorical variables are presented as n (%). Continuous variables are presented as mean (SD).
Pain: Number of pain types.
Association between CLAS and depressive symptoms risk
Table 2 presents the results of the logistic regression analysis examining the association between CLAS and depressive symptoms. In Model 3 with comprehensive adjustment for all variables, Group 4 was associated with a lower risk of depressive symptoms (OR = 0.58, 95% CI: 0.44–0.76, P < 0.001). Trend tests indicated a decreasing trend in depressive symptoms risk across all models.
Logistic regression analysis of the association between CLAS and depression
Note: Model 1: Unadjusted; Model 2: Adjusted for age and sex; Model 3: Adjusted age, sex, educational level, marry status, sleep, smoke, alcohol drinking, BMI, hypertension, diabetes and dyslipidemia.
Subgroup analysis of the association between CLAS and depressive symptoms
Supplementary Figure S1 illustrates the relationship between CLAS and depressive symptoms across different age and sex subgroups, as evaluated through subgroup analysis. The results showed that in individuals aged 60 years and above (60–70 years old: OR = 0.38, 95% CI: 0.22–0.66, P < 0.001; Age ≥ 70 years: OR = 0.36, 95% CI: 0.14–0.92, P = 0.033) and overall population (female: OR = 0.63, 95% CI: 0.43–0.93, P = 0.019; male: OR = 0.53, 95% CI: 0.36–0.78, P = 0.001), Group 4 showed a significant negative correlation with depressive symptoms. There were no notable sex-specific differences. Furthermore, no subgroups were identified with vital interaction effects (P-interaction >0.05).
Regional differences in the association between CLA and depressive symptoms
Supplementary Figure S2 displays the proportional distribution of CLAS groups across different regions of China. The Western region recorded the highest CLA nonparticipation rate (82.2%), and the Central region the lowest (69.4%). Conversely, the northeast showed the highest active participation rate (9.1%), compared to the lowest in the west (4.6%). Results from the logistic regression model, as presented in Table 3, indicated that compared to Group 1, Group 4 was associated with a lower risk of depressive symptoms in both the Eastern (OR = 0.47, 95% CI: 0.29–0.78, P = 0.003) and Western (OR = 0.33, 95% CI: 0.17–0.63, P < 0.001) regions.
Logistic regression model for the association between CLAS and depression across regions
Note: Group 1 (0 points), Group 2 (1 point), Group 3 (2 points), and Group 4 (>2 points).
Mediation analysis of pain in the relationship between CLAS and depressive symptoms
The path coefficients showed that the three paths in the model, namely from CLAS to pain (β = −0.093, 95% CI: −0.150, −0.037; P = 0.001), from pain to depressive symptoms (β = 0.198, 95% CI: 0.158–0.238; P < 0.001) and from CLAS to depressive symptoms (β = −0.163, 95% CI: −0.230, −0.096; P < 0.001), were statistically significant (Supplementary Table S2). The results of the mediation analysis (Figure 2 and Supplementary Table S3) indicate that CLAS had a great direct effect on depressive symptoms (P < 0.001), as well as a significant indirect effect mediated through pain (P = 0.002). The mediation proportion was 10.0%.
Mediation analysis. In the diagram of mediating analysis, path C represents the total effect, while path C’ represents the direct effect. The indirect effect is estimated as the product of path A and path B (path A*B). The mediation proportion is calculated as [indirect effect/(indirect effect + direct effect)] × 100%.
Sensitivity analysis
The multiple imputation results showed that compared with Group 1, the CLAS of Group 3 (OR = 0.78, 95% CI: 0.62–0.98, P = 0.031) and Group 4 (OR = 0.61, 95% CI: 0.47–0.77, P < 0.001) were associated with a lower risk of depressive symptoms (Supplementary Table S4). This discovery enhanced the robustness of our conclusion.
After adjusting for confounding factors such as PA (OR = 0.54, 95% CI: 0.35–0.83, P = 0.006) and cognitive function (OR = 0.70, 95% CI: 0.52–0.92, P = 0.014), the results showed that the group with higher CLAS scores (Group 4) was still significantly associated with a reduced risk of depressive symptoms, and the research results were somewhat robust (Supplementary Table S5).
Discussion
By analyzing the CHARLS cohort data from 2011 to 2018, this study systematically investigated the association between CLA levels and the risk of depressive symptoms among middle-aged and older adults in China, and validated the mediating role of pain. The results demonstrated that a higher CLAS was notably related to the reduced risk of depression. This association exhibited heterogeneity across populations of different ages, sexes and regions, with pain playing a mediating role in the relationship between CLAS and depressive symptoms.
This study found that CLAS Group 4 is associated with the alleviation of depressive symptoms. This finding is in agreement with previous research suggesting that social activities can effectively alleviate loneliness and social isolation among the elderly, thereby reducing depression risk (Wang et al., Reference Wang, Liu, Yang, Chen, Wang and Liu2024). A study conducted in China also showed that higher participation in social and leisure activities was correlated with lower levels of depression among older adults (Gao et al., Reference Gao, Jia, Zhao and Han2023). Furthermore, activities included in this study, such as playing mahjong, chess and cards, possess strong social attributes that can increase social interaction frequency and reduce feelings of loneliness, which are closely associated with increased symptoms of geriatric depression (Robb et al., Reference Robb, de Jager, Ahmadi-Abhari, Giannakopoulou, Udeh-Momoh, McKeand, Price, Car, Majeed, Ward and Middleton2020; Lee et al., Reference Lee, Pearce, Ajnakina, Johnson, Lewis, Mann, Pitman, Solmi, Sommerlad, Steptoe, Tymoszuk and Lewis2021). Meanwhile, activities like stock investment and internet usage require the mobilization of cognitive functions, including attention and memory. Long-term engagement in these activities can enhance cognitive abilities and delay brain function decline, aiding in preventing adverse outcomes related to depression (Lu et al., Reference Lu, Wang, Wang, Du, Li, Wei, Yao, Zhang, Yin and Ma2025).
However, it is worth noting that we found a significant negative correlation between CLA scores and depressive symptoms only in the highest CLAS group, while no clear effect was observed in the middle two groups. This suggests that CLA may require a high level of participation to bring significant psychological benefits. This result is also consistent with some studies on leisure activities and cognitive activities (Zhu et al., Reference Zhu, Qiu, Zeng and Li2017; Alves et al., Reference Alves, Dias, Neiva, Marinho, Marques, Sousa, Loureiro and Loureiro2021). This result indicates that, for middle-aged and older people, in addition to encouraging them to participate in CLA, we should also focus on helping them reach and maintain a relatively high and stable level of participation in terms of frequency and duration, to reduce the risk of depression. It should also be noted that although the effect of the highest CLAS group is statistically significant, its proportion in the overall sample is relatively small. Therefore, this discovery should be regarded as an exploratory result. Future studies with larger sample sizes are needed to validate the robustness of this association.
Further subgroup analysis revealed that the negative correlation between CLAS and depressive symptoms was more pronounced among individuals aged 60 years and above, while remaining consistent across both sexes. This pattern may be closely related to life-course characteristics across different age groups: those over 60 years are typically retired and, consequently, have more time to participate in various social and leisure activities (Kalbarczyk and Lopaciuk-Gonczaryk, Reference Kalbarczyk and Lopaciuk-Gonczaryk2022; Tunney et al., Reference Tunney, Henkens and van Solinge2023). A study from America indicated that the frequency of engagement in leisure activities is linked to depressive symptoms in older adults. Compared with the elderly without engagement in sports, social or other clubs, those who participate monthly are lower after 2 years, and the same applies to those who participate weekly (Bone et al., Reference Bone, Bu, Fluharty, Paul, Sonke and Fancourt2022). In contrast, individuals under 60 years are generally still in their professionally active years. Those employed may have relatively fewer opportunities for social and leisure activities due to busy work schedules. Work pressure and family responsibilities might exert stronger influences on depressive symptoms risk, potentially diluting the protective effect of CLA.
Regional analysis revealed significant geographic heterogeneity: high CLAS in the eastern and western regions was significantly associated with a reduced risk of depressive symptoms, while no statistical significance was observed in the central and northeastern regions. A previous study has suggested a correlation between socioeconomic status and depression risk (Lorant et al., Reference Lorant, Deliege, Eaton, Robert, Philippot and Ansseau2003), and some epidemiological studies have also suggested that socioeconomic status, education and access to health resources may affect leisure activity participation patterns and their psychological health effects (Henking and Gondek, Reference Henking and Gondek2023; Wang et al., Reference Wang, Liu, Yang, Chen, Wang and Liu2024). Based on this, we speculate that the differences in socioeconomic development level, cultural atmosphere and resource accessibility in different regions may, to some extent, shape the relationship pattern between CLAS participation and depression. However, it should be emphasized that this study did not directly measure specific environmental factors at the regional level, and the above explanation is still a hypothetical inference. The relevant mechanisms need to be further tested in studies that include more empirical variables at the regional and community levels.
It is noteworthy that this study revealed the partial mediating role of pain in the correlation between CLAS and depressive symptoms. These findings suggest that higher levels of CLAS participation are associated with a lower risk of depressive symptoms, and this association may be partially mediated by reduced pain perception. Active participation in CLA may not only directly improve emotional states but also indirectly reduce the risk of depressive symptoms by alleviating pain perception and enhancing quality of life. Previous studies have indicated that pain not only significantly impairs physical health and daily functioning in older adults (Li et al., Reference Li, Zhu, Li, Huang and Yang2021) but also directly contributes to unpleasant experiences (Wu et al., Reference Wu, Cornally, O’Donovan, Kilty, Li and Wills2025). CLA may promote the redistribution of cognitive resources in the brain, thereby attenuating the processing of pain signals. This mechanism could ultimately achieve effects such as distraction and improved pain coping strategies, reducing pain perception among older adults (Hartwigsen, Reference Hartwigsen2018; Khera and Rangasamy, Reference Khera and Rangasamy2021). This series of associations may help reduce the feelings of loneliness and helplessness caused by pain, thereby achieving the effect of preventing depression. Furthermore, participation in physical and cultural leisure pursuits is also associated with functional mobility and self-care capabilities among older adults (Komatsu et al., Reference Komatsu, Obayashi, Tomioka, Morikawa, Jojima, Okamoto, Kurumatani and Saeki2019), which may reduce the vicious cycle of pain exacerbation caused by restricted activity, thereby reducing the risk of depression symptoms.
This study revealed that pain partially mediates the relationship between CLAS and depressive symptoms (with a mediation ratio of ~10%). This suggests that higher levels of CLAS may partially reduce the risk of depression by reducing pain perception. Specifically, CLA may alleviate the experience of pain and its associated negative emotions by distracting attention, improving pain coping strategies (Hartwigsen, Reference Hartwigsen2018; Khera and Rangasamy, Reference Khera and Rangasamy2021), and enhancing daily activity function (Komatsu et al., Reference Komatsu, Obayashi, Tomioka, Morikawa, Jojima, Okamoto, Kurumatani and Saeki2019), thereby indirectly reducing the occurrence of depression. However, a 10% mediation ratio suggests that pain may only be one of the pathways through which CLAS affects depression, and its effect is relatively limited. The psychological benefits of CLAS are more likely to be achieved through multiple channels, including enhanced social support, improved cognitive function, improved self-efficacy and sense of meaning in life (Quaglia et al., Reference Quaglia, Zeidan, Grossenbacher, Freeman, Braun, Martelli, Goodman and Brown2019; Weziak-Bialowolska et al., Reference Weziak-Bialowolska, Bialowolski and Sacco2023). Pain can be regarded as an important but not the main link in the process of depression affected by CLAS. Relying solely on pain improvement may not fully bear the psychological benefits of CLAS. However, combining pain management with promoting sustained participation in CLAS among middle-aged and elderly people may have a synergistic effect in reducing the risk of depression (IsHak et al., Reference IsHak, Wen, Naghdechi, Vanle, Dang, Knosp, Dascal, Marcia, Gohar, Eskander, Yadegar, Hanna, Sadek, Aguilar-Hernandez, Danovitch and Louy2018; Cui et al., Reference Cui, Duan, Du, Yang, Tian and Liu2025). It is worth noting that the measurement of pain in this study was limited to the number of affected body parts and did not include dimensions such as intensity and duration; moreover, the proportion of mediating effects was not high. Therefore, the results of this study on the mediating effect of pain are better regarded as exploratory findings, aimed at providing clues for subsequent mechanistic research. The study is not suitable for drawing strong conclusions on causal mechanisms. It is necessary to replicate and validate this finding while incorporating more potential mediating factors in future studies.
Our study is subject to some limitations. (1) Limitations of measurement and variable manipulation: First, CLAS may have recall bias based on self-reported activity frequency; second, the measurement of pain mainly depends on the number of painful areas. Although an increase in pain sites is an important indicator of pain spread and central sensitization, due to the limitations of the CHARLS database, research cannot collect specific intensity, duration and functional impact data for each pain site. Therefore, this mediation analysis can only explore the pathways of pain burden in multiple sites and cannot reveal the specific impact of pain on depression through dimensions such as intensity or loss of function, which may affect the estimation of mediation effects. Third, because the proportion of samples with CLAS scores ≥ 3 was extremely low, we merged this population into the “CLAS score ≥ 2” group to ensure statistical validity. Although this approach is necessary in practice, it may mask heterogeneity within high participation groups, thereby potentially weakening the precise dose–response relationship between cognitive activity and health outcomes. Finally, the cognitive leisure activity index constructed based on existing literature (Sheng et al., Reference Sheng, Chen and Qu2024) and the CHARLS data in this study is a more comprehensive proxy indicator, and its effectiveness and reliability may be limited. (2) Research design and sample representativeness limitations: First, to meet the data integrity requirements of longitudinal analysis, over 75% of the initial sample was excluded, which may lead to selection bias and affect the representativeness of the results for the overall middle-aged and older population. Therefore, caution should be exercised when extrapolating the conclusions of this study to the overall middle-aged and older population in China, especially to subgroups that are weaker or have more missing data. Second, the queue study design cannot fully establish causal relationships, and future randomized controlled trials are needed to verify the intervention effect of CLA. (3) Limitations of data analysis and statistical validity: First, although this study used multiple imputation for sensitivity analysis and confirmed the robustness of the results, it initially relied mainly on complete case analysis. If the mechanism of missing data is related to unobserved confounding factors, there may still be potential bias in the results of the complete case analysis. Second, although we observed stronger association trends in some subgroups, due to heterogeneity and statistical power issues within subgroups, the CI of the relevant estimates was wide and the statistical power was insufficient. Therefore, these findings should be considered as exploratory results, and their stability and universality still need to be validated in future studies with larger sample sizes. Third, the results of this study only showed statistical significance in the group with the highest activity level, but the sample size of this group is relatively limited, which may affect the accuracy and extrapolation of effect estimation. It is necessary to develop more structured and refined standardized tools for future research to more accurately capture the heterogeneity of CLA and its impact on mental health. (4) Limitations in research scope and conceptual coverage: The study only used three cognitive activities with a focus on social nature and did not include other types of cognitive activities; this may limit the comprehensive capture of cognitive stimulus exposure levels in the study.
Conclusion
This study found that a higher CLAS was strongly associated with a reduced risk of depression, particularly among people aged 60 years and above and those residing in Eastern and Western regions of China. Pain was confirmed as a mediator in this relationship. Promoting diverse CLA among middle-aged and older groups represents an important public health strategy for preventing depression. Simultaneously, greater attention should be given to pain management, as alleviating pain may enhance the mental health benefits derived from CLA. Future research should further investigate the differential effects of various CLA types to inform the development of personalized intervention strategies.
Open peer review
To view the open peer review materials for this article, please visit http://doi.org/10.1017/gmh.2026.10145.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/gmh.2026.10145.
Data availability statement
The data and materials in the current study are available from the corresponding author on reasonable request.
Author contribution
Min Fang and Dan Chen conceived of the study, participated in its design and interpretation and helped draft the manuscript. Yaqin Wang participated in the design and interpretation of the data and drafted the manuscript. Hongzhu Zhang conducted the statistical analysis and critically revised the manuscript. All the authors read and approved the final manuscript.
Financial support
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests
The authors have no conflicts of interest to declare.
Ethics approval
Ethical approval and consent were not required as this study was based on publicly available data.
Open access
Comments
Dear Editor,
I am pleased to submit our original research article, “Association Between Cognitive Leisure Activities and Depression in Chinese Middle-Aged and Older Adults: A CHARLS Cohort Study from 2011 to 2018,” for consideration as a Research Article in your journal. This longitudinal cohort study of 4,136 middle-aged and older Chinese adults demonstrates that high engagement in cognitive leisure activities (e.g., games, internet use) significantly reduces depression risk by 42%, with pain mediating 10.0% of this protective effect. Notably, subgroup analyses revealed heightened benefits for adults over 60 and residents of Eastern and Western China, highlighting important demographic and geographic considerations for public health interventions.
Our findings provide actionable insights for developing region-specific strategies to mitigate depression through accessible leisure activities—an urgent priority for rapidly aging societies like China. Beyond documenting these associations, our mediation analysis advances mechanistic understanding of how cognitive leisure activities confer mental health benefits, addressing a critical gap in longitudinal geriatric mental health research. The identification of pain as a mediating pathway represents a significant contribution to the field, offering new targets for clinical and community-based approaches.
We confirm that this manuscript is original, unpublished, and not under consideration elsewhere. All authors declare no conflicts of interest and have approved this submission in accordance with your policies. We appreciate your consideration and welcome reviewer feedback to further strengthen this work.
Thank you and best regards.
Sincerely,
Dan Chen